LIVE from SaaStr Annual 2024- Workshop Stage - Thursday

Sa. Sam. Ram. It. It. Ladies and gentlemen, please take your seats.

Our show will begin in a few few minutes.

It. Sam sa. Sam. It. Ladies and gentlemen, please take your seats.

Our show is about to begin.

Sam. Dancing. Please welcome to the stage philip wraithly, cto at neo4j.

Hello.

Great to be here. So today I want this to be a conversation. I'm going to warm things up and just share what to me are some of the key takeaways from my experience at Neo4J really starting 2012. So a lot of you either are past the stage or maybe at the stage where as a product founder you want to find past a certain point your product becomes not just the product but your company becomes the product and there's no longer enough time to actually be involved in the day to day roadmapping and so on. And so you hire a head of product that's like minded and they can sort of channel your vision and work together. So that was me. I was the first product hire in 2012, ran product for around 11 years. So till, till last year moved into a CTO position and really thought today I would reflect on my lessons learned and some of the key key points. So Some of this Neo4J is a horizontal infrastructure company. I realize a lot of you may be more verticalized, but I think some of these lessons will still hold and then I'd love to have questions for probably 15 minutes or so at the end, give or take. So let's, let's dive in. The first one is work out whether your value prop is horizontal, have conviction in that or vertical and it's really easy to get tempted into doing more than just one. So like speaking as a horizontal company, literally every week this happened to me earlier this week someone said, well Neo4J is so great. I can see how I can just through the visualization tool just easily be able to take my domain and this was something in finance and trading and being able to look at companies and ownership and so on and hey, can we do, I'd love to do some kind of co investment. I'll throw you some money, we'll build a new product. This is very, very tempting and of course it's great to get feedback, right? So I love that sort of feedback. Sometimes feedback leads to more incremental changes that are in the direction of your vision and where you want to take things. In this case that feedback was leaning towards something that could be a huge distraction right from a horizontal opportunity. A I could end up competing with my customers. B if I end up taking this on and building some kind of narrow solution in a single space, then there's an opportunity cost and, and that takes me away from any energies that I might spend building the horizontal stack. And then furthermore, like if you look outside of product, while something may look trivial to build inside of product, it's always not right. Then you end up having to market, yet it's a different value prop, different price points, different market. You're going to need people who are experts in finance and so on and so forth. So, so we fought the urge to go vertical, but there is one case where it's actually been important to do so to a degree is our motion's been very developer first, but then a developer will find the technology, bring it into a project, and inside of that project you end up needing to talk to the business or people in it who are talking to the business. So while I'd strongly advise either staying in a horizontal lane or staying in a vertical lane, and if you have conviction in the horizontal opportunity, then there's no need to dally with other opportunities or likewise you have conviction in your vertical opportunity and someone's pulling you sideways to do something in a different domain, different industry, there's probably a stage where that makes sense, but I'd say that stage is probably posted 100 million ARR, again depending on the domain and on the TAM. So that's the first.

Oh, right.

So on the marketing point we found it's actually really helpful to have like a use cases section and an industry section on the website so that people who've heard about the technology, who are oftentimes looking at things from that angle, can actually see examples. So that ends up being a mapping into case studies and common stories and patterns. And then there's another thing that tends to happen is inside the company you can have a group inside of professional services or if you're open source, anyone from the community can add stuff that includes people inside the company that will add things that are maybe more vertically oriented. And I think that's fine, that's healthy, so long as there's a really clear differentiation of what's in the product and what's not. Like once upon a time we had some, you know, kind of like well meaning salespeople who would go off and take some solution IP that was developed like for basically to demo and to bootstrap a project and to say, all right, we're going to put this, you know, on a price list and okay, but then if you get, if you sell that, then you have to maintain it, then you have to support it. Support people have no idea what's going on. If someone calls in, generally not a good thing. So I think this is something to really look after and be very rigorous and clean with what your boundaries are. So that is number one, number two, also a choice. So it may be the case that your priority product is equally good. It's serving people inside of SMB and people inside of enterprises. And up until we were probably 20 million in ARR, we just catered to whoever's going to come knock at the door and you know, we're fine taking on both. And what we realized is these are materially different certainly from a go to market perspective. So enterprises much longer sales cycles, a lot more checkboxes, a lot more compliance kinds of things that actually ends up driving product significantly. So you end up doing a lot of stuff inside of your product that is undifferentiated. Users won't understand the difference, but that is table stakes for the enterprise. Whereas with SMB you tend to have a more high volume game, which means the product investment depends more on learning and good documentation like in product learning, make ease of use everything around the PLG motion. And that PLG motions are also very different than the enterprise motion. So my advice here would be, you know, probably somewhere around 10, 20 million if not sooner, depending on how clear you are about who's going to buy your technology and what's going to fund the growth of your business to a larger stage to just make a decision, focus on one and then it's always possible to make small investments in the background to handle the other one. So in our case we chose enterprise. Our go to market is entirely enterprise. We won't say no. However, if an SMB comes in and we want those people to succeed, we just don't have all the motions in the company to do that. But we would, you know, it kind of had this slow burn on making the product easier to use, easier to self onboard, et cetera, et cetera. Now we're at this stage, you know, we're, you know, kind of cross 100 million ARR in 2021. Can't talk about revenue now, but you can imagine, you know, north of that we're just now finally starting to get our PLG and SMB motion into play. So it took a really long time to get to where we could do both. And then related to this, you need your product to fill your pond. This is the PLG motion. And then you also need to be.

Able to Fish it.

One thing I often see in the Valley is people getting really excited and VCs also getting very excited about how many users you can bring in and how many downloads you can get and how many registrations you can get, how many monthly active users you can get on free versions and on open source versions. And that's great, that is filling your pond. But you also need to think very early and again draw very clean, a clean philosophy about what you want to monetize, in what situations you want people to use the product for free, in what situations you want them to pay. So it may be that you're okay with people using the product for free for anything but up to a small size or up to a small number of users or for things that are maybe not mission critical. So in our case we drew a clean line early on. Actually we experimented with a license that ended up being confusing to people. So we retired it. But the philosophy was we're happy for. So we've got the Community Edition, that one is let's say feature complete for a developer. And then you have the Enterprise Edition, which is feature complete complete for it Ops, lawyers, managers, sysadmins for doing like larger scale kind of stuff. And the Enterprise Edition, actually we used to say you have to pay in one of two ways. You either have to pay with money because you're earning money off of our technology. So either you contribute back towards the development of the technology with money, or you contribute back by contributing in open source. And so we actually used the AGPL license for this, which is the closest open source license to affect this particular philosophy. Again, turns out it was confusing. It turns out the AGPL may or may not, you know, hasn't been tried in court. Exactly do what we wanted it to do. So we ran this for a few years, but that tweaked the philosophy a bit and said all right, let's, you know, if let's make open source contributions, contribution to the community more something that happens independent of what edition you use. People in the enterprise will do it, people using the Enterprise Edition will do it, or you know, our SaaS version will do it, as will Community Edition users. But the having a clear philosophy led to a clear, clear set of feature differentiation. So basically things that were more security oriented or ops oriented naturally landed in the Enterprise Edition. Another way I looked at it was community users are in the realm of like being untethered and doing anything they want. And then enterprises are. This is a bit tight, intrigued. But like enterprises are there to prevent people from Doing things and like lock people down down and make sure you stay in the lane. So, you know, it turns out that a lot of the enterprise features are ones that are about restricting and governing use, whereas community features are more about, you know, hey, anyone can do anything. That's kind of the point and you know, is very much within the spirit of Open source. So that's on the differentiation point back to filling the pond. Open source fills the pond in a way, but then you don't know who the people are that use the technology, right? Like if you put a paywall or even a registration wall in front of Open source A, it doesn't make sense because people can go to GitHub and whatnot and just do a poll and get it. But on the other hand, that's kind of like against the whole spirit of what Open source is trying to do, right? Like get the technology, we don't care who you are, there's no barrier. So then you need other ways to actually get people into your community, understand who they are and fill the pawn and actually give like, you know, sdr, someone to engage with in cases where there is a commercial opportunity. And so in those cases what I found is, you know, there's a bit of an irony that, you know, I kind of live by, and probably many of you live by is our expectation is you can get a software product that someone's invested millions, tens of millions, hundreds of millions of dollars without giving your name and then someone can write a book and like that $10 book, make it available as a PDF download and people are very happy to provide their name and get the book in exchange. So you sort of need ancillary ways and motions to be able to.

Get.

People to engage with you in a non anonymous way. And I'll come back to community and the important role of community there. Great. Let's see here. Next one is roadmap. Prioritization requires whole team involvement. And my lesson here is that it's there's two reasons for this. So one reason is you're trying to build a product that caters to all stages of the customer life cycle and each person in the company. It's rare to have one person in the company that has a deep understanding of what customers need at each stage in the life cycle. So in a way I view each part of the org as like a sensory appendage that can detect information about what customers need at different stages. So there's that. There's also the psychology though is people in different orgs end up Being on the hook every day with customers who are facing the same set of problems. And that can become, over time, just really stressful and trying and annoying, right? That hey man, just fix this. I know what it is. I've reported it back. And so it's important. And while it may take years and years to actually knock out all the things that you want to do right, and then customers are always trying to do more. So if you've ever completed your roadmap right, then your company is gone. That means that people aren't pushing the limit. So what I found to be helpful in getting buy in is making sure each person gets at least one thing every release cycle. So every quarter it doesn't have to be the biggest thing. But then you can come back and say, hey, look, we know you've got these five things or these 10 things that are really important, we did one. Or you know, you might be able to do two or three. And depending on the company priorities. Of course. Roadmap should be a reflection of the high level company plan more than anything. But then you should find ways to fold in at least one thing from each department so people feel heard and then communicate it back and then have a venue for people to communicate. And I find the smaller venues the better because people feel like their team and themselves as individual are more heard. There's one related thing I wanted to mention around involvement which is involving the engineering team. So the engineering team should definitely be highly involved in prioritization, obviously because they're responsible for the cost. Engineers often have good feedback and input into roadmap and what the priorities should be. And then of course you want engineers involved in the design and actually talking to the customers and end users so they understand how the feature will be used. One area where I found it's best actually to shield engineers from not involve engineers in is more open ended conversations with customers around prioritization because customers will throw out each customer. You could talk to five different customers and get completely different sets of requirements. It's like, you know, the blind men and the elephant. Like each person has a totally different perspective on the thing and their needs. And engineers are never going to have the breadth of conversations that PM's will and won't be able to contextualize a single conversation. And you want engineers to be up at night thinking about the things that are next on the roadmap that are the company's priorities and are the product priorities. You know, maybe minus some percentage of, you know, free, open, refactoring, whatever you know, creative time. So yeah, so I've learned this a little bit the hard way. Don't bring engineers to individual open ended customer meetings because then there's a risk of that empathy for those issues turning into staying up at night worrying about something that in the larger context is actually not a priority. All right, this one. Anyone read the book Confessions of a Pricing man by Herman Simon? This is one of my favorite books on pricing and he has a section where he talks about three kinds of pricing. Discount, premium and luxury. Luxury. If one of you can figure out how to make a luxury based software product where you know it's like scarcity and status are the things that sell, then I'd love to hear about it. But I think the two that are applicable to us is premium or discount. And understanding who you are is really important with respect to what you build and also how you market how you sell. So premium product tends to be high quality, no compromises. Oftentimes it's higher value in the grand scheme, but it tends to look more expensive at the front end. Whereas a discount product tends to lead to. And then premium product oftentimes are the innovators, though they may not be the first to market with an individual feature. Like you can look at Apple as one example of this. Whereas discount tends to be fast follower. Oftentimes it's like every feature on the sun, but under the sun, but they don't necessarily work that great. But it's. They tend to compare themselves more to the leading product that is the one that they're following. Whereas with a premium product you're looking ahead, you're focused more on your customer, you're obsessed about your customers and you're kind of mildly aware of your competition. Whereas with a discount product you're really in the value game or pricing game and you tend to be very obsessed about your competition and what they're doing. So I found this is an important one. And if you're a category creator or a category leader, then you have no choice, right? You need to be a premium product. So these are. And I'd also say that there's the concept of whole product, which is the customer's experience with the product is not just the software, it's their experience with billing and support and customer success and so on and so forth. I'll just add one super quick bonus topic and then we can go to questions, which is I hear a lot of people building community oftentimes around open source and then immediately going and like expecting all these things of community and that's not the way to think about it. Like, that's not the way it works. You should think about how do I give and contribute to the community and trust that over the long term. Like having a community has all these advantages, but if you go in expecting things from the community, that tends to lead to behaviors that are not very community like and won't be appreciated and tends to work against the community building. All right, let's jump into questions. I think I've probably talked more than I planned to and we have five minutes for Q and A. There's a mic up here if you want to stand up. If you don't, I've got a couple other bonus things. But here, please stand up if you've got. If you've got questions.

Hi.

So question. When you were talking about premium, I could easily identify if you're considered a premium or discount and how you're doing.

You talked about the various ways of considering that.

How do you think about a company.

That you would say has a premium.

Product because it's more flexible and agile than the new thing, but you're able to deliver it at a lower cost than the current think? You know, pricing is very important and emphasizing, but it's not following what's being done.

So I'd be curious.

Yeah, that's a subtle point. You can have a premium product, I guess, if it's the same class of product. Pricing ends up being one of the indicators. But if you're disrupting an existing product and creating something entirely new, then in a way you've created a new class of product and that can have a lower price. I would say in a way, like we're an example of this to the degree that a lot of the technologies that one might alternatively use that are more traditional, say relational technology that are trying to solve a graph problem, both the total cost. The total cost is way lower, but the per unit cost is also lower. So, yeah, that's a subtle and I.

Think a good call out.

What's your recommendation if you go horizontal too early and realize you've built a.

Lot of features that really provide no value, but you do have customers using them.

Oh, but you have a vertical, vertically oriented product.

Yes.

Yeah.

I mean, I think there maybe it's look for partners or, you know, other companies that you can transition or hand them off to and do that as fast as possible, but as like cleanly, transparently. You know, tell them as soon as you can that you're looking for a path and then they can start looking for Solutions as well. So the burden isn't entirely on you, but you know that that creates a lot of drag. So you want to be good to your customer at the same time, you need to do what's right for your. Any other questions? Doesn't have to be related to any of this. I'll throw a couple others out that occurred to me while building this out. One is often we get really tied up in innovation and thinking that innovation is always categorically good. And in my experience, you actually want to depend on familiarity as much as possible possible and then innovate only in cases where that innovation and that new way of doing things adds significant value. So, you know, if I have a product and I can do 80% of the things that are just my normal workflow, and then just 20% of the things are new, but they're like so much better, then it's worth learning the new things. But you don't want to force users to spend energy learning to do things the same way in a different. Same things that are already doing in a different way. This actually came up a lot in language design for us as well as API and UI design. But from a language design perspective, when we designed the Cypher query language and then ultimately the GQL query language, which is now an ISO standard sibling to SQL, there was a. We had this tagline, which is there will be no idle variation from SQL like SQL's familiar. Don't reinvent anything. Like should be the same data type, same keywords, except in cases where you're doing something materially different, like the match clause in Cypher and gql, which is kind of the differentiated heart of the language. All right, that's time. So if any of you want to feel free to connect with me on LinkedIn or come find me and be happy to chat more. Hope this was useful.

It's. Sa. Sam. It. Sa. Ra. Ram. Sam. Sa. Ram. Ram. Sa. Sam. Sa. Sa. Sa. Sam. Sa. Sa. Please welcome to the stage Coby Conrad.

Head of growth at 1 leet.

Hey, everybody.

Whoa.

Hey, everybody. This is a little different than a zoom call. All right, guys, welcome to my presentation on five growth channels that help Rupa Health build hundreds of millions of dollars of equity. Let's get into this next slide. There we go. Cool. Hi, friends. My name is Kobe. So I currently run growth at OneLeats. One Leads is a platform for SoC2 penetration testing. But for the past three and a half years, I was running growth at a healthcare company called Rupa Health. So I helped us scale from about $20 million in seed valuation to being worth hundreds of millions of dollars. Did this with a very, very good product. But also we scaled user acquisition by about 4,000% over three and a half years before that. I went through YC 2019, raised $2 million to build a CRDT as a service. Built half a million social media followers before I was 20. And then I've been working on marketing campaigns with budgets up to about $5 million. So there's lots of cool frameworks out there for funnels and growth loops. My brain thinks in channels, so this presentation is going to be primarily centered around how we can leverage channels to drive growth. So some basic concepts, all users come from channels. I usually think there's about an 8020 to channel, although in B2B it's often very overlapping. Once you have identified your channel, you can optimize it to drive a massive amount of growth. And I'm going to give a number of examples throughout this presentation. So just for context, what exactly is RUPA Health? Rupa, you can think of it as like Amazon for lab work. So doctors log on to our platform. You can order 35 plus different labs on Rupa, we market to both practitioners and patients and to the labs themselves, just for context. So through this we're going to talk about search, we're going to talk about social, we're going to talk about what's working in educational partnerships, we're going to talk about conferences, shout out to Saster, and we're going to talk about outbound sales.

So search.

