How to Use Data Feedback Loops to Build Massive Moats with Tomasz Tunguz, Redpoint Ventures

I think it goes back to that point that the cost of customer acquisition is increasing and this is a competitive advantage. If you can qualify a customer, convert a customer, or move them down the customer lifecycle journey using only the product. That's the leverage that you get from technology because you're not using people to do it.

I need some traction. You need some traction. Let's get some traction.

Hey, what's up innovators, entrepreneurs, visionaries and disruptors. This is your Traction podcast host Lloyd Lobo. We're a community of over 100,000 people just like yourself on a mission to help you get the methods, the money and the madness to explode your business growth. Featuring stories and tactical advice straight from those who've done it before, like Shopify, Twilio, Asana, and many more, this episode is brought to you by Boast AI. Each year the US And Canadian governments give out billions of dollars in R and D tax credits and innovation incentives to fund businesses like yours. But the application process is cumbersome, prone to frustrating audits, and receiving the money can take up to 16 months. Boast AI gets you access to research and development tax credits and innovation funding opportunities without the headache and red tape. Join thousands of North American companies leveraging Boast AI software to maximize cashback. Check out B oast AI. This episode is also brought to you by Launch Academy, an international tech hub that provides mentorship, resources, network and the environment for entrepreneurs to launch, fund and grow their startups. Since 2012, Launch Academy has incubated over 6,000 entrepreneurs, of which 300 have grown their startups past Seed and Series A and have collectively raised over $1.2 billion in funding. To learn more about Launch Academy's programs, check out launchacademy.ca Very excited for this.

Topic on moats and Tom is a prolific writer and a VC and probably one of the best VCs I know on the planet. He authored this great book on winning with Data. That's why everything he says comes with a lot of data behind it. He doesn't speak anecdotally and he writes a lot. And one of my favorite writers and we were talking about just a second ago, I read almost everything Tom publishes and very recently he came up with this novel take on SaaS architecture that you can use to disrupt incumbents and create great moats. So we're going to dive into all of it. And Tom, before we get into all the questions, give us your backstory. How did you get into vc? How did you get into writing into data?

I got into Startups When I was 17, my dad and I started a company in South America selling software to law firms. And that's what got me hooked. And then after college and grad school I went to work for a larger stage company called Appian. It was about like 100 people at the time. And I saw what a real company looked like. Now that the publicly traded business has been a hugely successful company, the got the itch to come out to California. Cause I got. I just kept writing about this company called Google, which was incredible. And there's. I just really wanted to be out here. So I got a job doing customer support for Google which is like just get in. And then I became a product manager. There was spent three years there and then been at Redpoint ever since. More than 12 years.

What was it about investing versus starting another company that drew you to it?

I love learning. It's one of the reasons I write is just to share and help other people learn. It drives me is that goal to learn. When I came out of California, I remember going to a party and I met a venture capitalist. Even working at Google, I had no exposure to VCs then. And I asked him to describe his job to me and I couldn't believe it was a thing. I couldn't believe he could be paid to do what he was doing. And I was just drawn to the idea that there was this unlimited learning curve that you could always be on the cutting edge of what humans were doing and entrepreneurs were doing. And so that's what drew me into it. And what I realized later on is actually there's a far steeper and more interesting learning curve which you and I were talking about before, which is the people learning curve as your company is experiencing hypergrowth. The company six months from now is going to be completely different than it was. And you as an individual have to change and evolve and the management team as the company. And what I saw across great board members is they were equipping those management teams to handle that change. And it's different in these situations. So that's awesome. I really enjoy that part.

And then sometimes as a founder you're wondering like what do I do anymore? Am I even needed?

Great managers manage themselves out of the jobs and then they go and they focus on even harder problems.

You wrote this awesome book, Winning with Tell us a little bit about that and what drove you to write that book.

I saw the impact and the power of data at Google, arguably the best business model on the planet. And it's a feedback loop. Users search for queries, users click on them and then you use that information in order to figure out how to rank and improve the rankings. And since joining redpoint, we've done a lot in data. We were early investors in Looker and Snowflake and Dremio and Monte Carlo and many others, CockroachDB and Timescale, which we announced two days ago. And so we just really believe in the power of data. We hatched this idea. Frank, who is the CEO at Looker, he and I hatched an idea to write a book to talk about some of the problems and the opportunities that a lot of companies face when it comes to data and to highlight how modern companies are excelling by using data. And so that's the premise of the book.

That's awesome. Yeah, it's a fantastic book if anyone wants to check it out. It's called Winning by Data and it's on Amazon, so you can look it up there and I'll drop the link after the session. You said that we haven't seen an architectural shift in SaaS since the advent of multi tenant database, and you have this great proposal for the future. Walk us through it.

