Messaging & Segmentation with Zach Vidibor for Octave - GTM Engineer School (Preview)

All right, so good.

So welcome everybody. Today we're going to introduce Go to Market Engineer School briefly. And then Zach here is going to share a bit about the workflows that he's going to be teaching in the first class he gets to go first in GTM Engineer school. One of the frequent questions that I actually get about the course is, hey, I've heard of clay, but what's Octave? And yeah, it's not as well known as clay, it's not as well known as airups just yet, but what it does is really no less important. So to me, it's like the source of truth for audiences, for messaging, for Playbooks, so that all the other tools in the GTM stack can be grounded and coordinated. And that's really powerful stuff. And so that's actually why we chose to start the course with it. Okay, so that's the little intro there. Now some faces. Zach, you want to introduce yourself quickly?

Yeah. Hey, everybody. Zach Goodbor, one of the co founders of Octave. First time building a company, but spent about the last 15 plus years in the go to market trenches, if you will. So, you know, 15 years ago I was a SDR myself, hammering the phones, being on the front lines, and worked at a bunch of great companies and great teams and trying to bring all those learnings into the product we're building and hopefully share some best practices with y'. All. But, you know, my joke is, you know, when I go to sleep, I don't feel like a founder. I feel like a sales guy that's still carrying a bag. And so we're really trying to live in the trenches and, you know, build tools that really make a difference, so. Exactly. Excited to share more with y'. All.

Thanks, Zach. All right, Matteo, you want to say hi?

Hey everyone. Matteo here, Italian, but living in London for almost 10 years. And as I actually like moved here, I started my career in tech. From day one, I've been more on the marketing side of the gtm, so mostly in product marketing and content. And I've been running my consulting for the last couple of years where I really got to meet a lot of founders and just enjoying helping them build out their GTM motions. And yeah, now very excited about building GTM engineering school together with Jared and super even more glad to do it with folks like Zach and the rest of the gang.

Awesome. Well, I'm Jared. I live in Silicon Valley. My background includes both product and growth at startups like Auth0, Pubnub and Alexa and then some established companies like Adobe Intuit Yahoo, Amazon and most recently Okta. But relevant here is for myself, things move so quickly. I've always invested in staying on the cutting edge by taking courses, courses like Reforge Sales, Impact Academy, Pavilion, CMO School and recently Winning by Design. And, and for me, it's helpful to have dedicated time blocked off on my calendar where my manager, my team knows that I'm locked away learning and to be taught by experts. And so that's my inspiration here for building out this course. So thank you everyone for, for your interest and excitement around it. Say a little bit about why what's behind go to market Engineer school at this time. So yeah, I've spent the last 10 years leading growth and marketing teams at startups and I've seen this go to market motion evolve from like brute force outbound to PLG to what we're seeing now with AI powered modular systems. And it's pretty clear that AI is changing how we go to market really fast. At the other end of the spectrum though, there's a reality which is that the tech landscape is overwhelming for a lot of folks and for every breakthrough tool there's like 50 more that seem similar and you don't know where to start. And so this is a 101 level course. You don't have to be super technical, you don't have to have a subscription to these platforms. We're going to provide access to all the tools. You just need to be curious and ready to learn. Matteo, you want to say a little more around how we structured it?