Okay, so rupa, we spend a boatload of money on search. If someone is searching for like a GI map test, we're gonna show up here on the top. So it'd be like order a GI map test online. So some key learnings in just like the last couple of years. First of all, PMAX got really, really good. I think for a long time Google really wanted you to be able to just click a button and make your ads go boom. This hadn't actually worked for a very long period of time. Now it actually works really, really well. Also some key learnings from this is there's going to be an overlap of doctors and patients that are searching for the exact same thing. I think you'll find this is a very common theme when you're working in like multi sided marketplaces where you have different users on the same thing. So we had to just deal with that is like cool, maybe 5% of these people are gonna be our ICP, but those are really, really valuable to us. So we gotta figure out how to handle that anyways. And then the most valuable search is for us were test names and individual biomarkers. So it would be really, really great if we were able to own that search traffic for free. Google is like our number one channel. It was driving a ton of users for us. But our lab catalog wasn't indexable by Google. So what we did was we expanded it into what we call our reference guide. So reference guide is over here on the right hand side, it's a list of labs. And then for every lab there is individual lab tests. And for every lab tests are collection of biomarkers. So individual collection of biomarkers. And now you might be thinking that this is a really, really great example of what you can do for programmatic SEO. So what we did here was we basically took all of the metadata that we had available in our lab catalog and then we created landing pages for every single one. So every single lag company had a landing page, every single test, every single biomarker, every single category of test. And then we enhanced it with AI as well. So we combined our programmatic SEO play with AI, AI on top of it. And then we had a small army of doctors that was going in behind the scenes that was making sure everything was factually correct, sourcing all of our sources. And yeah, this was super, super effective for us. Also on top of this, we were publishing about 150 articles a month written by medical professionals. Basically we had two approaches here, right? We had the keywords that we knew were going to be really valuable for us, so like GI Map, whatever your Google Ads are bidding on. And so we wanted to create basically, basically owned versions of those landing pages that we could start targeting for those keywords we knew were really, really important. And then the other half of our SEO play was basically long tail. So how do we generate a massive amount of content at scale, probably 95% of people that are trying to become better at SEO, you're probably publishing like maybe 10 articles a month. SEO is really simple, but it is hard for the same time because all that's required is for you to publish a boatload of content and for every single piece to be better than everything else that exists on the Internet for whatever it is you're trying to create. So the way that we did this was we basically brought in a small army of medical doctors. We empowered them with both AI and with medical editors to go over and source all their stuff. And then that allowed us to create really, really Great content at scale. That was better than what either a human could do or an AI could do by itself. Yeah, and this is our search traffic publicly available on Ahrefs. We went from basically nothing to over 300,000 organic views a month. This is actually closer to 500,000 now. Last couple of months has been crazy, but yeah, this drives a really, really large chunk of the acquisitions for Rupert right now. Next slide. So action items to take away for the crowd. So 95% of people listening to this probably aren't publishing enough content. SEO is hard because you need to create something better than what exists in the world, but then to also do that at scale. And then I would combine targeted landing pages with a strict ray and pray approach where you're going after the things you know are really, really important, but you're also trying to capture that land long tail. And then, yeah, pseo, like programmatic SEO has been around for a long time. There's a really cool opportunity right now to layer in AI on top of your programmatic SEO. So you take all the metadata that you have that's relevant to your product and then you layer in AI on top of that and you can create really, really great content at scale using that. So social, what is working with social right now? So at rupu, we built our social primarily for social proof. I think this is a super underrated piece of growth. Especially in the early days. If you're building a product, nobody knows who you are. You want to, like, drive direct acquisition. But for us, if you're in any industry where trust is important, having a certain level of social proof is just insanely valuable. So we did it. It was really important for us to, to build trust with early partners and with early users. So just going to go over really fast. What we're seeing is working organically in case this is helpful for anybody. So, first of all, on YouTube, what we did was, I want to say 80% of YouTube success right now is content based. You can kind of think of YouTube as collections of clusters of content, and for every one of those clusters, it's going to have different levels of competition. So what you want to try to do is you want to try to find clusters of topics that are trending and blowing up. For Rupa, it was really, really tricky because I'm not a doctor, so things like nitric oxide was really hot. Things like H. Pylori was really hot. Things like, like GI Map. Basically what I had to do is I had to hire doctors that was able to tell me what was Cool. So if you're in like a medical industry, you want to figure out what is the content that is trending. And it's cool that people have not, like marketers like myself haven't figured out to start creating content for. So we did that and then we basically worked with a bunch of doctors, found who was good on camera, dressed them up, gave them good cameras, taught them how to be influencers, and then that performed really well. But there's all of this. You should have a really good hook and you should have really good titles and like, you need to have really strong retention. But I swear to God, if you just have like really good topics and you create the best content available for that topic, your YouTube will perform really well on Twitter. Best thing we ever did at RUPA was we convinced our CEO to tweet. We were on a car ride and I was like, hey, Tara, she's super competitive. I was like, hey, I'll race you to 20,000 followers. We'll see who can get there first. If you can just. If you can convince your CEO to tweet, do whatever is required to get them to do this, I swear to God, it's one of the most valuable things your companies will ever do. People don't really follow brand pages even. Like, the trick to a brand page is you try to personify it, right? So you'd be like, hi, I'm the intern behind the brand page and I'm doing this like crazy thing. It's so much better if you just get like the CEO or someone that is an actual personality to start engaging with your audience. It's wild. So get your CO2 tweets, get them a ghostwriter, like, teach them how to make long form stuff. There's no substitutes of the actual authenticity of having a human actually push out content that's a leader within your organization. And then on Facebook, if anyone here is still on Facebook for us, for whatever reason, doctors are engaged very, very heavily on Facebook still, which was great for us because no one else was bidding on, on this traffic. So it was really, really cheap for us. Just fun stuff that works for us. Don't post links in like as a link or as like at the top in the body. If you post links in the comment section, that actually works way, way better. It doesn't get you penalized on the algorithm right now. And then you'll just get higher engagement. Your post will do better. We saw like a hundred x increase when we switched to putting links in the comment section. And yeah, SEO is cool. Right now, but I sometimes wear a tinfoil hat and I'm really worried about what's going to happen in the world of AI. I still think that search has a really big future within the universe, but I think that for the first time ever, search is starting to lose a little bit of market share. So for us at rupa, we're basically diversifying our bets and trying to figure out what is going to be really resistant long term term in the face of AI content. So we're going super, super hard on YouTube right now. The thought behind this was that video was going to be replaced by AI last, but this actually isn't quite as true anymore. There's some really good AI videos coming out. So I, I think the only, the only real answer here is bring in experts who can create new knowledge within the universe. If you're just relying on AI, you got to figure out it raised the fund floor. But you've got to figure out still, no matter whether it's video or articles or whatever, how to make it better than just like the base AI. So action items. If you focus on the topic for your YouTube videos, your YouTube videos will likely perform a lot better. Headlines, thumbnails, engagements, all that's like really important, but topic is absolutely critical. Dish the brand account, create followings that are people, founders, influencers, humans. All of that performs way, way better than just a generic corporate brand account. And then if you're still on Facebook, drop links in the comment section. Okay? And I'm gonna go through this really fast cause we got 13 minutes left. So partnerships. I think this is one of the most underrated channels that most people are probably missing is people will gravitate towards the really obvious stuff with like search and Facebook and social and whatever. But I've consistently seen, seen startups and organizations undervalue the impact of partnerships. So for rupa, doctors learn about lab tests not just in medical school, but in universities and programs after they graduate from medical school. So what we did was we went to all these universities and we started sponsoring them and we'll be like, hey, we'll give you a bunch of money if you will put us into your curriculum. So we did this, ooh, laser pointer. We did this with the Functional Medicine or Kalisch Institute. We did this with Functional Medicine Fast Track. We did this with Pastere, we did this with Interactive Functional Nutrition Academy. We basically went to all of these schools and we got ourselves inserted in them. And I was just, you know, sleeping one day and looking myself in the mirror thinking, holy cow, Kobe, you're spending so much money here. Wouldn't it be great if we just owned a university like this? So what we did was, was we built our own university. So we took this channel that we're spending paid money on and we were figuring out how can we create our own owned version of that channel. So we built Rupee University. It started with us just doing a weekly live call where we brought in a doctor to come do a talk about something. And then we did that every day for three years. And then eventually we're like, this is really great, but wouldn't it be better if they were actually going through some kind of program? So we started introducing paid boot camps. You can kind of think of this like Lambda School but for doctors. So it would cost $400. We would bring in an industry expert, shout out to Dr. Carrie Jones, she's like one of the world leading experts in women's hormones. And then you would get a lab test as part of this program. So this is beautiful for us because like Google Ads costs money. Right? It's great to scale Google Ads and make a positive ROI organic search even better because you don't have to pay for that traffic anymore. But this is profitable. This makes like millions of dollars a year with like good margin. So this is a channel of acquisition that is profitable, that drives users towards our core product. So users would pay us money to learn how to pay us more money is like I love it. So the moral of the story is to do don't be afraid of service business. When you're thinking of product led growth, it's not just like growth hacking across your like AAER funnel. Think through what are the things you can build on top of your core product that are going to drive more users to your core product. And then I think through where are you spending money and then how can you create like organic versions of those things? So you're probably missing key partnerships beyond like major influencers or like the really obvious ones. But think through schools, think through licensing organizations, think through different like media outlets, any, any kind of like affiliation like within your industry are usually a really great place for where you can create partnerships that will drive growth and then also give you a really good brainstorming session on what you should be building yourself so you own. Yeah, these can be wildly effective in very profitable acquisition strategies. Look at your product usage to determine good places for where you should build. Rupert University, it came from users were coming to us, they're ordering labs and the number one thing they were asking us for is like, I really want to order these labs but I don't know how. So product usage is a really great place to think about where to create things on top of it. And yeah, okay, next one is conferences. So Rupa, we go to about 20 conferences a year. It's honestly kind of wild. Here is our team. We dress up in shiny astronaut outfits. It's a lot of fun. Key learnings from conferences is number one. The most important role in my opinion of conferences is standing out. Be visually disruptive. So I'm gonna go back a slide. Imagine stuffy medical conferences, right? Everyone is wearing a suit. Everyone like looks exactly the same. Like even here, if I like squint my eyes, like everything kind of starts to look very similar. The number one thing you want to do is figure out like no matter what you can, what you can do, you have to stand out. If you're just like blending into the conference hall, you're probably going to just be lighting your ad dollars on fire. Number two, it's not just about showing up, it is about how you show up. If you show up as the headliner, that is amazing. If you show up in the back of the conference hall and like nobody can find you, that is also an impression. So I would highly encourage, think about not just, not just showing up to the conference. There's this tendency for people to show up, do a booth, stick their toe in the water. There's not always good. Often I found that the conferences where we like actually show up really, really strong are often even higher ROI than the ones when we're just in the back. And then number three, understand where your ROI is going to show up from this conference. So with Rupa, we had doctors, we had patients, and we had laboratory partners, right? So the doctors really are our primary person we're trying to acquire. But our laboratory partner also showed up to these conferences. And if a laboratory is spending or doing like tens of millions of dollars with us and we're able to convince that laboratory to pay us 1% more money, the doctors are great. But also there are other partners at the conferences. There's potential like go to market partners. If you're in multi sided market, there might be more than just your initial user that is there. And then number four, the most profitable perks of conferences aren't on the menu. I don't know if anyone here is like a sponsor at this thing, but almost 100% of the time I've seen like the things that will actually drive the most roi, they don't want to give you. It's, it's stuff like I've, I've had. If, if I can just get them to send out an email to me for me, like a dedicated email to everyone at the event being like, hey, this is our number one sponsor. This is what, what they're all about. I've had events where that was more profitable than like the hundred thousand dollar sponsorship of actually showing up to the thing, getting like on stage and then like having like a speaker from your company on stage and then able to go back to the booth and like hang out with people afterwards. Like that's usually something that's not on the menu. Do whatever you can to figure out what are the things that you can ask from this organization that they might not be telling you about because they're not giving the, those things to other people. Which by itself goes back to rule number one is it helps you stand out. And then number five is give them swag that needs to be carried and brought home. There is lots of great swag at this conference and I can't see any of it right now. You want to give them something that like is actually kind of bulky, that you can't just stick in a, like a backpack and like hide away. You want to give them something that is going to like stand out that you can like see across from the conference.

All.

So action items on this. Number one, if you can get a dedicated email that can often work even better than actually showing up to the events. Number two, the best perks usually aren't on the media kit. They're speaking opportunities, they're dedicated email sends. They are, they're flirting with someone and getting them to share you like the contact list. It's like what can you do to like actually get all this. Things that are going to maximize this event for you. Number three, it's not just about showing up, it is about how you show up. Bigger ticket impressions can often be higher roi because each impression is strong. The people that show up as like a headliner to this event are will usually I've seen that be actually higher ROI per user than just having a single booth bonus. I absolutely love lanyard speakers sponsorships. I'm looking at the crowd right now. The only thing I can see like from a sponsoring point of view is the lanyard around everybody's neck. If you're like a new person in, in your industry showing up and like being around everyone's neck like, oh my God, who are you? And be like, oh, hi, my name's Kobe I'm with Rupa. You're wearing me right now. It is like insane credibility. It's like one of the highest ROI perks I've seen. Okay, influencers. So we're gonna talk about influencer marketing for a second. In medicine and most industries where trust is important, it's not just about delivering a message, it's about whose voice it comes from. So with us, we spent a lot of money sponsoring Dr. Mark Hyman. Dr. Hyman, for context, he's the founder of our industry. So Hyman is founder functional Medicine. He has a really, really big podcast. We spend a boatload of money sponsoring him. Each episode costs us about $5,000. And yeah, so again, this is a recurring theme. Spending a lot of money with Hyman and looking at myself in the mirror and be like, damn, what if I took a couple hundred thousand dollars and I built my own podcast? I like spending money with a positive roi. I really like owning the asset even more. This is like, if you take one thing away from this presentation, let it be. This is when you're building out your channels, you're trying to figure out what drives growth for you. Paid is a really great place to start because it's immediate. You can figure out, how do I deliver impressions that will drive acquisition for me. Once you figure that out, paid doesn't scale. Like, each additional dollar is going to cost you more. Like, each additional user is going to cost you more and more money. You want to figure out how to create owned versions of that because that is how you scale paid acquisition. If you focus just on the paid version, you don't focus on the organic version, you will eventually your CAC will explode and, like, you're just gonna have issues once you get into your series B, series C. So we started building our own influencers. So we started with hiring Dr. Carrie Jones. Shout out to her again. She has like 150,000 followers on Instagram. So we trying to figure out, cool, I can't hire Hyman. Hyman is just like a God in the world. Like, he's post economic. He doesn't care about money anymore. Carrie still cares about money. So we hired Carrie, we made her our head of medical education. So we now had like a public facing brand. And then we started the Root Cause Medicine podcast, which is Now a top 10 medical podcast. And yeah, this was super effective for us. So with influencers, attribution is really, really, really hard to track. Honestly, the only way we really know that Hyman was driving users for us is we started really early and because we started so early, we were just able to see the bump on the on like the total signups month over month. If you actually look at who, how many people are saying they came from hymen, we're only able to attract about 50% of it. So just keep that in the back of your mind as you scaling influencers, especially on podcasts, is that it's really, really hard to track. You kind of have to just guess and do do some research and then you're basically going back to the 80s again and doing the whole, I know I'm wasting half of my budget, but I just don't know which half. But as long as your overall campaigns are profitable, you're probably going a good direction. But yeah, hymen was our number one source of acquisition for a long, long time. It's super important within the mess industry. Doctors learn from other doctors, patients will learn from influencer doctors. It is often more important than the medical studies themselves. I would like to say we live in a society where everything is like science based. The truth of the matter is that we just aren't so action items. There's likely a huge 80, 20 in your industry by simply going to the biggest influencers in your space. Hyman was about 80% of our influencer spend. Everything else from that kind of just like fell off. If you're dropping major money on influencers, be sure to build your own version as well. And then, yeah, attribution isn't real. You kind of have to go old school and trust your gut. Ask people, how did you hear about us? Create discount codes, but just know that you're only tracking about 50% of your ROI, most likely. Okay, very last section with one minute left. Outbound sales. So to do the data that you give your outbound reps is super, super important. If you give them a bunch of shitty leads, they're going to spend all their time trying to track down the stuff that isn't qualified. So what we did for Rupa was we scraped Google using like a Python scraper on upwork and we basically scraped everyone that said that they were in functional medicine on Google search like local businesses. We got their address, we got their website, we got their phone number, and then we layered in Hunter IO plus Phantom Buster for the emails we create. We clean them with zero bounce and then we enhance the data with OpenAI's API. So we scrape their website content to determine do they say, they order lab tests on their website, what type of medicine do they practice. And we fed all that into Our.

Sales reps.

For context, that doubled the ROI of our outbound sales team is when. When we layered that in. Okay. And we are actually out of time, guys. I'm sorry, but up on sales, take a photo. Biggest TLDR here is be really, really specific. Everyone that sends you these generic emails of like, hey, I would like to sell you something like talk to me about doing legion for a Sock 2 report and how you did this for Vance and now you can do it 10 times better for me with one leads. I'll answer that call if you just do. I'm a LinkedIn lead generator. I'm not. I'm not gonna pay up the phone. Okay, Action items on this. Is email so important? Make sure you're doing cold call. Separate your cold calling function from inbound sales function. Outbound shouldn't live in isolation. Make sure you layer in marketing. Plus blend the budget because your outbound efforts aren't only your outbound efforts and closing takeaways.

Boom.

And that is it.

Thank you everybody.

It. Please welcome to the stage Santhaya Hegday, general partner at Unusual vc.