Yeah, I wish I could take credit. This is something that I've just observed and I wrote about it. So I consider me a journalist rather than a creator of this idea. But if you look at like the first generation SaaS companies, they were really all about creating a system of record. And if you look at Salesforce, how do I have a really flexible database and store as much data, typed data, or enter data into a database and then build a platform on top of that? And it's like a post hoc tool. In other words, I do something and then I type it in so that the managers can know what's going on with the pipeline. And what we've been talking about for a long time is, hey, what do the next generation SaaS companies look like? And Jerry Chen had a great post on systems of engagement, systems of intelligence, and that was a wave for a while and. But it really hasn't happened. And I think it's starting to happen now. And the reason it's starting to happen now is because of the prevalence of Snowflake and other cloud data warehouses and all of the company's data being put there centrally. Once you have this huge amount of data stored inside of Snowflake, what ends up happening is that creates an enormous amount of gravity within the ecosystem because SaaS companies want to start integrating into that snowflake data. So let's make an Example, let's say I want a CRM that prioritizes leads for me based on product analytics data because I'm a freemium product. Yeah.

PLG or product qualified leads.

If I don't have that product analytics data in my CRM, if I'm an inside rep, I don't know which leads to call or I'm calling the wrong ones because I'm not benefiting from that data. Where am I going to get it? It's someplace else, but it's also in Snowflake. And so if I can create a product that sits on top of Snowflake and it takes data from Snowflake and a bunch of different sources and then it's going to be a superior product and then. But the unlock and the thing that's different is there. It's going to write back to Snowflake as opposed to writing to its own data store.

Wow.

And that, that's the second feedback loop and that's a new architecture. And the reason it's going to want it wants to write back to the data warehouse is because it's going to benefit from having more data with inside of Snowflake, its own data. The customer wants it and the customers want it for like data locality or privacy or compliance reasons. In order to be part of the next generation stack, you're going to want to integrate with all the other next generation companies and the easiest way to do that is just going to be by querying Snowflake.

I often wondered, and I still Wonder why in 2021 are CRMs empty house CRMs and you got to fill it in. So we bootstrapped the company. We use the cheapest tools. I was manually forwarding leads until March and now we migrated to HubSpot, Salesforce and huge migration effort trying to tie our database and our backend from the and everything else. And I wondered through this exercise, why doesn't something just automatically stitch it all together? Why doesn't this exist?

Because innovators dilemma. If you're Salesforce, why would you want it to exist? You have an entire ecosystem of people whose livelihoods are dependent on moving the data and keeping it in the places where it is as opposed to having it simplified. And so now you have the next generation who want to upset the Apple cart and take advantage of it. I'm really excited about this architecture. I think it's beautiful in theory. I think the reality is always going to be more complicated. If you're a brand new business, you'll be architected this way. If you're a big enterprise company. It's going to take you a while to get there. But I do think that the next generation SaaS companies, whether you're building like a product analytics company, a CRM, a marketing company, or like a BI business, whatever is being built on top, it's going to be built on top of Snowflake and have that read, write, relationship.

Since we started on that CRM theme. As an example, there was Relate iq, which wasn't an empty house CRM. And you wrote about them. Were you guys investors?

We were investors, yep.

I was so disappointed when Relate IQ was acquired by Salesforce.

We love the product. We used it internally. We were huge advocates and founders decided to sell the company when they do. And Steve decided that was the right thing for the business at that time and we totally supported him. And there's affinity now, which is a really strong product which we also use internally. So that's definitely an important iteration.

Cool. Yeah, because it was a great product. I used it back in the day and I was disappointed and that's where it needs to evolve. You're right. So let's walk a little bit into these data loops. But first walk us through what is a cloud data warehouse. Then dive into these feedback loops and the different types of feedback loops and what makes it efficient.

Yeah, for sure. So cloud data warehouse, we're investors in two. One is called Snowflake and one is called Dremio. And these cloud data warehouses are databases in the cloud that allow you to put lots of data in and query it out really fast and write and read, which is unusual. And the benefit of it is that I can put all kinds of data in there and it's in a single place and it makes it really easy to access for a company and I don't have to manage it. Here's how like the most cutting edge companies are building their business. So you've got a bunch of different data sources. Let's say you're, you know, using, you got a store on top of Shopify, you're using payments via Stripe and you're running Google Analytics. You want to take that data and put it into a place where you can analyze it. So you use an ETL tool like fivetran, which is really common, and you pump it into a data warehouse like a Snowflake or a Dremio. And then what you want to do is you want to take the rightmost column, a bunch of different SaaS applications that you've bought that are going to leverage that data. So let's say, let's go back to that sales use case. I want to see which products people are looking at on my store so when they call me, I know what to recommend. And so that's going to show up in my CRM and then that's going to change the way that I behave as a, as an ae. So I might recommend them a whole bunch of different products. I'm going to put that into the system. And what I want to have happen is when I say, hey, I recommended to Lloyd that he really ought to buy this jump rope, I want that to go back into the data warehouse and improve a recommendation system or create a marketing segmentation that's focused on people who, like Lloyd's a former bodybuilder. So let's do that like a set of protein shakes that Lloyd really likes. Let's go and make sure that people who look like Lloyd start getting advertisements that are associated with those protein shakes. And so you got two different feedback loops. The big feedback loop is the end to end, hey, how to optimize the customer journey. And then the second is given all the actions that I'm taking on top of my data, how do I improve my own internal operations?