Yeah, absolutely. And I think like we wanted to like with GTM Engineering school, we wanted to do something a bit different. Like we really wanted to make this an opportunity to get hands on learning and build like GTM workflows directly in class, directly with your, with your peers because, and also especially in the context of the tools where the work happens, because I think it's very easy or like it's easier to learn frameworks or strategies that are outside of the context of where the work is carried out. But I think if the AI shifts is teaching us something is that the name of the game is actually doing the thing is actually execution. And so we thought how can we help build, like help as many people as possible to build these execution skills with the AI tools, right, where the work is carried out and also from a fundamentals point of view. So how can we build? Can we master the fundamentals like the ABC and doing so while being helped from people that has already been doing the things for four years? Because there's not many GTM engineers out there and so we really wanted to gather around these experts that are pioneering this new discipline and learning directly from them hands on skills where in the tools, where the work happens basically. So, so this is how we're envisioning this program to be. Cool. And just to give you a bit of an overview of how we're or what you're going to be learning, we have a four weeks program. Every week there's two sessions of three and a half hours each again and we're going to have two hands on blocks of exercises or hands on building inside of each this 3 1/2 hours lot. The first one we're gonna, the first week is gonna be dedicated to Octave. So Zach is gonna teach everything about codifying your ICP and messaging and all this stuff that he's gonna like preview in a second. Then in week two we're bringing in, if you can go to next slide, Jared. We're actually going to be focused on cold outbound and personalization. So we're working with Yash from Clay and the Lemlitz team and we're going to learn how to build prospecting lists of companies and contacts, you know, how to then run custom research or lead scoring on these lists and then even more around how to create messages from, from each of these sort of personalized messages to target each of these. Then on week three, the module is dedicated on both Cargo and Lemlist where we're going to explore Cargo and its powerful GTM workflows capabilities. And again, how can we build custom tools to both research prospects, qualify prospects, but also like engage them based on whether they look alike, you know, like lookalikes of your best deals or even based on specific product signals if you have a PLG motion. And then in week four, we're going to like shift gears a little bit and move towards the more AI led content creation part of GTM engineering. And we're going to dive deep into Air Ops and explore things, how to use AI to create SEO content, repurpose that into social content or newsletter or optimizing existing SEO content, doing things like interlinking and other stuff. And we're going to probably come full circle and plug Octave into airops for some more advanced fun stuff and see how we can create more marketing assets by grounding airops into the Octave messaging.

And yeah, we have a ridiculously stupid, stupidly high student to teacher ratio because there's a lot of places in these tools for you to kind of get lost and we, we want to have people work in small groups with somebody there to help them, help, help them break through and, and really get through the steepest part of the learning curve. So a lot of great folks that have, that have volunteered to help teach. And with that, I'm going to hand it off to Zach.