Hey everyone, welcome to Sasta Day 3. Really excited to be here. I'm Sandhya Hegde. I'm going to be talking about pricing AI products, which has been a really favorite topic of mine all of this this year. And before we dive in, maybe a little bit of context about myself. I've been in the startup ecosystem now for just over 15 years, been a founder, worked in great venture funds like Sequoia, Khosla Ventures. I joined a very early stage startup at the time called Amplitude, which is now a public analytics company, which is where I really first dig into pricing and packaging. I played lots of different roles in the product and go to market teams. At Amplitude, we had a very complex go to market motion which included SMB, mid market, enterprise, free self serve, top down. We did all of it. So it was a real lesson and have taken that with me now to my work at Unusual Ventures where I'm investing in enterprise AI companies and really helping them figure out the right way to take their products to market. Thinking about pricing and packaging from first principles in a very complex, fast changing environment. So let's dig into that a little bit. So first of all, you know, where are we when it comes to AI right now? I feel like monetization has really become the number one topic for a multitude of reasons. Right. 18 months ago, when the AI hype cycle was really starting to accelerate, accelerate, the question was really, you know, and the question analysts Were asking was, you know, how much money are we going to make? How wonderful is this going to be? And that's really no longer the tenor of the conversation. Now, you know, people are thinking about companies that actually sell LLM models asking, you know, how are they going to make any money at all? Their margins are so thin. So companies like Anthropic, for example, are having to like, justify their margin structure. The incumbents that have shipped AI products like Salesforce are still trying to figure out how to price and package AI. And we are going to dig into, like the underlying reasons of that conflict and why they are struggling. And really, so we're left with kind of this big question, which is, okay, all of these GPUs that industry has bought and investors have financed, what's going to be the way we monetize this, right? Like, how are we going to fill this $600 billion and growing gap? And I think if there is a silver lining here, the silver lining is this, right? So for SaaS founders, this is all still good news, right? The effective cost of like LLM inference just for the GPT product from OpenAI has like decreased by hundred X in two years. Even this kind of the cutting edge model, you can see is just continuing to decrease in price. And that means that if you think about the opportunity builders have, this is not a great chart for investors necessarily, but for builders, it's a great chart because it says that you can think about the opportunity around how to build AI products and how to take them to market without necessarily having to worry too much about the marginal cost of inference for the LLMs that we have today, which is huge. This is not clear, this is not obvious 18 months ago, it is now. And I think that's actually going to be a huge advantage for everyone who is actually building SaaS products on top of LLMs. However, the fact remains, pricing AI products is incredibly hard, right? And so confluence of three things. One, yes, the dollar per token cost is decreasing. Great. But that means your customers expect it to be free, by the way, right? Your customers are looking at that and being like, hey, what are you talking about AI being expensive? It costs like $5 to record everything a human being says for a year. Why are you charging me so much? And the reality is that if you're doing any fine tuning, you're curating data to train your own models, it's still really expensive. So the upfront cost of making these LLMs really work for you is still expensive, still kind of unknown exactly how much you need to invest up front, right? So cost still kind of dynamic and unknown. But customer experience expectations are now that this should be cheap. You can't be charging me too much for it. Issue number two, very volatile usage, right? So suppose your notion you have launched AI features and your typical customer who has bought it is using those features like twice a week. You can't assume their usage will remain at twice a week. Maybe they start using it every single day, multiple times a day. The adoption curve is still early and unknown, so you cannot make big assumptions about what the actual usage profile will be. Higher usage is great for the company because it increases stickiness, it increases the perceived value of the product, but not great for your margin profile necessarily. And then last but not the least, the ROI is still unproven, which means a lot of the enterprise sales processes. The conversation around how to price and what the impact on the company's business of adopting these tools will be is a very stressful conversation to have. Because your customers don't know what to take to their cfo. They can't say, you know what, because of this, this will be the impact in the next six months, 12 months, three years. It's still unknown, it's still a moving target. So you know, all of this makes it very difficult, difficult to really pick something and have a lot of confidence that it's the right model, it's the right number. And the, I would say the impact of all this is that in general, like AI has not really driven any kind of revolution in pricing. Right? 18 months ago, analysts were like, well every company will have these AI workers and AI workers will like price based on like success. And all of, all of SaaS pricing will be success based on, you know, outcomes. Like literally zero of that has happened because all of these underlying dynamics make it so complicated. So there's, you know, very little success based pricing happening. No, nothing personalized, nothing dynamic. It's all still being figured out. And the reality is taking a first principles approach is what's required in the moment, right? And first principles, SAS CEOs really always have to answer three questions when it comes to pricing and packaging. So question one, what's going to be your model? Right? Is it seed based licenses, consumption or usage? Is it, you know, services bundling? So dig into that. How do you stop? Where do you start there? How do you pick that? Like, you know, for company like Salesforce is struggling to answer that question. How do you make sure you are making the right choice? What's the actual price point? Are we optimizing for the highest possible ASP we can charge, are we optimizing to capture the most market share we can capture in this moment? Right. Extremely important question. You can see kind of what that dynamic is leading to in terms of pricing from say OpenAI and Gemini, et cetera. And then if you already had an existing product and you're layering on AI, is it an add on? Is it included in the core product? Again, like incredibly important question with a lot of impact on how your customers actually end up adopting and using the product. And I think one thing I want to really call out is the reason I said these are the questions SaaS CEOs need to answer is pricing and packaging is a full company strategy. If I look at the pricing and packaging page of a company, I can describe to you what the company strategy is. I can also describe to you where they have not made clear choices and are trying to compromise, which makes their pricing and packaging look muddled and confusing to their customers. So you can't hire a pricing expert to answer these questions for you. You can't hire a head of pricing and delegate this job. This is the a CEO level job. And if you have been hired to, you know, tackle pricing and packaging at a company or you are, you know, part of a product team and you have been given this task, you have to work with the CEO, you have to work with council across the company leadership because the decisions you make around pricing and packaging drives all three of these things. It drives your LTV cac obviously, right, because it defines how you are, what kind of customer segments you're going after and how much you can charge them. So that defines your go to market strategy. That defines your company's ltv. Gross margin is very much a function of how strong your engineering is, how low can you go but still preserve that gross margin. And last but not the least, where competitive win rate is also driven by how your pricing and packaging and how it compares and contrast trusts to other similar offerings and companies in the market. So the decision you make on this like one page affects all three of these things and hence the full company strategy. You cannot delegate this. Right? This is the strategy. So an important thing to remember is good pricing and packaging drives enterprise value, not asp. Some of the most successful software companies in the world have grown their enterprise value by decreasing prices and not increasing them. So pricing and packaging strategy is not just about like figuring out how can what's the biggest number I can put on a slide and charge. It's, that's not it's a very short term view to take to this strategy. So keep this in mind as you kind of go through, you know, the first principles approach to pricing packaging. Now this slide is just a universal kind of questions to answer that has nothing to do even with AI, right? These are the questions you have to answer for SaaS software to figure out like what the right pricing packaging is for you. And I divide it into two buckets. One is how does your customer perceive the ROI of adopting your solution? And two, what's your company strategy? So the combination of these two needs to kind of really align into an answer so you know quickly kind of going through these customer roi. Can you directly attribute revenue cost impact to the usage of your product? The number of pitches I'm in where someone will be like, well if a sales team uses my product to close one extra deal, it pays for itself and maybe we could like charge like a percentage of the deal deal. Like even Salesforce doesn't charge based on deals, right? Like remember that like you're working in a ecosystem where your customer needs to be convinced about your pricing model. You can't argue your way out of it. They need to like what they are seeing. Which means, you know, the question is not whether the company can directly attribute impact, it's whether your customer directly attributes revenue cost impact to the usage of your product. Similarly, can your customer easily quantify improved productivity if you are taking like a licensing seed based approach? And then last but not the least, like how are they thinking about what functionality enterprise buyers really need versus not right? These three fundamental questions really are the foundation of your strategy. And then of course with the company itself. What are your strengths as a company? What are you really good at? Do you have amazing user experience so you can get very good at self serve? Do you have a really agile engineering team that can do amazing custom deployments for really big companies? These are kind of intrinsic strengths that can be converted into a great pricing and packaging strategy down the line. So let's kind of take a look at the very first question, right? So we had three questions to answer, which is what's your model? What's the price point? What's the bundling? Let's take a look at model. I would say majority of AI SaaS companies have gone hybrid. So it's maybe 60%, 65% have gone hybrid. Which means they're combining two of these three levels. And the levels have been pretty small standard, right? The first level is seat based licenses. The second lever is usage or consumption. And the third One is what I call solution. You could call recurring services package, like lots of names for it, but I think of it as like the full solution. How do you kind of price based on what's being delivered in terms of the end result? And you know, most companies in the AI industry, no matter kind of where they are, like a majority are doing some combination of these three. There are exceptions. So the product is so simple or you are just, you know, consulting service. And you can fall on kind of the extreme ends of the spectrum. But most people are going hybrid and the biggest companies in OpenAI is a great example are, you know, have multiple products. So each product has its own unique pricing and packaging strategy based on what their customers need. And they're kind of doing some combination of all three for almost every product line. So how do you make your choice, right? What are kind of the two things to figure out? And the two things to figure out are like, what's the customer value? What's the kind of level that your customer sees as value? And then how much support do they need to get started and get deployed and up and running? So let me kind of compare these two companies, right? They're both prosumer consumer companies. Character AI is an entertainment product. They are doing seed based pricing, but the value that they create for their customers is a higher quality free time. And you can't monetize that as a customer. If I love my character AI, you know, friend, you know, maybe have a really great librarian friend who recommends great books to me, I can't monetize that yet, right? If they come up with a model for it, maybe they can do usage based pricing. But for now, like they can improve the quality of my free time, but they cannot help me make more money using that free time. Which is why it's easier to just do simple seat based, fixed cost pricing. I know that whether I spend five hours or five minutes a day on this tool, I'm only going to be spending, you know, $10 a month. And that matters to most customers when it comes to entertainment, right? That's why even if you are spending, you know, I don't know, hours a day on Netflix, you're still only really paying, you know, 15 bucks a month for it, right? So that's the reason why most entertainment, entertainment is kind of fixed cost. Now compare that to mid journey where a lot of the end users are contractors and they can potentially sell more clients because they are able to use midjourney, right? They are doing book covers for Kindle Books. And now they can do 10 times the number of books that they were doing yesterday. I'm kind of making this up, but you get the idea, right? They can actually convert the time that they are getting. And all of AI is about time. It's, you know, it's at its fundamental core it's automation which saves time. They can convert that time into more revenue and hence actually have a very wide range of dollars that they are choosing to spend on mid journey in like month A versus month B. Right. It could be very different. So we'll dig into that a little bit more. But that's kind of comes down to like what is the value you're creating for customers when you free up their time, can they do something with that free time that you can charge for? Then there's support. Right. So these are both legal AI companies with very, very different pricing models. Even though at the heart they are using this very, very simple, similar technology and adding similar levels of value. But the way they are going to market is very different based on who their customer is and how deep into the workflow they are integrating. Case text is completely self serve users learn, apply it on its own. Whereas Harvey's very consulting first, probably doing the service bundles needed to actually make the enterprise they are working with like actually succeed and adopt a tool. In fact, companies like Cohera are going a step further and partnering with consulting services that their target customers are already working with, have deep relationships with, like Accenture. I think the consulting businesses did like 2 or $3 billion in new AI service revenue last year year. So they are definitely contributing to the, you know, $600 billion question. And there's a big reason why, which is that their clients cannot really afford for like a big AI project to fail. It has to succeed. So they will pay whatever is needed to get people in to make the project successful. And really that's kind of how I think about these three boxes. Right. What is it that your customer needs certainty around? If your customer needs certainty around cost, which is, you know, I can only afford to spend 100 bucks a month on this. I cannot take the risk that I'll, you know, just spend more time on it and you'll bill me more. If that is your customer profile, you have to go with fixed cost seat based pricing. If your customer is looking for ROI certainty, which is, you know, for example data infrastructure or you know, anything that you are using, API calls for even customer support ticket resolutions. If your customer is happy to spend more as their infrastructure scales, they're looking for a Fixed ROI model, not a fixed cost model which lends itself really well to consumption and hybrid pricing. But if what they're looking for is just project certainty, like I will get fired if this project isn't rolled out successfully. They need services, that is what is most important to them. And you need to have a recurring service bundle associated with your offering. And this is very common in extremely big enterprises where there is so much stakeholder management that if you just buy a new software solution it could easily be ignored by half the stakeholders required to make the deployment successful. And now you're stuck with a vendor that hasn't successfully deployed and no one is using. This is where services come in. This is why they want to buy the services, because then you are accountable for the deployment. I think Rytr is another great example of this. RYTR is a full stack generative AI company and they started out with this kind of, you know, enterprise only package, then launched a seed based small team package, now have usage based, you know, packages for developers with the companies that they are selling to. This is a great example of I would say what's kind of working at scale in AI today because the customer needs are so varied that you have to be very, very accommodating with your pricing and packaging structures to make your clients successful and basically tell them they will do it whichever way you need us to do it to make this project successful. Unclear, like that's what the future will look like as things get more standardized and more reliable. But this is definitely what's working today. Next comes price point. I'm going to share an example. So I'm on the the board of this company, Vizcom, amazing company and I wanna share an example of how different price points can be with the exact same underlying technology. So you're all familiar with Mid Journey, which is why I'm using them as a comparison. Vizcom also is based on the stable diffusion stack, but they are kind of figma for physical product design and they're an end to end workflow, starting from sketching to everything to like generating a 3D model version of a product. But they're using stable diffusion for everything. So the same underlying technology, but a full end to end workflow dedicated to one particular type of designer. Whereas Midjourney is a horizontal platform that does one thing really well for everybody. So they have used the same technology to build very different products, have very different target customers, and completely opposite pricing and packaging models. So if you look at Mid Journey on the surface of it, it looks like seed based pricing Right. It's like, you know, $10 per month per user that seed based pricing. However, if you look more closely, you will understand this is actually usage based pricing because the only difference between the four packages is the number of GPU hours you get and you get some level of discount based on how much of it you're buying up front. So this is actually completely usage based pricing set up to look like seed based pricing for the sake of simplicity and clarity. But the underlying logic is simple. You get roughly four hours per hour per gpu. You get a higher upfront discount if you buy bigger bundles. And you know, you also have access to, to a free version, but it's slow and public and I think this might be a little out of date. They have made some recent changes, but the underlying logic holds. Right. So completely usage based actually, and we talked about why. Because their main customer is ideally a freelancer who can decide whether they want to spend more or less based on how they are monetizing their own usage of midjourney. Vizcom is very different. So Vizcom works different with enterprises where, you know, a head of design who is buying Viscom cannot allow for there to be a gap between the budget that they ask the CFO to approve and what they actually spent. That could be a fireable offense at some company if you as a functional leader say, you know, I need $100,000 budget for this new design investment and then go back six months later and say, oops, we actually spent $150,000. You're just not allowed to do that. And what that means is that Wizcom knew from the beginning that they needed to enter this market with a fixed seed based pricing approach. And so they have a number that's kind of on the higher side to accommodate all you can eat usage for small teams. And for enterprises, they actually have a much higher price point, like maybe you know, 10 times what you see on this slide for enterprises and service bundles that include custom models, maybe you know, virtual private cloud deployments. Right. So we talked about like how for some, some companies that is, you know, very important. If you stack rank all their priorities, that's actually priority number one. And so you need to have a recurring service bundle arrangement for a company like that. And so the ASPs that these two companies actually sell the same technology at could not be more different. That's like two orders of magnitude difference in ASP because their target customers are different. And then we come to packaging. I think this has probably honestly been the simplest and this is where we have the most data. So if you look at whether you should choose like an add on package or include it in the core package, it's just a question of adoption. The majority of your existing user base wants the AI feature, is adopting the AI feature. You include it in your core package. I think there's some optimization math people have done. There's a great post on Lenny's newsletter about this recently. It's like if you know over 60, 70% of your existing users want your AI features, you just make it a part of your core packaging and hopefully they find it valuable. So you can actually increase the price of the core package, right? That's the dream. However, if only 20% of your users are interested in your new AI launch, you make it an add on. You know, this is for example, this is a table from Lenny newsletter. I have a link for it in the text below. You should check that out. But yeah, essentially you can see this example where you know, notion has just made it a little add on and I think that says something about the level of adoption versus someone like Intercom who's made it a part of the core package. So I'll, you know, essentially wrap up at this point. Important things to remember are as follows, right, which is pricing package packaging is full company strategy. Don't think about it as asp. We need to figure out how to what price point to sell it at. Think about it as like how do we price and package to make this the biggest, best possible company. Most people are using hybrid. So don't shy away from it and you know, think about what your customers want predictability around. Do they need predictable cost? Do they need predictable success? And last but not the least, like make sure that it's aligned to your product workflows and what your customers want to buy and use together. So I don't think we have. We are doing Q and A. So I will wrap up here but I'm Sandhya @Unusual VC Sandhya on Twitter. Happy to chat more about pricing and packaging in AI. And thank you so much.

Ladies and gentlemen, please take your seats.

Our show is about to begin.

Sam, Please welcome to the stage Adam Gross, board member and former CEO.

Advance Vimeo.

Thank you. Thank you everybody. Good afternoon. Welcome to my session on plg. Thanks for spending a couple of your saster minutes with me. I hope this is a useful and educational session. Okay, a little bit about me. Why am I here? Why am I talking about PLG with all of you? I'll give you a little bit my background so I Was early at Salesforce did the platform developer marketing bits all the parts of the platform that you like to use. I will claim some authorship of all the parts of Salesforce that you dislike or you know, not a big fan of. I had nothing to do with from there. I was the 11th employee at Dropbox. I was the first business hire from there started another company and went on that was acquired by Heroku. Went on to to run Heroku for a couple years and most recently just finished a gig as the interim CEO at Vimeo where I am now still on the board. There is a through line through all of this, even the Salesforce bits. And the through line is how are go to market models changing and how did the change in those go to market models get implemented in organizations and how we kind of think about all the bits and parts like the pricing that, that we just heard about and of course in particular plg. The thing about plg and I guess I should start by saying if you guys watch Silicon Valley, you remember the Conjoined Triangles of Success. That's going to be one of these talks, right? This is going to be a kind of a framework thing, but I hope it's still tactically useful for all of you. The thing about PLG is has a lot of different definitions, a lot of different different ways of thinking about it, but it really owes its roots to developer marketing. Right in the beginning, the first people we plged to were developers because they had both the kind of technical ability and organizational autonomy to adopt tools, new technology tools inside of their organization. And so if you want to understand where PLG go to market kind of started and kind of what's working, being in developer marketing is always a good place to be. And that's how I ended up in this because I started off in the developer marketing business. Okay, so hopefully today want to talk a little bit about kind of take a step back, maybe offer a little bit from kind of the leadership perspective of plg. What should you all be thinking about in order to make this work? One of the themes I've heard of SAS from Saster this year is in the world it's not getting any easier, right? Executing, achieving our goals, being successful, it's kind of only getting harder. The bad news I have for you, I guess first with PLG is it's hard. It requires a level of organizational alignment, executional proficiency, thoughtfulness that is hard for the best organizations and as all of you know, can be mighty struggling for startups and organizations that are just building. So understanding that it's gonna take a little bit of effort. Hopefully some of these ideas can help you think about how to be more successful and have a little bit more positive experience on the journey. So why does everybody talk about plg? This part you already know. I probably stole most of the stuff from Sastr anyway, that if you do it right, you can have much more satisfying customer journeys, much lower customer acquisition costs. Certainly at Dropbox, if you look at companies as a percentage of revenue that they spend, I should have this on the slide. You can actually look up companies, public companies, by a percentage of revenue on advertising. And if you look at your PLG companies, it shows, right? You can have more efficient go to market motions that are fundamentally, I think, more satisfying and more sticky than some traditional models. If you do plg, right, it has some nice factors in terms of predictability and a reason why it's attractive to a lot of startups and a lot of startup founders is it's an organization and operations problem, right? It isn't a relationship problem. So it's appealing to people who like building stuff as opposed to people who like spending time at dinner talking about deals with, with people. So in order to kind of get a perspective on plg, I think it's useful to really kind of reframe it because I don't broadly think about plg. I think about Go to market and I think about kind of the eras of go to market. And really for me, what PLG is, is go to market 3.0. So if it's 3.0, what's 1 and 2? So those of you who are old enough to call yourself a Gen Xer, remember that in the beginning in go to market 1.0, the way that we sold software was on a golf course to a CIO wearing Gucci shoes, right? Usually by a rep who worked at a organization across the street from us. That's kind of the typical kind of canonical. This is what it means to do enterprise software sales that a lot of us have in our head. And a lot. The myth of this model, right, still lives large. And it's a reason why so many startup founders are so allergic to sales. Because it's so unsatisfying, right? The person making that decision, the cio, huge asymmetry between the buyer and the user of the software. So the user doesn't have any power, they are just getting, you know, whatever the buyer happens to choose gets forced on. And, right, this kind of very top down dissemination of technology and what ends up suffering, stuff the users care about. Namely experience. So rightfully that has a bad rap. And then in around 2000, there was a new model that came up. And I think it's helpful to remember how new and different go to market 2.0 is. And go to market 2.0 is fundamentally what we all practice here, right? I was there at Salesforce when a lot of this stuff was invented. The idea of a trial that you could sign up for online, that you could discover via SEM or SEO, right? Also new communications methods, new go to market methods, new technology delivery methods, all of these things tend to happen together. These new go to market motions don't happen in isolation. They're not just a different way of thinking about doing a sales motion. It's a fundamental different technology mix. And it's something that maybe is kind of implicit, but I don't know, gets as much kind of thoughtfulness as it probably deserves that one of the big deals about SaaS, one of the big deals about cloud, is because of there was no software, because there was nothing to install. It enabled a whole, a whole new generation, a whole new way of thinking about selling and go to market. But. But we didn't just do 2.0. When we got really good at 2.0, what did every single successful SaaS company, including Salesforce, do? It layered 1.0 back onto the top, right? And that was the journey. And that's the punchline. Because of course, we are now firmly into the 3.0 era. And what's different, right? In 2.0, you start not with the.

C level on the golf course.