Definitely. And those two are key. What are some issues or problems you've seen with people that are trying to implement this?

It's really new. And so I think the first is just education that this is possible. You're seeing the data stores or the data warehouses actually start to rise up and respond to this because they want to see this happen. It just drives more and more engagement. But it's fundamentally that and then also movement of the data to get it into the right place to be able to support this architecture.

Definitely that's. That's phenomenal. So what about the ownership of data? Isn't that something worth protecting rather than writing back to Snowflake? Or does data on its own have very little value? And it's the value creative created by the creativity of the use cases.

It's a really. So this is actually an architecture where you have an advantage for data ownership because you want it in one place, like today. The problem if, let's say I'm the CTO or CIO of a really large company, I have probably 150 different SaaS applications. They have all their own databases and they're all stored in a bunch of different places. And if I want to use that data, I need to do the work in order to move it from all of the 150 different database into a single database. If I want to be SOC2 compliant or ISO 27001 compliant, I need to make sure that CCPA or GDPR, all those regulations are applied to each database, like data residency. If I've got customers in Germany, I need to make sure that the data is stored in the German data center. And that's a lot of work. And so the bigger the company and the more SAS tools you have, the greater the challenge and the headache of managing that. And that's super important. And so if you can move all of this data into a centralized cloud data warehouse and then manage it there once, then you're in a much better position.

You shared all the tools here. Now are you guys investors in a lot of these companies? I know Snowflake, I know Looker you guys invested in. But like some of the others you listed in there, are you guys investors? I'm curious.

Yeah, we are. We've done a lot of investing in data. So we're in Snowflake and Dremio, which are cloud data warehouses. We're in Transform Data, which is a modeling company where investors in Looker, investors in Monte Carlo, which is a data observability company, which is like datadog for your data pipeline. So if you want to know if something's broken or the dashboard is wrong, Monte Carlo tells you that. And then we're also investors in Cyril, which is a code based platform to secure the data, to make sure that the people who are looking at the data are ought to be looking at the data.

Why did you say that this is going to transform the future? You made a few previously you talked about another architecture, right? I think last year it was, yeah.

The cloud prem architecture. It's a little bit, it runs parallel to this. So it used to be in the first generation of software you would have, you would write an application like a Microsoft app and it would talk to a database that it was in a server sitting in some closet at the office that was called on premises software on prem. And then Salesforce is a canonical example of moving that database and the application to the cloud. So both the app and the database went from local machines to in the cloud and then they made the database a multi tenant. So you had multiple customers running on the same database, which gave them huge leverage. And then now it's being, now what's happening is people are splitting the application and the database and it's being called cloud prem. So if I'm a large customer, I want that database to exist in my Amazon account or in my Snowflake account or in my Dremio account. Like I want control over the data, but I want the SaaS vendor to run their app, their application on their.

Own cloud with the feedback loops.

Yeah, yeah. And so what I'm doing is I'm splitting because let's say, let's take an example. So let's say Mattermost is deployed like this in lots of different places. We're an open source, Slack and for very large government contract, they want absolute security. So the software and the database are both deployed inside the Pentagon, but there are a bunch of customers who want the security to control their own data. And so they want the database to be under their Amazon account, but they want us to operate the application so that we can do all the updates, keep the application going. And so that application is running on our cloud but talking to their database, on their server, on their servers. And so that's what's called a cloud prem architecture. And that fits into this new feedback loop with Snowflake and Dremio at the center. Because all these SaaS applications, these next generation SaaS applications, they're talking to the Snowflake and the Dremio instances that are managed by the customer, but the application is running in the SaaS cloud.

So Bharat has a good question here. Feedback loop requires a schema awareness that is dynamic. How do you achieve it without manual intervention? Humans in the loop.

Yeah, this is a really interesting question. I don't have a great answer. I do have a hypothesis which is I think GraphQL is probably that layer. So GraphQL is a technology that Facebook created, a guy named Nick Schrock amongst others. And it's a new way of creating an API. And so what I can do is I can ask a system via API, hey, what data do you have and how is it structured? And then I can look at the response and then formulate my query based on that. Which is very different than like a Soap API or Rest API which are the two previous sort of technologies. So my sense is GraphQL is probably the answer there. But yeah, it's cool, it's great technology, check it out. But it hasn't been adopted at broad scale yet.