All right, cool. Jared Mateo, thank you guys for putting this on everybody. Thanks for being here. We are really excited. This is the first, I don't know, event isn't the right word the first kind of program like this we've been a part of. And I just want to say, like, we're really excited. We are all about hands on work. The, the magic easy button isn't here. It takes, you know, art and science and I think this is a cool opportunity to blend all that together. So I'm really excited about go to Market Engineering School and today in spirit of that, we're not going to do any slides. We're going to jump straight into Octave and I'm going to start showing y' all some of what we can do and what we'll be learning about. So I'm going to share my screen and somebody give me a thumbs up. It's coming through. Cool. All right, so we're on the home screen of Octave. Some new functionality you're going to see today. We actually haven't even announced yet, so I'm on dev. So in case anything goes wrong, I'm making my excuse already for any of folks on the phone who are running agencies, consulting businesses, you can now have multiple workspaces within Octave. So today we're just going to spin up a new workspace. The other cool thing we did is we made it. So especially if you're working with clients, you might want them to pay for a subscription or you can use yours across. So we're going to use one of our existing subscriptions and we're going to start a new workspace. Let's go. Today, since they were just brought up, let's use one of our friends, Air Ops, let's build a workspace for them and we're going to call it Air Ops. And then we're going to walk through the process here. Basically right now, Octave is confirming we found the website. We're able to scrape it. Just so you guys know, we can ingest a lot of another additional information kind of into Octave's brain and you can go here and you can give us a bunch of other stuff. You can always do this later too. But basically we're going to start very simply by just going, all right, Octave, ingest my website, try and figure out what I do, start trying to model my icp. So you'll see here in a second, it's going to kind of represent its first take at like, okay, what, what is this product offering? Right? And so we can see it's breaking it down in this, like, opinionated way. What is the thing? What are our capabilities? What do we help customers with? Like, what are the benefits? Right? And all this just, you know, like, you can refine this stuff now, but nothing set in stone in the brain. In Octave, we're going to kind of keep going through this. Everything is editable, refinable. And I think that's a big part of what we're trying to build is go to market, is iterative, like, nothing is static. This all needs to change. So you can now it's trying to figure out like, okay, who are the key Personas we might sell to? And you can even see we can generate more suggestions. But let's say for Air Ops, definitely cmo. I know. Head of content, director of SEO marketing operation. Sounds good. Marketing operations leader. Great. And you can always add more later, by the way. This is just like an initial setup and then it's going to do the same thing for me here. It's trying to figure out, like, okay, what are specific use cases we can talk about? So we can do, you know, lifecycle management, you know, brand, voice of AI, real time content. We can again come up with some more if we want. You could add your own too, et cetera. We'll go with those. I hope my friends at Air Ops are happy with my choices. And then here is where we just want to ground the model, right? Like we found these are really important to like, get a model that understands your offering as, you know, what is different? What is the status quo, what are people doing today? And you could even again, if you want our help here, you can ask us to start generating a suggestion for you. But these are kind of the last two steps and hopefully you saw in a couple minutes here that quick we can finish our setup. And now basically we're taking in all those inputs to build kind of the initial footprint, if you will, the kind of foundation of like your ICP model in Octave Brain. So in a second here, it's going to spit us out. It's going to let us know again, unless Dev fails me and I can blame it all on that. It'll tell us, hey, we're done, and I'll start going in and showing you the different building blocks. Great. We can go to our workspace. Obviously you can invite collaborators, but let's go check it out. We're going to explore the library first. Here you can see this is our library. This is the foundation of the brain. You can see we have all these elements here. One thing I would highlight, reference customers. Really important to add. So think of these as like your case study customers. You can come in here and you can add. Let's see, I'm going to go, I'm looking at their website so Air Ops, I see like Statsig. So we could find Statsig and we could add them as a reference. The power of this is twofold. One, you're like helping the ICP model start to understand like, okay, here's who, here's who customers that we've actually sold to, what do they look like? And then it makes them available when we start generating messaging to go, hey, if we're reaching out to bank of America and we know Wells Fargo is a reference customer, we're going to talk about Wells Fargo. If we're reaching out to Home Depot and we've sold the Lowe's, it's going to talk about Lowe's. So it starts making the brain kind of more flexible and creating that really bespoke version of your story. So you can see it's already gone in and tried to come up with hey, what's the story about Statsig? And you can always come in here and refine this stuff. So the best part about all these elements. So you know, let's go into like our product definition. So this is really representing, you know, what, what the model. When we say Air Ops, here's what it understands and we make it really easy. If you want to fine tune your model, you don't need to do no prompting or anything like that. You can come in and we could go like make this more about content for. I don't even know what I'm saying here. For E commerce companies, I don't know, something like this again, bad example probably. But you can see here you can easily like chat with all of these elements to quickly give your team a way to like refine Okta's understanding. So you can see here we can even have this like show changes and you'll see like what it's gone through and tried to like update our kind of core positioning. It's like this is again meant to give you a really way to kind of get a fine tune understanding in the brain of what do we do? Who is it? For, and as we know, like, these products are developing and markets are changing faster than ever. So to give you a really, like, fluid space where we can keep tweaking and ensuring we're. We're keeping everything sharp and dangerous. So that's the library. I'll pause for a second, please. Anybody jump in anytime. We'll save time at the end for Q and A. But as I'm going through this, if anybody wants to shout out or slow me down, I know I have a. I have a habit of talking fast, so please don't be shy. This is the library. I would say once you're up and running here, this shouldn't take too long. And this is not where you live in Octave. Day to day. This is the foundation. This stuff over time changes at a lower velocity. What comes next would be creating a playbook.

So here, Zach, I just wanted to like, give a shout out to what you guys have built so far again, because as someone that has been founding marketer and product marketing manager many, many times, like, this is a dream.

Okay.