You start with a department head on a Google search, right? That discovers your landing page, lead capture, maybe SDR motion, whatever it is, right? Trial, and ultimately typically an online sale. Then the 1.0 motion on top of that to get to the circle or the CIO or whatever the C level is. So with 3.0, we take that another layer down. We say we're gonna go directly to the end user, give them something that they use, not buy. So the measure becomes adoption, not conversion, not closing. And once we have that adoption, we convert to the department and then we convert to alt, ultimately the Enterprise. So punchline, plgie go to market 3.0. It does not exist in isolation. It is not an end state. If you think you are just gonna practice that, I'd say you are probably thinking about it wrong. And we could probably do a whole conference on companies that have tapped out by their inability to get beyond kind of plg. This was very much part of the heroku journey that I was part of. Maybe I'll get to speaking about. Okay, I'll put in a plug here. As I use up all my time on my one slide, we are definitely starting the GTM 4.0 era. I think that's an incredibly interesting thing. But we're all still trying to work out this 3.0 thing as well. So if you take a picture of one slide today, I would recommend this. I have talks and slides and stuff like that online, so if you don't, that's fine too. But so I've seen this model at a couple places at scale at publicly traded companies. And the thing is it in a very satisfying way fits into a pretty neat pattern. Right? So we talked about 1, 2, 3. I'm probably going to invert that a little bit here and confuse everybody with the numbers. But another way of thinking about layering on those cells motion or those kind of idea of eras is into kind of a holistic view of the customer journey, which really contextualizes what does PLG look like, what does it mean? And what it means is, right, you start with typically something that's free to the individual that you then via some kind of sass ish, maybe higher velocity inside motion, sell to a team and then enterprise via more traditional sales method. Well, okay, maybe that's not rocket science. One of the things though that's very satisfying is pretty much across all businesses you can think about what the value proposition is and how to measure these businesses in a consistent way. And I encourage you to kind of think about this framework and apply it to literally any business that's kind of practicing PLG that you know, which is to say the way you run or think about, right, this free pillar, right? What are you ultimately organizing and optimizing for? It's Mao, right? You're trying to do adoption. What are you doing in team? Typically, it's mrr. You're typically on a monthly, you know, high velocity kind of sales motion and enterprise acv. Well, the other fascinating thing, and again, I encourage you to kind of try this out as you are perusing the Internet is there tends to be a pretty consistent value proposition that's married with each one of these pillars. So when it's free, it's typically some kind of creation value proposition. Right? And I use kind of creation broadly here, but you can think about kind of single user mode of Dropbox or Heroku or Vimeo. What am I doing as an individual using those free products? If I'm Heroku, I'm creating an App. I'm. Or even like think about free mode of GitHub. What am I doing? I'm finding libraries. It's a fundamentally kind of creation activity in team mode, where typically is where monetization starts. What is that fundamentally about? It's collaboration, right? And you know, pick your favorite example, anything from Google Docs to Figma. Having that kind of multiplayer collaboration mode is usually the value proposition that kind of gets you to that next phase. And then finally at enterprise, what is it? It's compliance. So this was very much. This gets nuanced, but certainly in the Heroku journey, you know, we started it free, we had self serve. The big thing that we did, at least during my tenure, was add the enterprise, which was how do we allow you to create secure, compliant applications with the same kind of experience and ease of use that you can do core creation or development collaboration around. So it turns out PLG go to market 3.0 has a lot of these very kind of consistent shapes and characteristics, which is really helpful if you're thinking about kind of building this motion into your business. And it also kind of points to one of the core punchlines, I think, which is PLG is not a sales motion. Go to market 3.0 is not a sales motion. It is a business model. It is a business architecture. It is fundamental to your customer journey. And this of course, is why so many organizations trip over themselves in trying to execute this. Because if you think about it as magic pixie dust that you're just gonna sprinkle over your business and say, okay, we're gonna just, you know, change the price on the webpage and all of a sudden, you know, we're gonna have all this PLG goodness. Well, that's not gonna work because these are fundamentally products. These are distinct products with the same distinct value propositions, with distinct price points and organizational structures and processes behind each of them that need to not only be constructed and operate successfully in their distinct pillars, but then need to be married with each other so that you can have a holistic user journey that spans the entire thing. That's hard to do. It's hard to do because, you know, I've been doing this for a while. I'm sure many of you as well. Just doing one of these is really hard, right? Just doing. A lot of us are kind of spending time here in two, or maybe in three, maybe we've got a little bit of both of those going on and keeping our collective head straight. Making our marketing programs just work in two, right? Is just. It's a huge task. And all of a sudden what we're saying is we're gonna multiply that by three. And that is both the promise and the peril of this. You can do it, but it has to be a holistic intention that this is the architecture of your business that you want to pursue. It's not for everybody. You can do just three. You can do two, three. You can do one, one, three. Right? That's how a lot of open source businesses look. Especially historical open source businesses where you have free open source and then you have kind of enterprise layered all the way on top. And what you saw over the past couple years is as they then offered their products as a service, think about, you know, your classic kind of database companies, open source database companies here, right? They layered two back in huge amounts of work, huge organizational transformation. So there's always a mix of these models that you can employ. But just being having kind of a common language for thinking about it and an expectation of what it means to execute on this basis is really helpful. Okay, so if that's kind of the theory, what are some things that maybe I've learned in practice and hopefully can share to help you spare some paint? Should look at my slides ahead. At least I'm consistent. It's not a channel, it's a business model. I really think about this a lot. And I also say this to all of you. I'm kind of a product marketer in my heart. And for all of you who are kind of looking to become great product marketers or marketing people, I really encourage, encourage you to hone your thinking and your skills about kind of the holistic kind of architecture of the business where pricing and packaging, value proposition, messaging aligns to the customer journey or different sales motions. That is the single hardest thing for companies to get. Right. If I look at what was the big impact I had at Heroku or what was the impact I had at Vimeo, it fundamentally boils down to that. And I look at my successors who, who have done great work at both those places and that's the work that I see them engaged in. How you kind of holistically construct this offering and fundamentally think about how your business model relates to that. Now, as somebody who spends a lot of time in sometimes small, sometimes large organizations, and as I'm sure many of you experience in your kind of day to day professional lives, the single hardest thing for an organization to achieve is alignment. Right? Are we all pointed in the same direction and kind of doing the same thing? And the thing about go to market 3.0 is like it's an alignment randomizer. Right? Because I'm a marketing person and I've got, you know, sales as a customer over here and I've got customer service and I got the CEO and all of a sudden I'm trying to juggle what are my priorities, how do I think about my work arrayed across these kind of three fundamentally different modes of business. And so the core thing that organizations need to do in order to be successful at this is be good at that alignment. And what that means in practical terms is of course, right in 1.0 the way the world worked was you had a bunch of engineers who stayed up late and created a CD and then they threw it over the wall and then right, the salesperson picked it up and took it to the golf course. Now we have these incredibly intricately linked processes internally that ultimately create a holistic customer experience that goes from the website all the way to the product. How do you even think about doing that? I'll give you my one cheat code. I have done this at series A startups. I have done this at publicly traded companies. The quickest thing you can do to help get alignment is to create some kind of squad or team model around the pillars. Where typically you have kind of two in a box or representatives from different parts, functional parts of the business organized around those three pillars. So literally the way I would run my business at Heroku or at Vimeo is I had a free squad, I had a self serve squad, I had an enterprise squad. Now it's not exclusively it, but that's in terms of the cross organizational piece. They had their own functional priorities, but they also met as a team. They aligned and synchronized their priorities as a team just on the kind of project outcome, quarterly planning basis, not at like the detailed project. They still went off and did their own things. So they had some common way of fundamentally reporting on the business. So there was some even virtual or matrix function that you could say who owns this thing? Because the hardest thing in 1, 2, 3 is who is sitting atop each one of those pillars when implicitly they're highly interdependent. So as complicated as all that sounds, a simple meeting where you have key stakeholders. So for free and team, it's typically somebody from marketing, somebody from product from. For enterprise it kind of works how it does today. Typically somebody from, from marketing, somebody from sales and you can get a rational shared view of what the organization is doing to drive those businesses forward. Okay, As I mentioned, it's this Whole transition is particularly challenging for marketing, right? Because kind of marketing sits as the recipient of all of these requirements across the business. And so now marketing is kind of being pulled in all of these different directions. And how, Mark, how you think about kind of structuring marketing and how you think about the marketing activities that you need to be good at and keep yourself kind of sane across these different functions is really, really key. And I will. One thing I spent a lot of time talking to companies about, and this is true both kind of locally within marketing and kind of holistically across the business is the key thing to be aligned on with the company. Once you have those kind of structures and the ability to kind of drive the car in a direction, hopefully all on the same line is what is your prioritization across one, across two and across three. And this is like I have not, I've never been in strategic planning where we haven't struggled with this, right? If you're the product leader, if you're the marketing leader, the salesperson typically doesn't care because they just have their piece. They're typically only focused on three or two and three. But anybody who has kind of ownership across all of those pillars, right, trying to negotiate that balance and oh, we've got to honor our community, but oh, we need to grow. And the future is enterprise. And right. Enterprise always has a bigger voice. How you have a rational, effective alignment across, so the organization is consistently allocating its resources across those pillars in the same way is hard and it's super important. Here's the good news. There's only one wrong answer. The wrong answer is a third. A third, a third, right? Because if you're just peanut buttering across, you haven't made a decision the other. For those of you who are kind of, it's more rare for a company to be kind of fully up and running with all three pillars. Usually where companies are is you've got one or you've got a pillar and you're thinking about adding another one. Usually you've got two and you're thinking about adding three. And I'll give you my simple cheat code and my remaining three minutes for that. For those startups in the house, if you are, if you're doing two, if you're doing one and two and you're thinking about adding enterprise and you're under 10 million in ARR. Don't. The cost and complexity of building everything you need to do enterprise, it's so alluring, right? Because those customers are right there and your VC just introduced them to you and they're gonna sign $100,000 if you only do X, Y and Z. There's a little bit of kind of what I call the enterprise mirage that startups often fall into because they can get this initial boost of enterprise customers through inorganic or unnatural means, VC introductions, friends or family. But what it doesn't mean is they built all the repeatable systems underneath that and so they think they've done enterprise when they really haven't. So that's my resource allocation spiel and my kind of last point here. I mean, hopefully most of you know this. I stole this from Saster so you can ask Jason to make an updated version because I certainly couldn't do the legwork. I'll tell you the good news is I've been doing this for a long time, right? Salesforce in 2003, Dropbox in 2010. You know, it was very hard to have any shared consciousness about what this model looks like or find people who can bring best practices. Good news, this is becoming more common so you can hire for it and kind of build these capabilities in your organization. And that's reflected in how common it's becoming even in publicly traded companies. And it's still hard, but at least it's becoming more common. And I'll end with kind of my last bit of advice, slash thing I will implore you is in the 3.0 world, operational excellence becomes paramount, right? Because if you think about the go to market motions from one to three, one is really kind of consumer ish, right? It kind of feels a little more B2C ish. It's high, it's high transaction volume, it's very programmatic. Online automated marketing versus right, all the way to three is back on the golf course. In order to do one well and in order to connect one all the way up to three, you need to be excellent at your customer data systems. And I say this as a CRM nerd and it's really, really, really hard. Usually organizations don't prioritize that kind of operational efficiency. Have the CEO really give a shit about what your customer schema is. If you're a CEO and your CEO doesn't care about these things, I would lobby them intensely or you as a marketer, make sure that you're helping raise the fluency in the organization for that. Because it's so key, key to make this work. And guess what? It's going to be the foundation for 4.0, which of course is the new AI automated piece of this. So operational excellence and proficiency sales ops, Rev Ops, all those pieces, operational data, warehousing, your customer, data platforms, all this stuff just gains much more relevance in this new world. I am at time. There's more stuff. I've done a whole bunch of stuff like this online. If it's interesting to you, just search my name, you'll see it. I have all separate time talks about what it means explicitly to add the enterprise pieces and I'm doing an AMA in 30 minutes if anybody has questions and I hope this is useful for you. And thanks so much.

Ra. Ram. Ram. Sa. Sam. Sa. Sa. Sam. Sa. Sam. It. Ra. It. Ra.

Ram.

It. Sa. Sa. It. Sam. Sa. It's.

Foreign. Please welcome to the stage sam liang.

Co founder and CEO at otter AI.

Cool.

All right, thank you everyone for being here.

I'm Sam.

I'm co founder and CEO of Otter. How many people have used Otter? Raise your hands. Awesome. Thank you all for being our customers. Today I'm going to try talk about how to build a bottom up AI SaaS startup. Of course you know there a thousand ways to build a startup or AI SaaS startup. I'll just talk about how we built it and share Otter as a case study. So I don't claim this is the the only way or the best way. Of course. Just a quick introduction. You know we started 2016 with my co founder Yuen Fu. Both of us studied computer science, electrical engineering, engineering. We're you know, technical co founders. We try to solve a problem, to capture, transcribe, analyze conversations and solve a memory problem and the information sharing problem. We want to make conversations more valuable. So we built the first AI meeting assistant on the market. We start 2016. We have a mobile app, we have app that you can use for your Zoom meetings, Google Meet or Microsoft Team. Between 2017 and 2023, all the meeting transcription you see on Zoom was generated by our system. So we have processed all over a billion meetings and probably multiple billion meetings. So we're also building collaborative AI to help people collaborate better. So okay, today I'm going to just talk about only 25 minutes, so not a lot. I'm going to talk about three topics. Number one, how to pick your vision and market. Number two, determine your go to market strategy. Number three, how to build your AI infrastructure so that you have a technical moat. First of all what do you do? So do you pick a niche market or a, you know, a much larger broader horizontal market? So of course there's successful companies using either strategies. For us, we decide to work on a horizontal market because the potential time is so big, because everyone talks. So on average you talk 10 to 20,000 words a day. Everyone here probably have 10 meetings at least every week. Probably more than that if you are a manager. So we decide to work on that market, you know, say, hey Otter. Our goal is to put Otter in every single conversations. Put Otter in every single meeting. So if you think about it, that's just is tremendously huge market. Of course, you also want to think about any specific vertical use cases, whether you are a product manager, whether you are a sales manager, whether you are a marketing team member. So different vertical use cases. So our strategy is to build a horizontal platform first and and then build vertical solutions on top of the horizontal platform. In terms of horizontal platform, we need to have some basic functions we build like transcription, summary, and we also build a workspace with a channel system, very similar to Slack channels, to allow people to share meeting content. They can also use AI chat on the meeting contents to query the entire meeting database. Right on top of that, we actually released that in 2018 and continue to build the system on top of the initial app. So and then we observe how people use the system, you know, do they use it for what type of meetings? And we also asked the user what their roles are. Are they in the sales department? Are they in the marketing department? Are they a recruiter? So we collect a lot of data and use that to help us determine our strategy. So we actually notice a lot of salespeople use Otter because they have tons of calls with customers and after the call they need to extract data out of of the conversation and put that into CRM. So then we decided to build a vertical solution for the sales department, right, to help them extract data out of the conversation, such as, hey, the competitors customer considered the pain points of the customers, what their budget, who are the decision makers, follow the message methodology, we extract the information and automatically put that into Salesforce. So again, based on the horizontal use case, we actually discovered a lot of a vertical use case. Then we tried to get deeper into each vertical and make each vertical use work a lot better. Then we want to build a product that ourselves can use every day. As a startup founder, I talk to our internal team all the time. I join sales team meetings, I join marketing team meetings, I talk to a lot of VCs, I talk to a lot of customers. So we actually decided to build this product to partially solve our own problem. A big advantage of this is that you are your own customer. So you use your product every single day. You Feel the pain point. You come up with ideas to enhance the product. So if you decide to work on the product, that can solve your problem, own problem, I guess you are a lot more motivated. Your whole team can, you know, think about use the product. They actually QA the product every day. They come up with creative ideas to improve the product every day. So it doesn't feel like a work. It feels like you are building something for yourself. You're building something to make your your own life better. And of course you need to build a system that these days it's all about generative AI. It needs to be able to use generative AI natively. A lot of people try to put AI into their product. Sometimes it works, sometimes it doesn't work. Is there intrinsic reason that AI can make your product better? You know general AI is all about large language model, you know, while conversation is about talking. So it's actually very natural match to use genuine AI to build our product. We need to use gen AI larger language model to understand the conversations. What do people talk about? Right? What are the action items for salespeople? After a sales call they want to generate a follow up email immediately. So it's a very natural use of generative AI. You also want to think about why is your startup defensible. A lot of startup actually I see are just trying to call build a very thin layer by calling OpenAI or anthropic API. Sometimes it can work, but it is so easy for you to call someone else's API. It must be very easy for everybody else to do the same thing. You really need to think about what is your moat, what's different, right? Can you fine tune the model? Can you use the customer's data to actually enhance the model? Basically can you build a flywheel that if there are more users use your product, can they help improve the model to create more values for them? So the idea is that you use AI to build a product. You get a million users, the million user use it, they give you some data. The intrinsic of that data is.

Used.

In to enhance them. Of course you want to protect the user's privacy. You know, it has to be anonymous, there has to be no PI in the system. And for our system we actually ask the user to contribute. If they're, you know, for non confidential meetings there's a button and they say hey, I'm okay if you use this train your data. We have a lot of users voluntarily contribute data and of course for enterprises they are okay to use the data to enhance a Model just for their use case and only used inside that enterprise. And you can now use their data to train model for other customers, of course. So next is go to market strategy, right? There's definitely a thousand different ways for us. We want to build a product that's intrinsically viral and can also help us generate a free PR for us. So in terms of virality, if you think about the use case, when people use Otter in meetings, meeting is all about communication and collaboration. It's about talking to multiple people in that meeting, right? So if you talk to 10, 20 people in the meeting, the content is useful not just for the meeting host, is useful for everyone. So if one person use Otter, it's very natural for that person to share the meeting notes with 10, 20 other people. When they share that meeting to the other 20 people, many of them decide to sign up because they need to take a look at the summary, they need to figure out what action items they need to follow up with. So they sign up on Otter, then they decide to synchronize your calendar and use it for their subsequent meetings. So we build that natural viral loop into a product so that we can grow virally really fast and doesn't require a lot of marketing budget. The other one that's unique for Otter is that Otter was adopted really fast in the early days by reporters because reporters do a lot of interviews and they need to write a blog or article really fast right after the interview. So they found Otter super useful and when they use it and find it useful, they decide to write a lot.

Of.

Articles about Otter. So without spending marketing dollars, we got a lot of free pr. Of course it doesn't apply to everyone, but you want to really think hard. How can you build a product that can get people talk about it organically? So we got tremendous organic virus loop. The other thing is Otter has been used in a lot of conferences. In 2018, TechCrans Disrupt decided to use Otter to provide live transcription for all their speeches, all their panel discussions. So that gave us a lot of a free pr. You know the obviously that's the probably the biggest startup conference in the world and they use Otter. Then they announce Otter to all the attendees and ask everyone to download the app so that they can get access to the transcript of Otter.

So.

Yeah, pricing. When we first released Otter, we decided to price it very low. Back in 2018, the cost of transcription is extremely high. Normally you have to hire a human to transcribe your recording. The average cost is about $1 per minute. So when we Release it. We say hey, we, we are going to completely disrupt this market when we release it. We decided to provide 600 minutes for free every month. If you think about it, it's almost worth like $600 and people can get it for free. So that help us get a huge user base very fast. So, so you want to think about your pricing. Really hard to figure out a way to get a massive adoption. Because you want to do it bottom up, you need to get massive adoption as fast as possible. Of course that's actually related to my third point I'm going to talk about later is like you want to build as much in house AI as possible so that you can control the cost. Because we build our own AI technologies rather than using a third party API that enable us to provide the service at extremely low price. So if you call, if you solely rely on external API, you won't be able to afford that. And, and let's see. And we started with pro filmers, tons of organic adoption and then when people use it we actually one critical thing is at first we just let people to sign up using email, Gmail, Hotmail, anything. Then we realized that actually most people use it for work. If you use it for work, it's actually best to get people to sign up use their work email address. The reason is that all their meetings is on their work calendar. So we actually use her to synchronize a calendar with Otter. If they use their personal Gmail, they cannot use Otter automatically with their work emails with their work meetings. So that's a critical thing. If you want to build an enterprise bottom up startup, you absolutely need to incentivize the user to sign up using your work email address. So when they do use their work email address address, synchronize the work calendar, Otter get access to their other meetings and then the autopilot can join their zoom meeting, Google Meet meeting or team meetings automatically at the beginning of the meeting. They don't even need to do anything even if they're on vacation or they are double booked. Even if they can cannot join the meeting personally, their autopilot can still join the meeting on their behalf and take notes. When we do that and we look at our database or look at all the email in our system, we actually see that there are tons of people use it in the Fortune 500 companies and most of them are free users. But there are a lot of pro plan users, a lot of business plan users and that help give us a lot of information and see the user distribution among the Fortune 500 companies. And we can also try to understand because we do ask the user to tell us their role. Are you in the sales department? Are you in the marketing department department? Are you a recruiter? That could give us a lot of information and help us determine our strategy. And then our sales team can look at that database and determine, hey, which company should we focus on? Right. If we see a thousand users in a company, they will try to figure out who to contact. Do they go to the VP sales of that company? Do they go to the IT director or do they go to the recruiting team leader so that help us build a motion to penetrate enterprises. So again, this is the advantage of the massive adoption. No matter which company you or at, if we look at the database, we probably can find a lot of individual users. So this strategy is actually used by a lot of other successful companies. Dropbox is one of the best examples. Right? Dropbox is hugely deployed in a lot of companies before they build an enterprise product and then when they decide to go for the enterprise market, they go to the cio, the CTO and say, hey, you need to aggregate all these users into enterprise contracts. So that's what we're doing. So I touched on this earlier. Earlier. There's a lot of large language foundation models out there, but the biggest deficiency they have is you can ask ChatGPT about anything on the public Internet, but it doesn't have anything in your company. So for our use case, ChatGPT can answer any question about meetings in your company. So we have built a system to of course have as many people use OTTER for their meetings in their company. But then we need to build a model to allow the users to query all their meetings in their company. So we're also adding additional data into that system incrementally in the future we can allow you to speak in the meeting and say, hey Otter, hey, what's the growth rate last month? The OTTER system can query your system and answer the question for you. So it's very important to get access to enterprise data and build a model on top of that so that people can take advantage of it. Let's see, I talked about the cost earlier. That's super critical. From day one, we decided to build our own model so that we can actually provide higher accuracy at much lower cost. And also when you control your own model, it gives you the flexibility to customize it if you're only using a third party API, if that API doesn't support certain use case, there's nothing you can do. And you Cannot create your unique differentiation because again, that API is available for everyone. If you want to build something special, it doesn't support it. So this is why it's important to build your AI team as much as possible. Let's see. So yeah, that's the key points. So you want to build a unique product, you want to have a differentiation that solve a real problem, ideally solve a problem for yourself as well. And you want to build proprietary AI model based on proprietary data. That's it. I can take a few questions. And use the microphone here. Thank you for that. I used Otter before. One question I had was Sunday.

Of course everybody has that now, right.

Like Zoom, but everybody has a recording assistant, et cetera. So how do you actually compete with people who own the meeting infrastructure? Yeah, we definitely see no matter what you do today, there are always two camps of competitors you have to compete against. One camp is the big techs, Microsoft Copilot, Zoom, AI Companion, or Google or other large tech. Or there are other. There are probably at least a dozen other startups who are doing the same thing for us against the large companies. We just have to innovate faster. We have to build a better product, a better user experience against Microsoft or Zoom AI Campaign. I don't know for other people who use Otter, why don't you use Microsoft or Zoom? I think we have a better user experience. Even the feature wise. I think I still haven't seen Microsoft Copilot or AI Companion enable you to query your entire meeting database. They may build that later, but we have that for. When did we release that? I forgot. At least six months now. I still haven't seen them build that. The other major differentiation we have actually is the mobile experience. We support both online meetings and in person meetings. So when you meet people in person at Starbucks, for example, you don't use Zoom, you don't use Microsoft. So that's another advantage. Of course, when you build your own startup, you have to figure out your own differentiation.

Can you give some examples of fundamental.

Technology limitations you have run into? Like you mentioned in person meetings or multi speaker separation?

Multi speaker separation is actually a major one. When you're on Zoom, if you have three people in the same room, they actually cannot separate them. So we actually build our own technology to separate each speaker and identify the name of each speaker. That's a unique thing we did. Of course, even on ASR we are one of the few who provide a real time transcription. And very soon you're going to see real time speaker identification as well. In the Future, we want to build this voice assistant, real time voice assistant in the meeting. So not only you can transcribe the meeting using Otter, you can also ask OTTER questions in the real time meetings. So all those require your control of the model. It's not very easy to use a third party API to provide that. Yeah.

Hello.

In terms of building your own LLM, did you come to that conclusion when some use case wasn't satisfied or did.

You come with the philosophy that we.

Will have our own model? I'm asking because there are now multiple methods, right? You can train or you could fine.

Tune or you could prompt engineer.

You could use multiple models. So what is your path and what's your suggestion? I guess for others too.

Fine tuning or continuous pre training using your proprietary data can definitely help a lot. There are many different ways, but for us we see a major deficiency of the large language model on the market is that most of them are trained on written documents. There's some conversational data, but not a lot. So their model is biased toward the written document. So for us, we really want to understand the multi speaker dynamics in the meeting. So we need to do something special there to understand that. Our goal is to eventually build an agent avatar that can simulate a real person to speak up in the meeting and interact with multiple people. So we don't think the current large language module can support that. Yeah. So I guess I'm at time. Thank you ever very much.

Sam. Sa. Sam. Ladies and gentlemen, please take your seats.

Our show will begin in a few minutes. Please welcome to the stage carrie lou.

Dietrich, cmo advisor, formerly at atlassian.

Nicole baer, cmo at carta.

Gail moody bird vp marketing, LinkedIn sales solutions. And denise pearson, cmo at snowflake. Just taking my right. I'm last. You're last. Okay. All right. Hello. So nice to see all of you guys here. We had a wonderful CMO brunch earlier today and some of you have come and some of you, we get to talk to you for the first time. I have some of my favorite people on stage here today with me because it's one of the hardest times of the year for marketers. It's the time of the year where we have to wrap up an assessment of what we've done so far and decide what going to pitch to invest in next year to really make a difference. And this year is harder than any past years because the metrics haven't trended consistently right. Conversion rates are down. Performance marketing is harder than ever. Many companies are trying to reach out to new audiences or with new products or new verticals or new geographies. There's so many things that changing and yet it's the time of year where we have to propose to the board the budget we need for next year and commit to the revenue that we think we could get with that budget. And so I asked these ladies to come on stage and talk to me a little bit about how they're thinking about the changing nature of their businesses and where they think they can invest to get the greatest growth. And one of the things that stood out to me is as marketers, sometimes we get put into the box or we put ourselves into the box of talking about marketing tactics. But really these women are talking about growth tactics. They're really talking about the drivers of the overall business and then how marketing will operationalize those changes. Really how the fundamentals of the product, of the audience, of the market is changing. So Denise, let's start with you. You know, AI has really continued to change the nature of your business and yet there's so much more opportunity than ever before. How are you thinking about the growth levers for your business this year?

For us, I mean, we've been through everything from breaking into industries, I mean verticalized, the sales organization and the market organizations. We of course expanded into over 40 countries, right, in the last, you know, eight years. So we've done that part for us now our growth next is really coming from multi product adoption. So that's what we are really focused on. And of course AI is a big part of that. So it's our AI offering is of course growth opportunities for us also. We're really very focused in the machine learning space. So that's really where we're focused on and therefore we're very focused on our existing customers. We've really been so focused on new business, right. I mean, over the years. And we still of course are, we're not slowing down there. We're at the point where we need to break into our organizations, customer organizations, wider, deeper, with our new products. With that as well, we're looking at of course, how do we scale further. And also over the years we've invented so many new products and where we are now is that we have this big franchise product programs that we run now all over the world. So right now it's the world tour in spring, there's another program, in the summer is another one. I mean it's, it's evergreen program. So at the moment we, we just can't invent more. It's A focus on which are the programs that we can scale even further. What are the programs? You know, we can stick scale with our partners. It's all about scaling what we have, breaking into, into our accounts at the wider and deeper level. That's really where our growth is going to come from next.

And some of the changes you've made there have been organizational to try to align that tell us a little bit. As early stage startups, everyone's in the same office and you all get to talk to each other and then you get into these functional teams and then at some point of growth, when you're slowing down, inventing new things, you either have business units or regions. Just tell us a little bit about how you're thinking organizationally about aligning people to those goals.

And also our headquarters right here. So Matia, we're sort of in hybrid model. Everyone is in the office three days a week and sort of the philosopher on here like headquarters. Central versus decentralization. We really want to keep the central team as lean as we possibly can. Right. It's like people, it's a people themselves with it, with, with a quota and non quota carrying people at a headquarter. We have no quota on the marketing team, right. The regional teams, they have have quotas. So our role on the central team is really only to focus on those scalable programs. Of course, everything around positioning, messaging, the brand sales enablement, all those sort of things. But everything else we really, really want to push out to the regions so they have more autonomy. And as you know, as you grow, right. In the early days everyone is, you know, wearing different hats for different things, right. And then some people become leaders and some people become more specialized and so of course they're more, some of the specialized resources were more central. Right now we have more special resources out in EMEA and apj, right. They have product marketing teams now they have customer marketing, they have partner marketing team, they have brand resources which they didn't have in the before but really want for us to be able to move, continue to move incredibly fast. The regions cannot just rely on us here in Summertia, right. That would be a massive bottleneck. So now it's more about more power to the regions, let them move, move faster. We also work very closely. The MIA leader, the APJ leader is my direct reports, right. I mean we talk several times every week. So they are also, they are running some programs that are global too. So some of our global training programs is run of APJ because they're so incredibly good at scaling things sometimes some great innovations comes from Europe and they get to lead that, you know, globally. So again, as little people at the headquarter as possible, you need to have a quote on the market marketing team. But again, of course. Right. You want to be able to have those shared resources so we don't have to redo things. Right. It's all about dora. Do it once, repurpose and it work.

Yeah.

So effectiveness and efficiency. Now Carta is in a very different place right now. Right. Nicole, you guys have had this traditional focus on the cap table business, but now you're expanding new products and new audiences. So your growth levers for next year talk us through how you're thinking.

Yes.

So we're very well known as a cap table business. Hopefully many of you are Carta customers in the audience if you're a founder. And we're making this strategic transition to creating software for the office of the fund CFO. We see fund CFOs for VC and for private equity as being incredibly underserved. They're still using Excel, they're still using single points of failure with accountants. They use other vendors that are primarily service forward as opposed to software forward. And we really want to give our fund CFOs what we think all CFOs believe, which is a true ERP system. So that's really where our strategic focus is going to be over the next year. And I can say as a cmo, it feels so great to have a singular focus. It really does. We're really laddering everything we're doing in.

Marketing, in product, in sales, into this.

Singular vision around how we're going to really fall in love with the fund CFO and deliver something that, that they've never really had before in their career and create incredible value, which is what we think we've done for founders as well in the cap table business. And so your marketing strategy is really about a lot of new new campaigns, new new categories work, new audiences, all the new new.

Yep.

Yeah, we've been, we, we've had fund administration as part of our product portfolio for some period of time and things like waterfall analyses for private equity. But it's really about coalescing the entire offering for the fund CFO that we can then align our marketing organization around. So I'm hiring for product marketers. This is a small plug. Product marketers and customer marketing managers are really important to this overall vision. Really anything that's going to be at the heart of this customer we're focused on, which is the fund CFO as well as deal teams, finance teams and accounting teams. Great. And Gail, so Gayle is the vice president of marketing for LinkedIn sales solutions. So she's focused on the aspect of how Companies can use LinkedIn to sell better. So she's at this interview intersection from my perspective of what are the marketing strategies that are shaping this coming year and what are the sales strategies that are shaping this coming year. And of course the pitch for LinkedIn is we help you know better and target the exact right people and interact with them in a much deeper and more engaged way. So you must have a lot of pressure to like do great marketing and great sales being kind of at the epicenter of it. But it sounds like you too are trying to can figure out how do we like re accelerate growth and which new levers can we use on the on the team or as a company to do so. Yes, I sit in an interesting place in LinkedIn because we are known as a talent solutions business. So in the sales solutions business we really are sort of a startup within the business even though we're a billion and a half dollars hitting upon two billion dollars. So the question is how do we grow, how do we continue that acceleration? We think mining the customer experience. So CX is our path to doing that. 90% of our revenue plan is working with existing customers and going deeper. One of you mentioned penetrating customers and finding more pathways, more regions, more divisions within an existing organization. So we've firmed up a priority within in our organization that brings customer success. Sales, life cycle marketing and product marketing. To understand what are those levers we need to touch to get greater utilization of the people who already have our products, making sure that they understand the value. So a lot of time spent on strategic value reviews, making sure that the onboarding process is clear. And, and so we did a study with Accenture and we looked at all the customer experience maps we had. We had about 15 customer maps. So everybody's got their own view of the elephant and thinking about how they're approaching customers. We want to unify that and make a seamless customer experience focused on value so that when it's time for that renewal or that growth conversation, it's very clear that that there's been value delivered, that people are using the tool and that it's building growth in their organizations. So for us the mantra for the year is really CX investing in that and really removing all the barriers that tend to happen between support customer success and life cycle marketing. So we're pretty excited about that. I Think it's a nice range that we have here, kind of like scale what's really working, build out a whole bunch of new things and then like create this experience. When we were on the phone, we were debating because in CMO land, category creation is always a big question. And certainly for a lot of early stage companies that are trying to figure out how you talk about your business. Everyone should go look up Denise's Saster session from a couple years ago. That was the path from 0 to 100 million and it's 1 of Sasser's most popular sessions of all time. Nice job, Denise.

No idea.

But she talks a little bit about these stages of growth at Snowflake and how at first they had to really describe who they were in the context of the other data warehouses and databases. And then they had this move to the cloud and at each stage they were articulating a little bit differently. And I'll let her talk about that. But also Carta's working on that and also LinkedIn is trying to shape how do we talk about sales tools and data. Are we getting grouped with the wrong people? So let's take a couple minutes and just talk about it. Maybe we'll just go through this order again. But Denise, you worked with, I think, Zoom marketing on some of your early work and then did a lot internally. But how do you think about naming your category so that your customers can understand and buy their right tools?

Yeah, I mean, in the early days, I mean, Snowflake was so much more than a traditional data warehouse and. But when you're like, your name is Snowflake, nobody's ever heard who you are, you have absolutely no credibility to try to go out and say, oh, we're this enterprise, you know, data platform. No one would even, you know, believe us. Right. And it would take sort of too long to kind of explain what we really do. And it was better for us to really break into a category that people that already existed, a category that people understood, a category that there were budgets in a for. Again, sort of our nuance then was that, okay, this is the data warehouse built for the cloud. And all the differentiation of Snowflake and all the things you could do now that you've never been able to do before came from the fact that we were built from the cloud.

Right.

And this traditional on premise or those have been just migrating into the cloud. It just had major limitations. But if we had tried to say that we were something bigger, we just wouldn't get the sort of the traction it would Be hard for the sales team to even, even get a meeting.

Right?

I mean, you need to also go after market where there are, you know, budgets and actually people buying those products and they will spend 30 minutes trying to understand, say what we do. And then the customers say, oh, so you're like a data warehouse, right. So let's, let's do that from the beginning. And that worked incredibly well. I think it's much better to kind of break into an already existing category with something significantly better than in the early days, try to do something bigger. Of course, you know, we grew a lot and then again our customers kind of pulled us into this. Like we really need like a data cloud. Right? That's what we need. We need a data platform that sits across, across the clouds. And that's kind of how we then more evolved into become the data cloud company, the data cloud positioning, you know, we have today. But again, we say the most important thing I would say that we did well, that I have never done well before with previous companies is the extreme consistency. Because when you saw Snowflake, it was always the data warehouse built for the cloud, anything, whatever it was. So we just hammered that. It was the consistency, it was the frequency.

Right.

That's what made it work. It's almost more important to be consistent than being 100% right. I would say.

Yeah. Just when we get bored of it in marketing, that market has only just started to hear us. Right. And you had the same experience. So you tried to name something at an earlier stage startup and you just didn't have the money or the traction to get it. But then you've seen success in other experiences. So this is one of my favorite topics I've actually been to, like the mothership of category creation. If you follow me on LinkedIn, I have an 8 series newsletter about creating a category. So I'll tell you what worked and what didn't. So my second time working with category design advisors and creating a category was at LinkedIn and it was successful. It was actually pretty transformational for our business. We created a category called Deep Sales. It resonated with the market. It got our sales team so excited to have a North Star to talk about. And it also differentiated us. It was well chosen because it differentiated us from shallow selling, which was a subtle dig at our competitors, but not an overt dig. So we just kind of found that sweet spot. But because we were LinkedIn, because we're large enough, we have a platform that we can speak to on a daily basis as well as get our Server enthusiasts speaking about it, we had the leverage to make deep sales a thing that mattered. And I could roll it out to a sales team of 2,000 people, we could flood the feed with it. So I had that sort of the share of voice to create something that was really meaningful and it was a game changer for my career at LinkedIn. Conversely, the first time I did it I was at a, I was CMO at an AI startup and it was series B. And as you were saying, we didn't have the leverage, we didn't have, we didn't have the budget, we didn't have the platform, we didn't have the scaled sales team. And so it's very hard to strike in the market when no one knows who you are. You don't really have strong representation and create something new. So although we really believed in the positioning, we just didn't have the resources to make it stick. And so I think there's lots of great things about it but you have to know the sort of the boundary conditions of when it works and when it doesn't. And I've learned from both. But there's lots of notes in my newsletter that you'll see online about what are the conditions that help it work really, really well for a team that has the scale well. And one of the best ways is to extend what you already have, right? I mean going from the data warehouse, our data warehouse for the cloud, to the data cloud to the AI cloud, you know, you're kind of taking more space incrementally and for Carta, you know, you, you're highly respected. I have lots of stock that I log into Carta and I like it much better than the Shareway shareworks interface. I'm an advisor to a number of growth companies and I take equity as part of my aspect. So I'm actually unique in the market because I'm always like seeing different equity and equity platforms. But one of the advantages for you is that you're trying to say we're so much more than you thought before you know how are you, you planning to get that message out to the market this year? Are you going to do a lot of in person experiences? Are you doing brand campaigns? Like what's the investment that will drive that growth? All the things and they're going to actually going to be starting this month.

So you will be seeing our new.

Sort of category defining shaping exercise out there in the wild soon. I'm really excited about it. We've done a ton of work to think about. How do we, to your point, point, like clearly communicate and consistently communicate what we do in a way that resonates because the world of venture capital and private equity can be very opaque. Right. To most people. And we want it to feel like it's very clear that the value prop is there in the, in the category.

As we define it.

So, yes, I'm really excited. We're going to have a lot coming out in the next couple months. Exciting. Well, I don't think we could have a category or a section about growth next year without talking about AI. So let's like take a couple minutes to talk about where we see the biggest opportunities coming. So Denise, you guys are. Have seen some good AI traction in the SDR organization. Maybe tell us a little bit about that. Unless there's something hotter that you want to discuss.

And I think, I mean, everyone goes through a maturation journey here, right? I mean it comes back to. It is a data maturation journey you need to go through as well. And of course, we've been using AI across marketing for so many years because we, we have the products ourselves. But in terms of them, more generative AI where we're seeing success there has really been on the our SDR side. So we have about 300 SDRs and they are in seats for 15 months. Then they move into inside sales, which means that we're hiring 200 SDRs every year. So we have 10 different cohorts in our sales development academy join happening every month essentially. And for them it's hard, right, for them to write the right emails and get the right content. So that's where we see a massive impact. So we're seeing, you know, more than double the meeting rates for those teams who've been using generative AI to, to, you know, for their messages on their emails and everything. So that's where we can go and optimize even more.

And so it's not even really getting rid of them. It's making them all more productive and.

Scaling without scaling all the racing, raising their quotas, essentially.

Yeah, yeah, yeah. Welcome to the team. How about you guys? Where are you experimenting with AI and where I'm seeing traction? I'm the most excited about marketing mix modeling from companies like Hockey Stack and Alembic. That to me could be a game changer. Plus, I'm so tired of arguing about attribution with various folks. This is like can plus up beyond attribution where you really can get inputs from a variety of sort of surfaces and channels, including brand, as opposed to just being the Channels that we can sort of identify and own but across so many different areas where we can get signal into what's working with marketing. So that to me is like quite literally the most exciting thing I've heard in probably 10 years. And do you want to give any shout outs to a tool that you're using? We're evaluating right now Hockey, Stack and Alembic.

Yeah.

You know, with the complexity of sales, I think media modeling gets even more important. We know that it's many touches to many people to get to get there. So I think that'll be an exciting part this year too. Gail, how about you? Where are you guys experimenting with AI? Anywhere you're seeing traction. So I will do this now from my product point of view as opposed to to the marketing point of view. As a really like a seven year subsidiary of Microsoft, we've been steeped in AI for a long, long time. So with the association with OpenAI that Microsoft has, we have a lot of early access to lots of the tools that they're building. So we're really talking about our agentic future. There'll be lots more from us over the coming weeks about that. But the way we think about it is there are lots of tools that are going to be very efficient at creating more spam cannons. And what we really want to do is focus on an AI tool or a set of tools that really helps you decipher what's important and what's not. And once you identify those next best actions to have this agenda tool almost your sales assistant or your marketing assistant do some of this activity for you. And so the difference between AI features and an agentic assistant is this agentic assistant actually starts to take learn your behaviors, learn your priorities and take actions on your behalf. And so we are working furiously, which is one of the reasons we were talking about PMM being so important. We are launching and testing products at such a rapid rate that keeping up with all of the competitive noise around AI is something that certainly keeps me up at night, but in a good way. So we're embracing it full on and we think it's going to do great things for the future. So last question. I think the biggest problem about being a marketer is more is more is more with less. Is there anything this year that you're going to dial down? Is there any place that you've identified? You know, we're going to try to do more here and we think that we're going to do a little bit less here because there's limited resources and we have real trade offs anyone, we're doing a bit less with web optimization. I think this idea that someone sees piece of content, fills out a form, goes to a website, becomes fodder for an SDR is. It's not been an effective way. You talked about declining rates and so you know, we could spend hours and hundreds of thousands of dollars optimizing web. We're thinking more about building community, doing a lot of influencer marketing and doing more one on one because people are so excited to be back out in the market again. So we're dialing down some of the things that I used to love like building websites where we're dialing back sort of undifferentiated events, I would say where we had very top of funnel sort of all comer, very high level thought leadership type events. Even if they aren't necessarily like that expensive. If we're doing it online, it's just a massive amount of resources to pull those off. So definitely turning the dial down on those, but turning it up on community, that's an important channel for us and I define that kind of broadly. But our community of CFOs, our community of founders and making events that are very tailored for them, but also sort of straddle brand to demand. I really try to avoid the trap that many B2B marketers focus on, on being way too bottom of the funnel at the account based. And remember that we have to make sure we're covering the full funnel. Anything you guys are dialing back on? Yeah.

Coming back to where we started. We're dialing down on things we can't scale.

Right.

Things that you know, we can't, can't repeat and dialing up on the things we can scale also. Again, I think one thing we're investing more in is also our people significantly more. I mean we're investing a lot in learning on development. We need to take on entire workforce and the marketing into the new as well. Learn about all these new technologies and how to be more efficient and all those sort of things. So learning and development is actually something where we're scaling up.

It's never been a more exciting time I think to be both in tech and marketing. Right. With the changes in SEO and the behavior of search, I think that's massively going to change our website decisions. You know, I was part of a couple, couple of other AI panels during the session and one of the discussions was a lot of what we do now is trying to use AI to adapt to our existing paradigms like how do we just rewrite the SDR's email better? And again, as we move forward, potentially in an agent landscape, they'll say, well, don't start with the email, start with the LinkedIn notification or you know, all sorts of things that are going to be brand new pair like the fact that email is still one of the most productive marketing paradigms is kind of a joke when we think about buying email lists at the beginning of our list. So it's a great year of learning and I'm just so thankful to the Sastr team for bringing so many of us together. I think let's give Sastr a round of applause. Jason and Amelia have put so much effort into making this a fantastic series and you know, they do their weekly Wednesdays where we get to see behind the scenes. And I'm really thankful for Saster for all of the thoughts and thought leadership that they capture of the best technologists. So, you know, it's like a library for us all to learn and try to understand what's coming next in the next paradigm and how do we stay ahead. So thank you so much to my speakers and enjoy the rest of the show. As a courtesy to our speakers and fellow attendees, we ask that you please silence all devices.

Hey, good afternoon, Saster. My name is John Gleason. I'm the founder and managing partner at Success Venture Partners. We help founders with the 0 to 1 before they hire their first head of customer success. Earlier today we hosted the first CCO summit here at Saster. This afternoon we're going to have our meetup, the chief customer success meetup, join the broader Saster audience. We have some awesome programming planned for this afternoon where we're going to talk about how customer success evolves in different business models and then we're going to talk about how customer success proactively sells. And so let's kick it off, everybody. Welcome to the CCO summit portion of the Saster meetup. All right, here we go.

Voice of God, Matt o', Connor, SVP.

Global Head of success Slack tableau at Salesforce, Ishmit Singh, CCO at Mulesoft and.

Vanessa Gatihi, GTM head of strategic and.

Major customer success at OpenAI.