You know, and there's so much, you know, first principle thinking that has gone into like, understanding all the components of messaging, like making them, you know, cohesive, link them together and like the aha moment that you get from just putting a URL of your website and a couple more of your proof points from your customer references, like, the output that you get for the, for the effort put in is just outstanding. And it's not just me, but I've shared this. You know, you could say that with just a few clicks that you killed the PMM job description, you know, but you probably just also did the opposite, which is you 10x it and you made it even more actionable, right? Because when I shared this library with a couple of director of PMM at, you know, series B companies, like, everyone is like, wow, this shit is amazing. And this is just again, like the beginning. This is just the beginning. And that's where the scary part is. So I'm just going to like, shut up and give it back to you, man.

All right. New New Hero banner on the website is quote, this shit is amazing.

Zach, can I ask you a question?

Yeah, please.

So within this sort of example, Air hops is a prospective client or someone that, you know, you're trying to go after, right? Or is in this case is Air Ops the actual user of Octave.

The second, the last we're. I'm pretending I am Air Ops or.

Got it.

I'm doing work on their behalf, but I'm Spinning up the who is the ICP of Air Ops. How would we position our offering? And then now we're about to build a playbook which are. Okay, how do we take kind of that, like core positioning work and now start really like tightening the aperture and like, how do we deliver messaging for really like specific segments? So got it. Good question and good segue. Here is where again, yeah, the library. You're not going to be in there on a daily basis, but playbooks are where you kind of start using these building blocks and become more, more active. And so the playbooks are meant to be like, again, what, what really? What doesn't work today? Like, there's far too much noise to be, hey, we are, we are an XYZ company and we're great for financial services businesses. We make you more secure, right? Like, good luck with that message. No one cares. Like, you have to be really, really hyper targeted and relevant. And that is what playbooks are meant to help you do. So these are muscles here to like, again, we're trying to establish the focal point of our messaging. Like, do we want to talk to you because you're a specific type of business? Do we want to talk to you because about the function you lead. Hey, you lead a DevOps function or is there a specific business problem? Like, hey, you are. You have a high credit card fraud problem. We have a solution for that or some trigger or milestone. Like, these could be typical things we all know about hiring, funding, stuff like that. It could be more specific. It could be like, hey, we saw you just launched an app in another country and like, that's relevant to our offering. So this is meant to give you a bunch of ways to execute on those strategies, but also just brainstorm. So we're going to pick one here and you'll see the part here where we're trying to do something different is like break free from some of the thermographic characteristics. Like that are often like, people have been like, you build templates and you go like, all right, this is a great template for a industry tag manufacturing company that's, you know, employee range, more than 1,000 employees and greater than 1 billion in revenue. And that like oftentimes doesn't capture the story. You know, the story is maybe like, hey, and here's this. We'll take one of the suggestions, like e commerce retailers that are expanding, right? And like there's a qualitative dimension and we're trying to capture that. So then maybe we go, great, we'll take that. And then we go the insight would be, you know, you're just kind of putting your, your finger on the scale here and you go, you know, in Ecom, you know, lowering tack is really, let's say critical. And we can help drive, you know, long tail queries, you know, back to your products. That's good. I don't know, I'm probably not the best person at pitching Air Ops, but there's an idea and then what you're doing within the Persona or sorry within the playbook is picking Personas. So here again the beauty is like we want to prism this message kind of even further into like how would we talk to a CMO about