Hey, welcome everyone. We're so excited to be here today with you all. Lots of energy across these couple of days at Saster. And again, shout out to everyone who made this possible, both Saster as well as CCO summit and customer success meetup team. So, Vanessa Katihi, I lead the strategics team at OpenAI. We're really excited to talk to you today about a couple things I think are really, really important when we think about customer success. How do you think about starting a customer success team in Org? How do you scale that and how does that evolve? Evolve as your business evolves? Both of you have really thought about this from pre IPO through IPOs and into Salesforce. So really excited to join in today. So these two people, what we're really excited about in this panel are these are two of my former colleagues. So before I was at OpenAI, I was at Slack, so at Slack Salesforce. So Ishmi was there, mealsoft, and then Matt o', Connor, which a personal story, just a human. And I'm so excited to be here. He hired me as a humble CSM many, many years ago. I'll never forget being in that interview room. It was my dream job and you hired me. And over my five years at Slack, I had many different roles in your org and I led a scale team and a high touch team. So I really learned a lot about customer success from that. So before we dive in, I want to let them each introduce themselves really quickly and I put them online the spot right before we got mic'd up and I said, a lot of people here want to understand, you know, how do we get to your level? So maybe a quick way to get to that is what do you think is a CS leader superpower that you would recommend people to cultivate? So I'll start with you.

All right, thank you, Vanessa, and so nice to be here. I'm Ishmeet. I lead customer success at Mulesoft. Just a little over a decade ago, I was a CSM there, and there's a guy somewhere in the audience who hired me to do that job. I just saw him. There he is. So thank you, Brent and Superpower, right? You know, you really put me on the spot, Vanessa. But I what I could come up in the. In the two minutes I had, you know, as I was coming up to the stage was you really have to care. And I remember one of our former students CEOs said this to us, that in his words, I'm going to try and quote him. This was 10 years ago. Like, you got to get two hoots about what your customer cares about, and you got to deliver outcomes in excess of expectations. And you can do that only if you have this innate ability and desire to care and care about your customers and also care about your team.

Matt, over to you.

All right, so first of all, it's great to be here with Ishmeet and Vanessa. Vanessa is literally the best hire I think I've ever made. I'll probably be working for her someday, I would imagine in the future and happily so. I'm Matt o'. Connor. I lead the success teams for both Slack and Tableau at Salesforce. I spent about seven and a half years at Slack, so when I joined we had eight or nine mighty CSMs and obviously the business grew over time, so that was an incredible journey. But prior to that I spent about 12 years at Salesforce. So I had a row front seat as Salesforce started to build out and subsequently evolve its success motion. And the thing I love about success is that it's ever changing. So I get to work with people like Vanessa and Ishmead and we're always learning. So it's great to be here with you all. Now the question I think the superpower I would highlight is alignment. And where that comes from is I think success for our customers is truly a team sport and sometimes we get caught up in the silos of our own functions and that's actually impacting how we deliver success, whether it's through the product, whether it's through marketing, messaging, whether it's through professional services, through sales. So I think having a keen eye on alignment as you think about how to create successful customers is a superpower.

Great. Well, thank you all. Really excited to dig into some of the content here. So I think about customer success. A lot of people can think it's a once size fit all motion.

Right.

It's new gain site. We think we all of us understand that there's a lot of importance in customer success from onboarding and renewals. But a lot in the middle can be really different and unique. And really what we want to get across in this session is really great. High quality customer success needs to be adaptable. And so to start out, one of the first questions I'm going to ask is when you're thinking about starting a customer success success org or even if you just want to take a pause and recalibrate right re kick off a customer success means in your company, how do you define or hone in on what your charter is like? What truly is the mission of CS and how do you think about that? So, Matt, I'll maybe have you start there.

Yeah, I think I love that question because I think it's so foundational. I think having clarity around the mission that customer success is unique, uniquely fulfilling in an organization is absolutely critical and it informs all the decisions you'll make. Kind of Subsequent to that. So as Salesforce was starting customer success kind of way back in the day, SaaS was new, but CRM was actually not new. It was a mature category. So when customers bought licenses, they also bought professional services and they did a huge implementation. And that actually kick started the whole adoption motion. So then success would get involved. After the implementation, we'd work with customers on best practices, activating new features and also defining the value. All of that was in service of driving the highest possible retention with our customers. Now if we did a good job, that would help drive growth. But growth was squarely the remit of our sales organization. So success was 100% focused on retention and sales was 100% focused on growth. Now fast forward to Slack. It's a different category. We had a different revenue motion where a lot of it was product led growth. There was a lot of organic adoption. And so we thought about the success motion very, very differently because the way our contracts worked is you could buy a thousand seats, but you could actually use 1500 or 2000 with no additional cost. Now we would talk to you during the renewal and that would generate a revenue growth. But the way we oriented our success organization was 100% around driving growth. And what we talked about was valuable adoption leads to growth for our customers. And so we shared that mission with our sales partners. So very different than how Salesforce approached it. But it was because of the category, it was because of our revenue model, how we did our licensing, and also the maturity of customers actually buying our products. So those were some of the things we thought about as we thought about the mission. And I think it's important for every organization to be thinking about at the beginning.

Ishmeet.

Yeah, I mean, we did it exactly the same way as Matt shared. So at Mulesoft, when I joined cs, it was primarily a renewals function. And we realized pretty, pretty early on that, you know, CS could be a growth multiplier for us. And we were at that stage about, you know, 10, 12 years ago where we were trying to establish APIs as a category and APIs weren't a, you know, discuss a topic that was discussed in the boardroom at that time. So it was hard for us to be able to explain or it was imperative for us to be able to explain the value of integration and APIs, which are often looked upon as a plumbing. Right. And not a shiny application interface. So what does that mean for cs? For us it was that we have to be build an organization that can, that can educate and reframe business stakeholders so that they understand the value of APIs. Right. And so we had to hire people who could, who had both the business and the technical acumen to be able to do so. Once we were able to do that, we pivoted and looked at our go to market motions and our go to market motions were still maturing. Right. We were good at landing new logos, but once we had a new logo, how do you expand it? So we started focusing SCS on the expansion motion and that included a lot of delivery methodologies and expansion of operating rhythms and playbooks and so on, so forth. So that's how we evolved and actually established and evolved CS along the maturity curve of the company as well as the category.

One of the things just to double click on, I think both of you said this, right? You said expansion and you talked about how you really showed that CS could drive growth and drive value. I founded the CS CS team. It'll be a year ago next week at OpenAI and it was really fun to think about this right when you're starting from scratch and you've learned lessons along the way. You know, I really came in with a strong POV. CS is a growth value driver here at OpenAI and it's a key critical part of go to market. And so I learned a lot from the way that you talked about CS in that way. So super helpful. Okay, great. So you really taken the time to define your charter. You socialize that you're aligned across different parts of the business and what CS is really going to accomplish for the company, that's unique to the company. But how do you get started? Right. You know, I think a lot of people here really want to learn from all the things that you have been through. So I don't know who wants to take this one first.

Maybe I can get started. So there's two things I would highlight. The first was unfortunately more of a lesson learned and the second one was something that we did for from the outset. So I think the first thing is coming up with a hypothesis around what does it take to make a customer successful? So what are the conditions that are required for a customer to be successful? Is it a technical implementation? Are there regulatory considerations, security considerations? Is it more of a provisioning activity? And then there's change management that follows that. And that type of conversation needs to happen with not just within success, but also including the product team, the sales team, marketing support, and really developing a shared view of what do we need to accomplish as an organization, either through the product or through experiences we're creating so that we create successful customers. And a lot will flow from that in terms of how you design your success motion. We didn't do that at the outset at Slack. I wish we had done it earlier, but that was a bit of a lesson learned. The second thing is customer journey and it doesn't need to be at the outset. A five day workshop led by external consultants. I think it slacked on probably day 20 for me. We spent half a day in a conference room with our sales partners, our SE partners, services support. And we just defined a couple simple phases. Well, there's a pre sales thing. They buy a thing, we launch them, they adopt, they mature their usage, they create value and then they renew and grow. And then we started to talk about who owns the different things that we think are going to be really valuable for those customers. Well, sales is gonna do a thing. SES are gonna do a thing. Success. Well, wait a second. If it's a really large customer, we want that to be services. Oh, okay, they should do it then. And we really defined who does what and that doesn't need to take very long. And I think there's two benefits for that. The first is if your experience is anything like mine was at Slack, it's chaos. You are hiring all these people very quickly. Everyone wants to just help and support customers. It gives a little bit of order to the cast. And I say a little bit because a lot of order is not realistic, I don't think. But then the second thing is as you learn more quarter in, quarter out, you now have scaffolding that you can start to add to and all those learnings can be applied. One of the things we learned was customers weren't prepared for their own success. They didn't have an expectation about what it took to activate 5,000 seats. So we go back to our sales team and say, hey, upstream, maybe you can talk about that with our customers, get them more ready. And that actually accelerated how quickly adoption could happen because they were ready on day one that way. So I think those are two things I would think about.

Thanks, Matt. Ishmit.

Yeah. The way we started, we did not design customer success in isolation, right? Like customer success is a cross functional activity and every function in the organization is responsible for that. And so from very early on, we involved all the different functions, starting with sales presales and services channels and support. And we designed what it would take for us to make our customers successful.

Right.

And then so we started with that framework and then we assigned, okay, here's what each of the functions are responsible for and customer success. Also, the customer success function had its own sliver.

Right.

We codified that methodology, branded it, certified the entire field on it.

Right.

Every single person in the field should be able to speak that language. And that's how we got started. Right. So as part of that, the mission and charter of customer success became very clear and how it interlinked with and intertwined with the other functions. Right. So that also, and that I will tell you was the single is one of maybe the three things that sets us apart at Mulesoft. Even now when we talk about customer success, you know, Mulesoft's most, you know, Salesforce, after we got acquired, still recognizes that as being a very unique aspect of how Mulesoft used to drive success. And we're continue to try to scale it at Salesforce.

I love that idea of the certification. I might connect with you after on how you actually went about that tactically. That's really interesting. Okay, great. So now we thought about how do you get started? You have this vision you've got going. You started to hire the right people people. You're hopefully speaking the right language across your teams. But things change. And I think you two have gotten a front seat of what happens when things change for the better. So you grow, you scale, you IPO or we did a direct public offering, so we DPO'd. And then you get acquired by a big company like Salesforce after your public company. Maybe Just talk me through. How do you think about growing and evolving your teams? Maybe your mission, what were some inflection points along that journey? Don't know. Who wants to take this first?

Sure, I'll go. You know, I try to make this more structured in my response. And as I was thinking, looking back at 10 years, you know, when we were, I joined when we were a $50 million ARR company and from the 50 to 250Cs was all about driving, driving growth.

Right?

Yeah. We were solely focused on acv.

Right.

Whether it was attrition, renewal dollars or expansion dollars.

Right.

That's how we were oriented and how to build that APIs as a category.

Right.

As we matured. So we through that process, we actually wrote the value methodology. We wrote the cross functional success methodology that I alluded to earlier. As we matured and got to sort of the 2250 to 500 range, we started to think about our customers, became more complex as well. Right. And we had to focus more on the adoption motion and how do you standardize and train our people on that. And then we also sort of became more sophisticated in who we hired. Right. Like we couldn't find all, you know, at scale people who had both the business and technical acumen. So how do you structure that? How do you hire CSMs who can continue to push API as a category but also also have the right technical and architectural expertise. Right. So those are hard hiring decisions. And then as we grew even beyond that, you know, at the sort of the billion dollar mark and above that, the cost of sales starts to become very important.

Right.

And you cannot scale the CS team at the same rate as the business is scaling in a fast growing company. So that's when like, like how do you deliver the delivery models? That discussion became more important. And you know, we started as a fairly dedicated model, high touch, high cost, where we were scaling primarily based on ARR, but now you had to start investing in pool models and digital models, you know, monetizing cs, those kinds of things, premium support offerings. So that's how we evolve cs. And at some point I remember having this, it was a few years ago, it was an interesting exercise that we did as a leadership team. We actually put these four metrics on the whiteboard and we said, how do we prioritize what we want to drive first? Right. So the metrics were renewals, expansion, adoption, I think there was consumption was the fourth one and every leader in the room had a slightly different opinion. But it was so important to us align on what are we going to first be accountable for and then what would we contribute to the next.

Right.

So I think that was a useful exercise that we did that really helped us orient around it.

So to make that tangible. In the beginning you mentioned it was dollars. So the renewals, expansion and then maybe what is it now or how did that top metric change?

I think over time, like, you know, for a smaller company, ACV is the most important, important thing. So ACV was important annual contract value.

So how much are we bringing in from a customer?

Right. And as you grow and mature, your customers become more complex, there's more bells and whistles on the technology. You've defined a category. It's super important that someone's solely waking up in the morning and thinking about the customer success. And also your sales go to market motion has matured so your AES are able to better sort of focus and be more independent. And at that point you have to have a very sophisticated adoption methodology because that's what CIOs are asking us for. Like, yes, we get your Platform. But how are you going to make us successful? So it becomes a lot for one function to be spread thin across all those four objectives. So you should be very intentional about where you're going to focus your time on.

For, for us there was a really big shift in the success motion as we started to move up market and penetrate the enterprise. What we realized very quickly, and we saw this coming, was that the enterprise buyer, we were moving from kind of departmental buying to kind of top down CIO level buying much larger companies. So moving from SMB and mid market to proper Fortune 1000 companies, they just had simply different needs. As we started to work with them, we had to start to really over overhaul our success motion. So that kind of launch phase, we had to completely overhaul that because we realized that companies were interested in activating 50,000 licenses as opposed to 5,000 had a completely different set of needs. And so we had to talk to them about governance. Well, who can create apps, who can build workflows? What do we need to consider? And so we changed the way that we launched our customers the way they adopted. We also had to have a much bigger focus on ROI because CIO level buyers are now spending millions or tens of millions of dollars. They think about TCO renewals start to become very important. How are you justifying this expense? There's alternatives on the market. So we had to build into our success motion a heavy focus on capturing and documenting the value and then socializing that with our customer well in advance of a renewal conversation where we're just a line item on a spreadsheet. And so moving into the enterprise meant that we had to really up level our game. And it also included talent too. We also needed to hire people and change our hiring profile for folks who are used to having those executive level conversations, used to talking to sophisticated IT organizations with significant requirements. So that was a big inflection point for us as we moved into the enterprise.

Yeah, I love it. I think so much of what I really heard from you both is we need to change with what we know is going to drive impact for the business, business that's really key there and being aligned. I think one interesting dimension to add to this is sometimes I think customer success also needs to focus on what your C suite really cares about. And I mention that because at OpenAI, right, we really are a research company. Like our main mission is to really build the best models out there. But there is such an intense craving from then the leadership and research arm to understand how, how our customers even using the Technology. So that was a paradigm shift for me where I always thought about ROI and use case documentation and the value of that to bring back to your CIO or to that CEO. Where for us we realized a couple months in, oh wait, actually those are equally if not more important to bring back to our research company into our C suite and that's going to help justify the headcount that we want. So that's just another. Again, you have to kind of think about a lot of things when you define what's really important and what your CSMs should focus on. Okay, great. So now we're going to talk a little bit about. You know, I think one of the things that was in the title of this talk was customer success in different business models. And we want to get a little bit more granular and what are some of the accesses that are helpful to think about when you're building or scaling or evolving your customer success team? Ishmi gonna go over to you. I'm very familiar with the Slack business as all of you now know. I'm less familiar with mulesoft. And you sell to developers, it's API, it's consumption. Maybe talk to me and us about how do those different lenses of kind of who the customer is, that customer profile really help you think about what CSMs you need to hire and what they do?

Yeah, absolutely. So mut is a, you know, unlike a lot of different applications and which have more mature, well known good market motions, Mulesoft is a bottom up sale. You have to win the hearts and minds of developers. And developers have nearly unlimited choices when they go to work in the morning on how they connect to devices or two databases or systems and MuleSoft could be one of them.

Right.

So how do you make sure that the developer is using mulesoft and it's an everyday thing that the CSM of a platform or a developer platform, developer tooling has to deal with. So it's not just enough that the CSM is helping the customer or the developer learn how to use mulesoft, but then you gotta go back and help them realize why they should come back to mulesoft.

Right.

So that means how do you reuse what you built, how do you consume somebody else, what somebody else built? Right. And how do you build that community around within the company and drive that behavioral change that developers are going to keep coming back to your tooling versus using open source frameworks or somebody else's tooling that the company already has in the house. So it's a let Me help you build with my product. Right. Let me help you consume, drive, reuse, and standardization so your company and your business stakeholders see value. And then also on top of that, let me help you shift your operating model so that not just you, but everyone in the organization is using mulesoft and getting the value out of it.

Right.

So it's, it's a pretty daunting job for a CSM to do all of those things. And it was never easy for us, Right. To be able to find somebody who can do all of those things. So that's just to give you an example of how the model is different at companies like mulesoft. And I'm sure there are a lot of others who probably have some similarities in that regard.

Anything to add, Thermat?

Just one quick one. Do we have some product leaders or maybe product founders here in the room?

We've got a few.

So even if you're a success leader, one thing that I just so deeply believe, believe is that we've talked a lot about the success motion and how to think about it philosophically. But I think it's important to mention that the most profound lever that any organization has in creating value and success for companies and your end users is actually the product itself. And that product being elegantly designed so that the onboarding flow for users is easy and very intuitive, so that it guides them through progressively using more and better aspects of the application. Because all of that work that you can accomplish through the product is infinitely scalable. And then that leaves success with actually a much smaller amount of work to do. Very, very valuable work. But I would encourage all of our success leaders to work so, so closely with your product organizations, help them understand where there's opportunities to create an even more elegant or even better customer experience through the product, because that is the gift that keeps giving and it's the ultimate leverage. And Vanessa mentioned how important it is that alignment between kind of providing that research and that voice of the customer back to the organization. So for success leaders, I would say absolutely lean into those product relationships and really think about how they can accelerate a customer getting more value from your product.

Yeah, and actually we're almost coming up to time, our last, like, bigger question was going to be on alignment. But maybe. Matt, I'll put put you on the spot. I couldn't agree more. Right. Aligning with your product leadership team. Aligning with your sales leadership team. Sometimes that's easy and that's a natural relationship, sometimes it's not. Maybe you can leave us with any Tactics and maybe, you know, if you want to add here at the end, tactics or ways that you would maybe mentor people to help build those relationships in a better, more effective way.

Yeah. I think some of the things we've already mentioned so that that notion of what is it take for a customer to be successful product has to be a part of that because the power of that is the organization can start to assemble a shared view of what are the ideal conditions for customers to get the most out of our product. To me that's like a starting point because then you go into the field, you learn from customers experience and then you bring all that back to your product organization. But I would say really establish those cadences, those rhythms that you need with your product organization and make sure they understand, understand the feedback that you're getting from your customers. One of the things that we did at Slack and much of the credit goes to the product team is they were out there in front of customers and they were spending time with us and hearing directly from the customers. Because it's one thing for Vanessa I or Ishmi to say, hey listen, the customers told me we need better this, that and the other, it's another thing for them to hear that and really do their own discovery in that. So there's many mechanisms. We had experiments that we would run with our champions community so product could get that easy, quick access to the users firsthand early in the product life cycle.

Yeah. And where, you know, again we have product leaders to join calls, we can't get them to join a lot like all of them. So those gong snippets. All right, shout out to gong. They're built on OpenAI. Gong snippets also are a really great way to scale that and share them. Ishmael, anything to add in working with sales leaders, product leaders, anything?

Yeah. I echo what Matt said about product. Right. Product is often a recipient of the news about when you talk about customers health. Right. Like who is talking about customers health? It's the cs, the field teams.

Right.

But I think you have to pivot and product has to be a participant, not a recipient, an active participant in actually defining what is a successful customer look like. And if you just go back and ask your product leaders what are the most successful customers from your standpoint, I think you might hear a different answer compared to what success leaders might say. So I think that that pivot hasn't happened a lot as I've seen so far. But I think there's a lot of interest and eagerness in product to partner in that and provide signals that, you know, actually in flight will correct a course, correct a customer or a CSM to catch things when customers are using the product not in the ways that they should be. Right. So I think that's where we. There's some room for innovation. And then sales is super critical. Right. We started with sales alignment, sort of alignment. I think sales alignment is super critical at every stage of evolution of customer success.

Awesome. Well, thank you both for your time. Really appreciate you sharing your thoughts and your lessons learned over the years. Scaling incredible customer success. Org. We're wrapping up now and giving into the next group. We are going to come back on stage after the next session and answer questions. So if you have any questions, please save them, write them on your iPhone and we'll be coming back up in a little. Thank you.

Thank you.

It.

Sam.

Ladies and gentlemen, please take your seats. Our show is about to begin. Hi, everyone. Hello. Good afternoon. We are continuing through the season SEO summit slash customer success meetup track here at Zastr. So thank you all that are joining us from the morning session. For those of you that are new and this is your first session, welcome. I'm personally very, very excited for this next Fireside chat because it features my boss and Notion's Chief Revenue Officer, Erica Anderson. My name is Monica Perez. I lead customer success at Notion, but I'm very excited to introduce the two next speakers which will be Erica Anderson of notion and GitHub's vice president of customer success, Abbas Ali Haider. Thank you. Please welcome to the Stage Erica Anderson.

CRO at Notion and Abbas Haider Ali.

VP Customer Success at GitHub.

Hello.

Hi, everyone.

Yes, thank you for the warm welcome, Monica.

It's exciting, exciting to have everyone, everyone here. I see we have a, we have a big crowd. I see a lot of familiar faces in the room as well. We, we managed to make it the first step. I did not spill any water on the stage or on myself. I do speak with both my. You know, people say speak with their hands. I speak with all my limbs, including my feet.

So the.

This requires moving it to the side. Erica, it's great to see you here. I figured we probably should start with a little bit of an introduction around how we know each other. We're from two different companies, but actually Erica hired me at GitHub and we had sort of a unique first day experience, I would say. So a couple of fun facts I wanted to share. So first of all, Erica did hire me at GitHub, but we've never met in real life. This is actually the first time they were meeting here on stage, which is sort of very strange. Also, we had a very unique first day. I'll provide my take on it, then I'll let her provide her take. So, you know, I started it at GitHub in January of last year. Very excited to join the team. This is just after Copilot announced it's like AI Hype. I was super psyched on my first day, my very first meeting with Erica. Erica said, abbas, I have some news. And I was like, like, okay, that's interesting. Let's start with some news. And she said, plus, I'm leaving GitHub. That was my very first day at GitHub and it kind of went from there. Erica, does that sound right to you?

It sounds very right. I was in a very long search for a boss and was very excited to bring him to GitHub. But yes, that is definitely what happened on his very first day, which I'd never had that, that happen before. I'd been at get up for 10 years, so that was very unique. But look at us. We're very well bonded today.

We're good now. Now I will say there was a temptation my part, say I've been playing the long game and my goal was to get her on stage here and I was just going to get up and walk away and be like, let's see how you like it. That was part of the original plan. We did decide that. However we resolved those particular challenges and now we're friends, we can hang out and have a conversation around all things customer success. And that's really what we wanted to do today is really have a pretty open conversation around a few different customer success topics, starting with some really easy ones. Like, you know, what are all the teams that go into a customer success.

Organization from your standpoint?

Yes, there's many, many. And I think it evolves over time as your company's growing and you're seeing circumstances. But you know, across my most recent company journeys, you know, obviously you have customer success proper, like the customer success managers, professional services, your customer experience team, you can have more technical roles. What am I leaving out? Renewals, account managers, the list goes on and on. But definitely an evolution. For example, at GitHub from a customer success perspective, we started with professional services actually and support and then evolved from there. At notion, it's a relatively. We've got built out teams and more roles than like the GitHub journey, which has been interesting.

Yeah, I think, I think it is.

Sort of an interesting choice of options that you have when you're building your team. And actually a fun story here is around the origin of a customer success organization. So I think for companies at earlier stages, I think it's really important to write those things down somewhere because at some point someone's going to ask you saying how did customer success start on your team? And it'll be like, well we started with support OR services or CSMs just depending on what your business needs. It can vary and actually it's a pretty long list and in fact I try to keep this list in my head as I go through and talk to different CS leaders. I probably had this conversation with a number of you on here and so there's like a menu and it's sort of like you can get on support services. We talked about, some people think about training as a separate function. You have tams, you have renewals, account managers, customer experience, engineering and the scope sort of determines the leaders you need, the functions you have the roes. It just varies based on business and the company complexity of the product. So you know, that is certainly a very common pattern to look at. It also is sort of a pattern of growth over CS over time.

Right.

So your cs Org, I'm sure you saw it, you know, evolve over GitHub, right? Like I said, started small, grew from there. And you're seeing it evolve right now with Notion as well.

Yeah, yeah, it's interesting. I mean I think you want to match it to what you're literally trying to do for customers and obviously like that be the starting point at GitHub. As I mentioned, the big need was how do you get people onto a software development platform. And that's why professional services was kind of the orientation at the start at Notion. You know, support obviously like technical support was an underpinning but customer success was early because Notion is a wonderful product but there is a learning curve. And so how you onboard customers the first product and actually hands on support from a customer success management perspective was really important. And now Notion, we're in the journey of now layering on professional services, layering on renewals and that's like the present day journey that we're on. And that's all an evolution of like what do we need to meet the needs of customers plus what the business is doing from like a performance perspective.

That's great.

Well, speaking of onboarding onto Notion, I think there's probably a few people here maybe who are Notion users. Maybe we should do that as a quick poll I'm going to get my phone ready.

Raise a hand for notion.

Here is a notion user.

Okay, thank you.

There you go.

Afterwards, I will take any customer feedback. I love to chat with users. I actually do that a couple user chats every week which is so insightful for my role in particular. So catch me afterwards. I love to know the magic wand. If you had a wish for notion, what should we change?

That's right.

I mean ultimately I guess the CRO is the ultimate csm. Is that a fair statement? That's true. It all comes down that way. You know another common question that comes up all the time. I'm sure you hear about it from people reaching out to you, seeking your expertise as well as in your operator role itself. What is the right level of investment into?

Yes, I think you know, for me, I guess I think about it a couple different ways from investment you can look at it purely like the dollars you want to spend sometimes you can look at it as percentage of ARR professional services. You've got different ways you want to look at that investment. So that's, that's one way I can talk about. I know a boss has got good interest insights on the percentage of investment and how he thinks about that. I then also think about it in terms of investment in the different areas of how we want to support customers. So you know what's the like dedicated hands on support that we need to provide to customers. What's like the way in which we can pool resources and investment and a way to reach customers and then how are we investing? Investing digitally. So that's the other way I also think about it and that's dependent on different areas of the business and what's the level of touch and experience. So I kind of look at the two vectors like the pure numbers way and then like what's the models we need to have to support the business across all the different functions you saw earlier slide of the different roles you can have as part of customer success.

I think the functional piece is really interesting, right? Because you can think about the customer journey is really driving that. It's like well forget the numbers, right? Like ultimately if our business, if we're not, if we're not doing a great job making our customers successful, like the numbers stop to matter, right? You don't have a business essentially, right. So like having the customer journey to map out and really thinking about the functional things that you need is really important, right. So this is something that you know.

You saw at GitHub with adding in.

The layers it's something they're continuing to add into on the notion side as well. From my standpoint, I very commonly get asked the question, I'm sure a number of you do here as well, is how much should we spend? Ultimately, a lot of folks who are here at this event are business leaders within their organizations. And so I get asked a question all the time. And so for those who've interacted with me before, you'll know that I have no shortage of opinions and I'm also not shy about sharing them. So I actually have a very specific number. So I tell companies all the time, 10%, that's the number you have 10% of your revenue that you can spend on all things post sales. And I can debate what's in there or not in terms of the menu we talk about, but ultimately, if you're going to run a business that can raise money effectively, that is rule of 40 that can go public, whatever scale you're looking at, if you're spending more than 10%, you better have a damn.

Good reason to for why you're doing it.

You have to be growing much faster than your peers. The quality of what you're generating as leads from CS to drive sales has to be strong. All the things have to be better if you're above that number. And the more you can monetize the conversation. When you talk about professional services, you monetize and lose a little money, but then eventually start catching up. That protects that 10% for new things that you want to do so you can pull functions in and generally build a more sustainable business. And actually for CS leaders, this is a number that you should have on your T shirt, right? So actually I'll ask this as a question. If on the spot, if someone came to you and said, how much do you spend of your revenue on customer success? How many of you have that number right at the tip of your, right at your fingertips. Quick, raise your hands. So taking a look around, that was about 15 hands may be at best. And I know there's way more customer success leaders here in this room. And so as business leaders, as you look to mature from, you know, growing in the customer success discipline, that's a number that you have to know inside out and you have to really, really strong perspective on the ROI of that. So going back to the pieces we talked about before, you have to know what is actually going to do, what are you actually delivering for?

I'd also add like, know it by the breakdown of your different functions, right? Like customer experience, for example. At Notion, which is the support organization, we track that month by month. Right. And right now our goal is 3.9% of ARR, for example. And that's like. And I encourage exactly what Abbas was saying, but even for every function. So you kind of know the breakdown and you know where you're getting your optimizations and then where you're redeploying those dollars.

Yeah. Speaking of, since you gave the number and you're tracking it.

Yeah.

I'm curious, what are, what are your thoughts on measuring customer success? Right. How do you, what are the KPIs, how do you actually measure that customer success is working?

The age old question. I'd say, you know, several things. I think, you know, the trial and true things. And I can share like how we measure it at Notion right now in terms of adoption metrics and revenue expansion. And I'm a believer in both and I can talk about that. But I'd also say that I'm very tuned to metrics at the very beginning of a customer's journey. And that's something at Notion that we've spent a lot of time working on for the purpose product itself and what we need to make customers successful. Like the onboarding is a really critical piece of the customer journey, especially at Notion. And so that's a place where we spend a lot of time coming up with those metrics. And that to me is a really important indicator of what's working and that's at the earliest part of the journey. So for example, for Notion, you know, a metric is integrations that a customer's integration enabled, for example, you know, so it's not just the, you know, usage and what you're seeing, but it's some of these other tangibles that you understand that someone's getting off on the right foot and they're going to be successful long term with your product. So integrations is an example. We've also been looking at how do you define, even in the first 90 days, the onboarding journey, how someone becomes a more advanced user. Right. So for Notion into its, can you see their database usage in editing? And so I'm super interested in the earliest part of the journey to understand how we're predicting, you know, the longer term success. Ultimately the KPIs kind of at the far end. You know, we look at what's the monthly actives, right. In terms of next quarter renewal. So that's kind of the vantage point we take. So in the next quarter, what's the month, monthly actives we want to see that at a certain percentage. And then also what's our NRR effectively? How are we contributing to the growth of a customer overall?

Yeah, I think that's great. And I think you hit on a couple of key points there. First of all being that you're not focused on like the lagging indicators. I think very commonly when people talk about KPIs for customer success teams, people quickly throw up. Grr, nrr. Like, those are my numbers. I'm like that's cool. But that's, you know, you're driving by looking in the rear view mirror. You did the right set of leading indicators looking for activity, onboarding, velocity, those durable usage over time, increasing level of depth of adoption. Those are the things that tell you that you're doing all the right stuff. And I think that's the right way to think about it. You know, in terms of KPIs on the, on the GitHub side. I remember when I joined, it was sort of funny. I was like, I joined. Erica said she was leaving. I guess I'm not totally over it. I'll keep bringing it up every little while, just a little bit. So basically I joined, I was like, hey, where are my KPIs? And actually like a lot of organizations, a majority of them were sort of focused on the, look back, the lagging indicators. And so one of my first activities as joining a new team was going through and sitting down with the leaders and customer Success, which at GitHub is Professional Services Support, CSMS. I have an engineering team that reports to me a strategy organization and sitting down and saying, what are all of the KPIs? How do we know that we're okay Based on we've done historically, that's our lagging indicators and what tells us we're going to be okay next month, next quarter, for the next year. And actually doing that is an interesting exercise. If you join GitHub today in the customer success team, one of the things you get access to is a document that I started in my first 30 days. I describe it as my fever dream document of 30 pages of KPIs. And I was like, we should pick something out of this. And now it's a contributed document from everybody. It's 50 pages of missions, unique values, KPIs, leading and lagging indicators for every single function that we review on a monthly basis. And I think that's really important. You mentioned you look at it every single week. It's not KPIs for the sake of vanity. Right? I mean, I love to kind of stand up there and be like, look at all these great numbers. I'm so excited. But that's not the point, actually. In the business reviews we commonly, and I think you've done something similar, is like, I don't want to talk about the green. Right. The green means that we put the right targets in place and we're happy with them. I only actually want to talk about the yellow and the red. And if there is no yellow or red, it makes me really uncomfortable. I get super paranoid because, like, we should be, like, doing. We should be pushing enough where some things are hovering in the danger zone.

Right?

Like, and if we're not, we're not being aggressive enough in all things in our business, whether it's customer acquisition, pipeline revenue, and customer success. Like, all the things have to actually be slight, like, slightly on the edge. Otherwise you're just not trying hard enough. The other topic I did want to talk a little bit about here was there is always a little tension between chief customer offers and CROs, especially when they leave them hanging on their first day. That tension definitely exists on those things. And so as we talk about those sorts of friction points, what are the.

Opposite sense of those?

What's a signal that tells you that magic is happening between the organizations? And really, what's magic happening in the broadest picture of customer success?

Yeah, I think internally magic is happening, and especially as you're growing really fast, I think some of these when magic is happening are really important. I think it's around, does everybody know where they fit in the customer journey in terms of there's a lot of people that interact with customers and deliver customer experience? And does everybody have an understanding of, like, what they're doing and is that aligned across the organization? And are you, like, building on that momentum versus overlapping on how you're helping drive that journey? So I would say that's like, one thing in terms of the magic happening internally, and that's probably the most important. It's like, what's. Is everybody doing? The roles, the clarity? And then how do you share the success of what's happening across those teams and where you're improving and where you have opportunity to do more, to push the envelope. So I'd say those are things that stick out internally. I think externally, when you see magic happening, I had a very great example. I don't see Monica anymore here, but about a month, month ago, we had like a magic experience from an external perspective where a customer of notion that was recently onboarded to using the product wall to wall at their company. The CEO of the company went on LinkedIn and just posted. This has nothing to do with notion. I just want to share how amazing this journey has been to get on this product and the adoption at the company has been so amazing to see and wall to wall and people are so highly engaged. And that will probably be a memory that never leaves me that like externally we made magic happen because there was no conversation that happened with us at all. Someone, a CEO, took the time to write that up and to recognize the company and the team. So that to me is magic from an external perspective.

There's no validation like customer validation when it comes to these things. And definitely that little screenshot from LinkedIn goes into that CSM's brag document for review season, right? Like it's like, definitely remember this, erica.

For like 20 years.

That's right.

If I was a CSM, I'd use that.

That definitely shows up every time you have an interview, question or wrong those things. But there's lots of elements, right? You talked about the clarity of roles, accountability for the team, the joint recognition that the team show up together. I think a little bit of this is also about the mutual respect of everyone. Kind of like knowing that you trust each other in the game. You know, one analogy I use, like we don't always talk about customer obsession. Like that's a big part of every organization that is really focused on delivering value. But sometimes customer obsession can turn creepy is how I think about it. Right? Imagine your entire revenue organization. You're always in everything in front of the customer at the same time. It's like one of them and four of you in every meeting, that is kind of creepy, right? Like that is the wrong side of obsession on the customer side. So having those lines set up where you know what you're all doing and then you have a common language by which you talk about different stages of customer life cycle. There's no sales framework that's like pre and post sale. If you decide qualification is medpic, everyone speaks a language of medpic. Whatever you pick, it's common. And then the thing that I've always loved is this idea of leadership interchangeability. Sort of like, can you walk in the shoes of your revenue peers? Right? So as a CS leader, can you actually drive a sales forecast call? If you're a sales leader, can you run a risk review meeting as well as your peers? If you do, that sends an incredible signal for magic happening within the organization itself. So those are some of the things that I look at and of course, the ultimate one one is always the customer stories. Right. Every time a customer posts about GitHub on LinkedIn and someone likes it, I'm like, yep, that's a screenshot that's going in this view. Unsolicited public praise from a customer. Nothing quite beats that. So I'm watching the countdown. It's counting down very quickly. But we have a long list of questions, so I'm going to switch. It's okay with you? Let's switch to lightning round though.

Is that all right?

So I had a series of questions. We're definitely not going to have time. People have to get us afterwards and we'll run through this. So I'm going to go through our lightning round questions. I think we literally have time for maybe yes in a sentence, no in a sentence, something like that. So some hot takes.

Okay.

All right, here we go. First one. Customer success. Monetize or not? Monetize. Yes or no? That's all you got?

Well, there's different parts of that. Okay, yeah.

Yes.

Yes. Monetize.

Yes.

Monetize. Yes. That's good. It maximizes that 10 investment you've got. Right. What about professional services? Positive margin contribution. Yes or no?

Depends.

Yes and no. It shifts over time. Ultimately, I would love that to be yes.

Okay.

No to a yes.

Okay.

I'll take that as no.

Do a yes.

That's good.

It's a journey.

I'm going to keep asking the we weird questions that require like 10 minutes to answer.

Support.

Monetize, yes or no?

Yes.

I like that one.

Very good.

Customers always are willing to pay to be front of line for support so that no one wants. If someone says vip, will you pay for the pass? Of course you'll pay for the VIP pass. That's always how it works. This is an interesting one. So, you know, teams go through model customer success. Teams have this mode where you're operating in a GM mode where different geos and stuff are kind of grouped together as a pod of all the revenue functions. So sort of pod structure versus a functional structure where all of CSS together, all of sales is together.

Which way do you like it?

These require explanation.

Lightning Rons have no. No space for nuance.

I like both. It's dependent on what you're trying to do. So, you know, for example, at GitHub it was a function functional structure.

Still is.

Still. It still is. Yes. At notion or in a pod structure and a GM structure. I think it's dependent on your. Where you're at in terms of Growth. What are you trying to accomplish?

So I think that's right. I don't think this has to be a pendulum. Just go from one to the other.

Yeah, that's true.

Last question. Since we're right right at time, pretty much. So obviously we had a fun first day. We never got to meet Rose life until today. Do we call and Text more post GitHub or while we actually had overlapping.

Post GitHub? There you go.

See a valuable life lesson. Take traumatic experiences still can allow friendships to be formed. And so that's basically, I think that wraps up topics that I wanted to cover and make sure we stay on track from a scheduling perspective. Erica, thank you so much for joining the CRO CCO relationship. Super important to make a business scale. Thank you for. For coming out here and thank you all for. For joining us this session.

Thank you.

Ram. Sa. Ra. Ram.

Hey. Good afternoon, folks. This last AMA session is going to be really special. Today we've run programming that really covers the depths of customer success. We're going to close out the day by bringing back all of the speakers that we featured as part of the CCO Summit today. You can ask all questions. Customer success, it's all up for grabs. Let's welcome our guests, our speakers, back to the stage. So come on up, guests. We have Abbas Haider from GitHub, Matt O' Connor from Slack, Vanessa from OpenAI, Nishmit from Mulesoft, and my friend Monica from Notion. All right, so here we go, folks. Here's how we're going to run this. You don't have to raise your hand. You can just come up to the mic, ask a question, and we'll go deep on customer success. All right, take it away, sir.

Question. What are the tools that are getting you excited?

The ones that you just discovered or.

You feel that would perhaps give you.

An edge that you could perhaps share with us?

I mean, I. I know one Vanessa's.

Thing called ChatGPT, because if you haven't, it's gonna. It's gonna. I mean, I think it's been around for a long time, but I've been digging a lot more into what GONG can really do. So integration into your CRM, you can understand the different types of activities that your CSMs are actually doing without having them to actually go manually log in. We were talking pretty recently. He hosted me in the Salesforce Tower and let me pepper him with an hour of questions. Really starting to think about how do we tie activities your team is doing to outcomes, but really starting to think about that in an AI first way. And so I think GONG has some interesting tools there that I'm assessing. There's other companies like people AI who do that as well. We started using them a little bit at Slack years ago. So those are the ones that are top of mind for me.

Well look, I think I have a couple of examples here. So for me I'll start with the business piece first. I think in the, in the, in the modern and new wave of tech companies that are being built up, the expectation of efficiency for all the functions increasing. Right. So used to be like 250Kr per employee was, was a standard you could use the math with. That's just not true anymore. Right. Ideally AI native companies are starting at a much higher benchmark. So for companies with existing structure to go through and adapt to that, they have to go through and find ways of driving that efficiency. As an example on the GitHub side we built using OpenAI, our own support support copilot chat that we built in house and it is incredibly effective. It deflects 60% of the conversations that comes away that I've never seen a scaling lever that effective. Now when you think about applying that across all the disciplines and customer success, that is a huge efficiency lever. So I think it's really important for companies one to experiment but I believe very strict criteria for how they measure. I'm going to put out a couple of names here to check out. They are full disclosure. They are in fact portfolio companies that in my investor life I invested in. But they are, I believe they are incredible leaders in their own right. So the first of those is Maven AGI. So they're doing incredible work. That is a company that's founded by customer success leaders out of HubSpot to build out a an AGI engine to have a conversation with to handle support and CS questions. And the other one is One Mind, which is another company that's doing fantastic work in terms of providing more of a join my meeting as an expert scale type thing in terms of CSM SE work. And I think this is a rapidly evolving area and if you're not experimenting in this area, that is a mistake. And what you should be looking for is really high benchmarks for how you're going to measure the efficacy of these systems. I'm going to double in on that.