this? How would we talk to the head of content or VP of marketing Ops? Again we can, we'll pick them all just for the sake of today's demo. And then similarly we're trying to go how do we tighten that around the use cases that really matter. So again let's take the suggestions, let's go with those and then we're going to generate a playbook. So this is where in Octave I would say people are more active. So you can imagine you might have some playbooks that are more long living and evergreen. Like hey, this is, this is our play for a segment. We're always telling this story or hey, here's for, for people that come to our website and we identify them here. Here's our message we believe is like a good fit and we can do it by Persona. And then we might have really hyper targeted. Here's how we go after DTC E commerce companies. Here's how we go after scaling SaaS businesses, right? And you'll see here now it's basically spinning up like what's the strategy? Okay so like now we have this like high growth, you know, D2C E com retailers. What do we know about them? Like how do we position to them? What would the narrative be? What's the approach angle and like more importantly, how do we distill that into like really tactical messaging we can go use? So here in the value props, you can see by Persona we're giving again a distillation of all this messaging, of all the stuff in the library into again this isn't meant to be the actual copy yet, but what are really tight, discreet, pain centric, right Ideas we can go test. Not big, broad value props, but how can we talk to these folks about the problems they care about? And so this gives you again we'll, we'll go a lot deeper in the school if should you guys join Us but you know, there's. You can enable and disable these so you can really tighten down your kind of experiments and what you want to make available to the agents. But this is kind of your, between the library and playbooks, this is kind of the foundation of the brain. The next piece that's kind of cool. And to Matteo's point earlier, it's all about how do we actually connect this strategy to execution and not this be something that just spread like peanut butter in Google Docs and notion all over the place. So building agents is the next piece. So here you can see by default there's a bunch of different agent types. We'll learn about them all later. A sequence agent, pretty self explanatory. How do we start taking that strategy around how we go after DTC E commerce companies and turn that into a campaign. So we have different types of sequences, number of steps, different models. Let's you know, crank up the horsepower on this one. And then a bunch of like really granular. And again we'll go into all these settings here, different messaging methodologies, different calls to action. Like these I would say are, you know, where we've gone really deep of what are the knobs and dials that you need to make campaigns and really drive engagement. And so we'll go here, I'll just drop myself in as an example. And in this agent playground here we can start basically iterating on what would this agent actually do when we put it out in production and we let it start running our campaigns. And so here what you're seeing, basically it's going out. The agent is scraping my profile, scraping my company, processing all that data basically through the library and the playbooks. It's trying to go like, all right, I know we're air ops, like what about Zach's background and his company matters to us and is most relevant to in this case we've hard coded it to DTC E Commerce. So you know, kind of bad example here. Like obviously I'm not an E comm business so we'll get a slight Frankenstein here. But you can see it's like telling me I'm, I'm kind of confused, right. But anyways, good example. Like it starts showing you like here's what the brain is thinking and then giving you an example of here's what the emails might look like. And then from here this is where we can have some real fun. So this is where there, there is some artistry left. We are going to dive deep in this in the school but for Example you can go, hey, I really like email 1. You can add it as an example to kind of hard code it into the agent's brain. You can tune it there. So then when it runs in the future, it knows, okay, here's a golden example. You can also even use our metaprompter so you could go make the call to action softer and offer to send a video. Case study question, Zach. Yeah, please jump in.