We are both shared investors in one mind. Absolutely fantastic. There's a couple startups that we, we've had at the meetup who are really special as well. I think Reef AI is doing Incredible work around understanding the ins and outs of your customer health. I think Centralize, which was part of the last YC cohort, is also doing some really good work as well. So I would dive deep on those ones. I really like Thena as well. You know, answering your customers with their.

Questions, where they need to be met.

I think is very important as well.

So those would be three that I.

Would add to the Crunchbase search.

Actually one last one before I forget, sorry. It's an important one. It's a company called Chameleon, so I utilize their product as part of building a in app and email synchronized digital customer success journey. I think best in class in this sort of area as well. And another one worth looking at.

My comment is a little more general. So obviously at Salesforce we use Slack across the entire surface area. We've invested a lot within our lines of business, so think sales professional services support success in bespoke bots, a lot of them leveraging AI capabilities not only for question answer information and knowledge discoverability make that easier, but also we're moving towards like the creation of tasks. And so our services team, they close a deal that actually feeds like a bot message that comes to you, says hey, are you ready to set this up? You fill out a form that's all in Slack and then it creates all of the project materials and assets, hosts all that in the channel, creates a canvas. And so not just the information retrieval, but how does it actually automate tasks that usually would take like somebody 16 hours to go do all the things I just mentioned. And so that's just kind of happening across the surface area. So again, the advice of experimenting, you can never get too early a start on that.

And maybe one thing I'll just leave us on before we take the next question. I think as a leader I would say two things about choosing tools for your team that are going to like delight and create efficiency. Efficiency whether for your team or for your customer. One is the landscape is just changing so rapidly. This morning we announced OpenAI released our first model with reasoning. So this is a complete paradigm shift in how LLMs can work. So the tools that we just discussed today, those developers, if they're built on OpenAI and actually have some founders here who did it, they're like, we just changed our application to hit your new reasoning model and we're so excited to see see what the application can now do. So it's going to be changing. I would say stay plugged in where you can on how the technology is changing over Time to understand how you can leverage it for your business. And second, I do think it's important to have either a close partnership with your ops team, so whether that's rev Ops, or your own csops team, someone who's constantly thinking about these types of tools and how they're going to change and how you acquire them. And one question that I would ask when I'm thinking about any new vendor relationship, especially in this space, is what is your track record of continuing to iterate and develop as LLMs change? Because I think that's going to be more important when you choose a vendor and partner is how much are they going to be able to adapt to the LLMs adapting.

Cool.

All right, we'll get another question here. Make sure to say your name and where you're from as well.

Sure, yeah.

My name is Patrick.

I am here from Pittsburgh, Pennsylvania.

I'm the CSM at a company called Practice Beat.

We focus on patient access. We work with private medical practices and.

Lower the barrier between patients and care.

One of the things that I'm really.

Curious about, like, I have a wish list of, like, what I think AI.

Can do for me and for my.

I mean, my team is me and CS specialist. Right.

So, you know, we are currently like doing QBRs using like Google Slides, you.

Know, very like almost 20th century technology. But, you know, we're looking at, you know, proactive outreach, proactive cross sells and upsells and et cetera.

Like, we do that manually now. I want our system to do that. We have tons of customer data, etc.

When.

When I hear people talk about AI.

At Saster, everybody's kind of talking about support sales. I'm curious about your take on AI and how it can uniquely help customer success. Obviously we already heard a bunch of.

Examples and the sort of second piece.

Of that is like, where is Cesar going? Because there's like a little, a little bit of talk about, like, is AI gonna replace cs? You know, like, is the customer success dashboard gonna disappear and be replaced by, like, by an agent who actually can.

Provide more information in a shorter span of time?

I know that was.

That was a lot.

No.

Okay.

Does anyone else wanna.

I already said, I always said we've got opinions on things, so I think Ashmit gets to go first this time.

All right, Sure. I think of two areas, right, where CS leaders should be thinking about investing in AI. One is efficiency, as everyone's talked about, and the other is really customer sentiment.

Right.

And I think there's already some tooling that exists out There, whether it's, you know, generative AI, predictive AI, whatever it is, and that includes like the health of the customer. That's a space that is widely sort of wide open. My mind, mind, because how you measure health, there's no standard way. I alluded to it in my talk earlier, like how product thinks about health versus how field thinks about health is also different. How the customer thinks about health is different. And that's what some of you know, what John alluded to. Reeve AI, they're trying to solve that problem. So I think that is an area to keep a close watch on where CS can make an impact or AI tools can make an impact and then efficiency. I feel like as we think about scale scaling CS models, the problem that I've seen with a lot of scale CS models is that they lack the context. So you could throw a body, a human, and you could do a task or you could even have technology. Like there could be some digitally enforced tasks that the model could do. But ultimately customers are looking to interact with something, whether it's technology or a human that has the context of their business problem.

Right?

So if AI can be used to augment that, like for example, if you have a, you know, a person who just doesn't know anything about the customers, just asked to do a scripted engagement with the customer, that is often what we do with scale models where you could actually provide that person with a rich, on demand summary of the entire customer context so that they don't have to go and prepare for 2, 3 hours to understand what happened at this customer want. What was the history? Why were they brought in? I think that technology and that use of AI can be super powerful to help CSMs augment what they do and be more effective.

Right?

So it's efficiency, effectiveness and customer sentiment. Those three things are.

I'll jump in on this one too because I love that. I think the efficiencies are king here. I think when we think about like AI, right, the models right now are accelerating at a great pace and if we look at like the two areas and Vanessa, like definitely chime in on this. But the two things that the existing models do really, really well right now are one, write code very, very fast and number two, find the mostly right answer in very large, unstructured data sets. So I think it's like it's natural to believe that one, this will be used in support and it will free up people's time to find better answers, to spend more time with customers. If we think about like product release cycles, you can now release products at a faster pace than ever before. You can write code at unprecedented paces. And so if we think about that, like how do you know what to build now, right, when you can build faster than ever before, your technical moat is no longer as deep and wide as it once was. And so how do you stay relevant? How do you stay within product market fit? Well the way that you do that is through personal relationships, right? It's actually getting the information out of your customers to know what to build next. Right? If you think about why OpenAI has CSMs and now I'm speaking for you, but like you have the most powerful tool in the entire world, world right now at your disposal, but you still have a team of CSMs and those CSMs are going around the world to build these deep connections with customers. And it's in those connections that the.

Next generation of products will be built.

So I actually think like the role of customer success in the future only elevates and it's actually one of the best places to be in businesses if you really are getting close to customers. Customers.

Took the words out of my, out of my mouth please. I was just going to say I think my answer to that at a high level was I do think when we think about SMB Long Tail, I think that's where we're really, really going to see AI have a huge impact and I think a lot of our CS talent is really going to move up market. I also think with AI we're going to be able to do something that I don't think we've done really well. Slack ideas it a lot in this space which is meet customers where they're at. Like how can we actually understand where customers really just want and need AI and they don't need a human so that we can reallocate maybe smaller, more efficient resources and the customers actually do need our help. And I think the one thing that was on my mind that I think, you know, John just got out was CIOs and CEOs are not going to want to interact with an AI bot in the next three to five years. They're still going to want to interact with a human who understands their business, who, who has empathy for their pain points, who can tell them how other CEOs and other CIOs like them in their industry are handling these problems. I think that's all going to be really, really human. So I see, see us starting to focus and move and shift all up market on that while we do some of the enablement and all of that and low long tail be more AI.

One thing actually I'd love to add just build on to what Vanessa said. So just put like some very specific numbers on it. You're building a customer service success business and your annual spend from your customers is under 50k. There is no chance that you'll be able to build a sustainable business unless it's purely digital for csm. Right? So if you're, if you're trying to build, if your average selling price for on an annual basis winds up being between 50 and 150k, you need a hybrid model. You need sort of like digital support and some, you know, reactive engagement from a csm at 150k plus you get to have enough space end ratios to really build a high touch model. Now what does AI add into this mix? It makes all those things better. Your digital experiences become better. Maybe you can stretch it from 50 to 100k. Your scaled experiences become better. You can go from maybe up to 250k and increase the quality of those engagements at the same time at the top end you provide scale and support from all these tools to allow customers to run at higher rates. And the aggregate effect is your customers get people engagement when they need to. That engagement quality is higher and you can hire more and you can be more efficient at the use of your people resources. And ultimately when you add those things up, that's how you sort of lever up the impact that a CS organization can have by covering more customers, more revenue. It's actually the same number of people. It's not about people replacement, it's quality and spending you can actually cover with the people resources you do have.

All right, that was awesome.

Do you want to introduce yourself, say your name and where you're coming from? Sure.

I'm Raghavendra from Moxo. I help on the GDM side. I'm here from Bangalore, India. So my question is actually a point that Abbas made earlier. I wanted to understand. You spoke about the 10% metric, right? The spend on CF and this is for everybody.

I wanted to get your insight on.

How you think of a good ratio.

To have for CSM to book size. I know that can vary depending on.

The business but does that play into the 10% as well?

And how do you justify as let's say a CS leader to the revenue side for budget as showing a return especially if you have let's say not every function. Not every. If, let's say account management is not.

Playing within the CS function or if.

You have sales sort of taking care of the upgrades and responsible for the revenue number.

How are you justifying the budget? What is the return?

What is the metric you're using to show the return on investment for the function? So two questions really. What the 10%, the book size of.

The CSM and then how you're showing.

The return on the function.

Thanks.

I can maybe jump in on this one. So I joined Notion when we were at 40 million in ARR and we're at 400 million today. And so a 10x growth. We've had every possible model that you can think of and every kind of combination of that. I think that historic CS textbook would tell you, well, it's X million per head and then that defines the ratios and the types of accounts. But I think there's a lot of other factors that go into it. One matrix of how to think about this is something like product complexity and commercial opportunity on the account. At Notion, we're very lucky that we have high commercial opportunity on our accounts because it's a horizontal product. We also have high product complexity because it's infinitely configurable. And so for both of those, it made sense for us to specialize that into two different roles. We have a sales leader and a account manager and we have a CSM at other companies where maybe one of those is lower. You can have a single kind of owner of the account kind of owning all of that. So I think you have to think about like what are the specialized roles that make sense for your product and what you're trying to optimize for in the business model. So I'll start there. I will also say that like stage is really important. So like at the very beginning, very early stages, we're okay to be a little bit less efficient and maybe a little bit more expensive. We're learning as much as possible from our customer base and it's actually really important for us to be super hands on. I think as you scale, you understand like where to bake in those efficiencies. So I would say there's no hard and fast answer, but if you ask a good CS leader, they would start to ask you, okay, well what are the criteria and kind of the decision tree in their head that they should be going through. So that's, that's how I would think about it.

Yeah. One thing I want to make sure is really clear, so when you guys go back to your CEOs and say I would like 10% of my revenue for CSMs, they don't call Me and.

Say like what the hell boss?

So just be very clear. I was referring to 10% is covering all the things that fall under big customer success. All things that are post sale. Now where you choose to spend that depends on your business. Depending on stage, depending on complexity, you could spend it all on support, you could spend it all on csm, spend it all on services that you're comping to your customers. Ultimately that's the room that you have to play with to drive the outcomes you need for all of post sales. Now CSM specifically those vary at scale. Typically it's 2 to 3%. Sherrod in her AMA earlier was talking about some of those numbers as well. So definitely there's another person who's Erie can bend on numbers like she's right there. If you want to raise your hand. There you go, pointed right in the audience.

Follow up question.

And so you know, I think I just consider that you're considering all the pieces and you want to line that up the book. Size is just another variable. How you actually spend that to get the right roi. I would not start with that as any assumptive one layer or the other.

One thing I would add to it is as you're trying to figure out what should be the size, what's the spend associated with kind of a post sale experience. When we think about success in the world of Slack, we said, you know, I said in an earlier presentation we were about growth so we were able to drive adoption beyond contract and that actually generated 35% of the incremental revenue annually. So the math changes based on the contribution. So you said how do we justify, you know, our existence and investment? We pointed to 35% of ACV annual contract value comes from success. That changed the math substantially because now wait a second, that's a cost of sale. Actually we think about that bit a little, little bit differently. So once you figure out exactly what your mission is, whether it's retention oriented or whether it's growth or both, that can kind of help change some of the calculus in terms of investment from a CIO CFO's perspective.

Cool. We can go on to the next one. Please state your name and where you're coming from.

Hey, my name is Alex.

I'm from Livermore, California, about 45 minutes east.

So many questions I want to ask.

You guys but I'll just keep it to one.

It feels like there is a push lately for CS to be more sales driven with upsell and expansion opposed to strictly retention focused.

Some CSMs have a sales background, but many don't. When it comes to sales training for.

CSMs, do you believe they can go.

Through the same sales courses and mentors.

That a typical AE BDR would go.

Through or do you think there's a.

Need for CS specific sales training?

Thank you.

It's a great question.

I was putting this one someone else. Ishmeet, you take this.

Happy to. I don't know if this is just a recent thing but like we, we had a push for CSMs to be part of the sales process even like when I started Mulesoft a decade ago and, and we didn't really put CSMs through the exact same training that the AES went through. But there was enough overlap in the, in the fundamental training of how to like, like the level ones positioning of your product for example, like that training everyone in the company should be able.

To talk about that product.

Okay.

And then level two is that that's where maybe CSM should be part of that. But then I think beyond that maybe an overview of pricing and packaging.

Right.

That's important for CSMs because customers are going to reference those things. I think to the extent that the vocab CSMs need to understand the vocabulary that the customer is going to use with us that the AES have introduced to them.

Right.

And then I think there's different extents where I've seen CS organizations be part of the selling process. Right. It can be as fundamental as just I understand how the selling process works and I will only engage once this deal closes or I will engage maybe at the 50% mark or the 75% mark. And you have to make that determination based on your product and category.

Right.

And what we realized at was that before the deal closed, once the technical validation was done that the customer had bought into the product, the next question that came was solution validation. How are you going to make me successful with your product? And our AES weren't trained necessarily to be able to speak to that. So services was part of the conversation at times, sometimes they weren't. So we said okay, instead of being inconsistent about how we respond to, to that, let's bring in a CSM or a CS leader at that point where you're helping the customer to the solution validation. So that became a really powerful mechanism for the CS organization and the capabilities that we brought to the table to become a super helpful selling value proposition. You're still selling, but you're not necessarily like doing the commercial negotiations. Right. So that's how we sized that problem and that's How CS took apart of that. Now there's other models I'm sure like people use lead gen and leadpass and all kinds of things that CSMs are measured on. But I think those are more traditional ways of looking at the problem.

One thing I would add to that is I think there's a fundamental trade off between a commercial skill set. Meaning I'm used to positioning upsells, I'm used to negotiating and kind of driving a commercial kind of expansion conversation. And I'm in the business of driving it out as kind of their trusted advisor and partner. I think those are somewhat separate skill sets. You'll always find people who are unique and can do both. So you have to kind of decide, I think on a model perspective are we going hunter farmer or do AE's own kind of that commercial discussion. Because my own experience has been, and it's not to say this is the right way to think about it, is that that commercial conversation is a skill and if somebody's very good at it, you typically want them focused on it. And if adoption, driving and creating, you know, further adoption and maturity, that feels like a somewhat different skill. So you got to think about the trade offs there. Not to say one is better than the other, but I think there's a trade off to be had for sure.

Yep, we can go to the next one. We have a couple minutes left here so we're going to try to speed through the next few. I can't promise we will get to everybody.

Hi, my name is Vasu, I'm from.

Maestro qa and, and I mean I'm from New York.

Cool.

I think I saw on a slide.

From John earlier like numbers lie and.

I'm curious from your experiences, where have you seen numbers lie and what do.

You do about it?

If you can tell any stories, if.

You make bad decisions based on numbers.

Or you've caught it, you've caught the.

Lie and you did something about it.

I have an example going back to if you watched Erica's discussion around leading indicators versus lagging indicators for notion. A good example of this was how we measured onboarding. We have a huge focus on onboarding. We do think it's predictor of success. Initially when we built our onboarding metrics it was largely hypothesis based. It was like we think it's important if you start a document, we think it's important if you comment. So we had a broad sense of which metrics were were important and we built our onboarding, we call it scorecard based off of that. After we had enough data and enough customers and a data science team. Data science actually redid the analysis and they were like, hey, what you've been measuring doesn't matter at all. These metrics are not at all correlated with retention and expansion 12 months later. But we've done a massive regression analysis, we see the user base and we actually know now what are the top metrics if they happen in the first 90 days, are highly, highly correlated with metrics 12 months later. So we had to check ourselves. We were kind of surprised. As Erica mentioned, number of integrations was an important one. I thought that like commenting and interacting through the product was important. They're like not at all important. Doesn't predict any future growth. Throw it out. So we really had to be challenged to throw it out. We since redid that onboarding scorecard, it's now made up of six different measures that are proven to be related to retention and expansion. But it really questioned our initial thinking. Even as CSMs, we thought we had a good sense of that and we were seeing maybe false positives. And so having that go through like a data science team that can actually measure that against revenue outcomes, ultimately predictability is what we're hoping for is kind of what really allowed us to, to redo the metric.

Yeah. So I'll give sir a very specific example here. So I a couple of companies ago, I was at segment and when you looked at from the outside, the growth story was incredible. Right? Like huge numbers. And growth numbers are the least trustworthy numbers you can look at as a business leader because it masked like insane leaky bucket things that were going on in the business. You were like, holy crap, what is going on this business? In fact, one of your first, first hand war stories. I see Katrina right here. Katrina, raise your hand.

Real quick.

Raise. Hi, we got Sarah Burns. Where's Matt Smitterbush? Somewhere in the back.

They're talking back there.

Talking. We got Rob Schmelzer.

Where's Rob?

He's probably around here as well. Find any of them? Ask them like the war stories of what it's like to be in the system where you're like everyone. The outside is this is great. 3 billion acquisition. Amazing. You're like leaky bucket from hell.

Right?

Like that is basically a reality of how, how much numbers can sort of mask the reality of what's going on in the business.

I'll say, Matt, you had a quote. I don't even know if you remember saying this to me. But when we're talking about like, oh, you know we're hiring really quickly, but we didn't have the process. High growth hides all sins. Yeah, and that's always stuck with me, especially as I'm building out things at OpenAI. And so anyways, that's. I got that from you.

That at Slack. What Vanessa's referring to is every single chart about growth in the business. Seeds, ACV revenue, everything was up and to the right, like at an extreme level. And so that prevented us actually from interrogating more deeply to understand like the fundamentals within the business. Retention actually looked good, but we weren't seeing the maturity of usage that was probably going to be durable and healthy in the long term. And so eventually that kind of like, oh, we needed to address that. But when you're seeing a lot of growth in the business, you can ignore many, many things and everyone wants to celebrate the growth. Kind of to your point, you kind of put blinders on. So it's a lot of success very quickly is dangerous. That was kind of the same story.

Or the context behind that initial quote was good.

Numbers can, can lie. You see this especially in land and.

Expand businesses like Slack, right?

Where you land a small group, you.

Grow and you grow and you grow and you grow.

So that NRR number is fantastic. Now what eventually happens though, is you expand out all of your customers. There's no more land to expand, so all of your customers are fully saturated. So you hit this wall when you have nothing left to sell them. And so when one of those accounts does churn and you have nothing left to sell all the other accounts, that.

Churn counts for a lot.

And so what I saw in my, my own business was NRR went like this and then there was an arc to it. And then we had to actually wait till the product caught up to re accelerate that NRR number because we had to attach to all of those customers.

Once again to accelerate nrr.

That's also where the best offense or the best defense is a good offense comes from. Because if you're in a competitive market where you're expanding, if you don't get those new product attaches, your competitor might get that new product attach and now they've got a wedge into your business.

And so I found there was a.

Lot of metrics in and around NRR and GRR that could actually be rather misleading. And you really had to push down.

To the story level to understand what.

Was happening in your. And I think Jason said we could just keep rolling. Is that fair? No. Okay, so we're done. We're over everybody, thank you so much for joining us at the customer success meetup at Saster. Give a round of applause to our panelists.