So like when you say like scrape your profile, obviously you script your, your own LinkedIn, right? Like Zach's LinkedIn. And this is basically telling the agent, like it's grounding the agent until it, like this is what a customer of mine looks like. Learn about it, learn about their profile, learn about what they post or like, or what they do, whether like their product and so on and so forth. And, and on the back of that craft some first touch outbound messages that could get me a reply basically.

Exactly Right. Like what, what, what we're trying to do is like the old world, I would say like your messaging isn't really responsive to your target in real time. Right. It's oftentimes. What is it? It's like a, it's a template with find and replace variables loaded up in outreach and then the, those fields get merged in from Salesforce or HubSpot and like, voila, like everybody's kind of getting the same thing. What we're trying to do with the agent is really make that like your messaging and story should be responsive to each individual recipient. So before we generate it, it's going out. All right, let's make sure we know who Zach is, what's his company, what's his background, what's that company, what do they do? And then it pulls that through to generate the story.

And how does the agent like? Because like this is like, it feels like a one to one learning process, right? So you're feeding the agent one profile and then it learns, it turns back one message for that person. How do we keep that one to one level of personalization and scale it up to, I don't know, all my prospect list or.

Great, you're leading me to the next thing. So here, let's say right there in the metaprompter, we're just showing it, it was showing how to like update all the instructions so you can see it's going, it know how to fill this out and change the settings so that, that's again one way of tweaking this. We're going to see it run again and it'll update like now, our call to action is going to be, we want to send a video case study. What's cool here. And to Mateo's point, this is meant to just be the playground. Obviously, we're not going to do this for every single person. Like, that'd be crazy. And we want to hook this into the rest of our infrastructure, into Air Ops and Clay and Cargo and Lemlist and all these other tools you're using. So we'll see in a second here. I'll let it finish. And again, once, once this runs, you know, it takes a second because it's doing all that stuff in real time at runtime. But then we'd say, like, great, let's say we're happy, we can save this agent. And then we'll give this. We'll let it come up with its own name. Great, sounds awesome. We'll save it. Here's where. Now, basically, now this agent is something that we can call programmatically. So you see here we have using Clay. Soon there'll be using Cargo, using Air Ops, many other buttons. Here we basically go, give me an API key. So now you have an API key that knows, basically, all right, this workspace that's all about Air Ops, all our information, and then this agent has a specific id. So basically in Clay. And, you know, we'll. We'll learn about all the settings and how to go deeper here in the school. But you can imagine you have a clay table with 50,000 rows. Now all you need is an API key and agent ID, and you call Octave and for 50,000 people, they are all going to get a precise, tailored to them, unique message. So that's probably. I'm looking at the clock. Like, that's probably a good time to leave off. But like, this is where, again, a lot of the magic happens from here even further is, hey, we're in clay, we're scraping other data in real time. About this Customer, are they SOC 2 certified? Can we pull product names off their website? We want to weave into our messaging. Right? And you can feed all those into Octave as well. And really, you end up with. Again, just so you guys know what we're building, like, we are not trying to build AI agents that write and, like, deceive people. Like, oh, I can't tell if this is human or not. Like, that's not really the point with Outbound. The point is you want to end up with, like, can't fake it. Personalization. Like, you just want to inject enough data that the person on the other side goes this is unmistakably meant for me, like there's not a mistake. I know they've done their homework and I'm going to, I'm going to be interested or not, but like I'm not going to be pissed off. Right? Like that's the goal here is like, how do we tap people on the shoulder in a really thoughtful, high quality, high touch manner. And you know, outbound is tough. It has to compound over time, has to have touch points across many channels. But this gives you a way to do this like at scale with like a high degree of, you know, rigor and sophistication and really tight control over your messaging.

So, Zach, we got a question from Alex.

Yeah, let's, let's jump in and do some Q and A. Yeah, there's one.

From Alex in the notes. LinkedIn.

Let me open the chat or LinkedIn automation or email only. Good question. So we don't, just to be clear, we aren't doing any of the actual sending pipes. This would be an example. So let's say you hook up through Clay Cargo, whoever you use our other friend, Lemlist, you use smart lead instantly, outreach, whatever you want. Maybe use Heyreach for doing LinkedIn automation again here, in here, obviously the kind of language is more, you know, talking about emails and things like that, but at the end of the day it's words. You can shape these. We will give you a LinkedIn specific agent soon. We're working on it. But like we have customers doing that already where you just shape this to be what you want. You know, it's, it's, it's easy enough to turn this into what would be an email is we make it slightly shorter, trimmer, slightly more casual. There you go. LinkedIn message.

Well, what other questions do we have?

Yeah, what's the impact on the credit usages for, for running these sort of plays, Zach? Like, what's the. I noticed like your, for example, your, your credits up there, you started with.

I think 18, right?

40. Now so what's, what's the, what's the.

Yeah, so in this sequence agent, how the model works here is basically steps. We charge you per email and then multiplied by the model you're using. So in this case, this is a three credit per email times four. So that'd be 12 credits per run. And this, when you're using this, just to be clear, through our API, you're not consuming clay credits or cargo credits, right? You're just, you're calling our API, you're paying for Octave credits. For this.