Let's look at applying this to disrupt like a very services oriented industry. Maybe it's legal. Right. So there's lawyers here and it's traditionally very hard to sell to lawyers anything. And it's an industry that's mired by mounds and mounds of manual work. If you had to build technology there and AI doesn't Solve all the problems. Because you need human input, right? Like you need the human input so it gets more and more intelligent. How would you digitize legal or automate Legal using this architecture?

Yeah, if it's okay, we'll talk about the accounting use case, because I think that's a great one. Okay, so if I'm an accountant today, the way that I work is I need to create all the financial statements of the company. And to do that I need data from lots of different places. I might need data from Salesforce, I need data from our contracts database. So I need data from our expenses, what our amex bills are, if we've got a corporate card or ramp or whatever service that you're using, and so on and so on. And then what I do is I take that information and I manually input it into a system like NetSuite or QuickBooks. Yeah. And then I go through and I apply a set of rules that I've learned because I'm a CPA, onto those rules and I classify all the different expenses and the revenue streams the right way. And then I output my P and L, my SaaS metrics, my balance sheet and my cash flow statement. Okay, so that's the old way of doing it. The new way of doing it is all of that data that I need to collect is actually available via API. And so I don't need to do any data entry. I just need to hook up to my stripe, my Salesforce, my whatever, my ramp. In order to get all that information, I dump it into Snowflake or Dremio, and then I've got an application, a modern SaaS application that's sitting on top of the cloud data warehouse. I put in all the rules that I care about, and then the pnl, the general ledger, the balance sheet, all that stuff are basically just SQL views, materialized views on top of the data that's in Snowflake. And the great part about that is, unlike NetSuite or other tools that need like a bunch of integrations in order to solve problems, all the data is there. I can go back, I can drill up, I can go up and down, and then if I make a correction, the system is going to learn from that in the future. And then I can go back at any point in time to the state of my balance sheet, like four months ago.

It's all manual effort right now, totally manual.

Or if I want to run different scenarios, let's say I changed my pricing strategy, that's a four month effort in Excel. But if I have all that Data inside of my cloud data warehouse. If I, with the right contract, I can just execute the query and see how things would have flown.

Do you know any accountants that are.

Doing that right now? So there's, we're spending a lot of time in this category of understanding. Okay. Which are the big buckets of companies that are going to be built this way. And there's a lot that's happening in FP in the office of the CFO because I don't think like NetSuite and QuickBooks will be the systems for the future.

There's still systems of record and I think the future is a system of intelligence as you're describing it.

Exactly. That's the way I see it happening. We will see if it's true. Gary has a question. He says, so presumably the application has its own database for users, roles, permissions that isn't company specific. Yeah. So this is a simplification. So what you're asking about is if I've got a SaaS app like a next generation CRM, I'm going to need to store some of my own data, like who has access to whatever. What would be an example? Anyway, there's going to be a bunch of app specific data that's probably not going to be stored in the customer snowflake instance. It's still going to be stored in the SaaS applications database. But my data as a customer is going to be stored within my cloud data warehouse. That's the big change. But it's a good question.

Yeah, definitely. I want to go into moats a little bit. You wrote this great blog post a little while ago and that was focused on moats and you talked about data networks, you talked about network effects, you talked about ecosystems, and you brought up Jerry Chen a little while ago. And Jerry from Greylock also has similar thesis and he came to traction a few years ago and said all the existing moats are dead. Right? Like basically you need to build new moats and they're going to be based on a very few set of things. Walk us through that. Walk us through what as being moats going into the future. Does your thesis of data networks, network effects and ecosystem creation still hold going into the future?

Yeah, I think there's an open question about like whether or not data is a competitive advantage. I think the ecosystem has come to realize, and I too, that data network effects are probably not as great of a moat as we once thought and that availability of data is much easier or access to data is much easier than most people thought. I think Ecosystem building remains. If you look at the success of open source companies or you look at the success of any of these companies within the ecosystem, quicker is a great example. One of the things that App management team did really is like we became the, the, the tool that the Snowflake account executives used to demo Snowflake. And so people wanted to buy the whole package and then worked with the DBT team and, and just really integrated there. So we became part of that next generation stack. And so I think that the ecosystem part is going to remain. Maybe the one that I would add to replace the data network effects would be plg. So I think that there's a broader backdrop. If you look at the cost across the industry, the cost of customer acquisition for SaaS, it's basically just going up.

Yeah.

And so if you're building a business there, then the question is, okay, what do I do in order to mitigate that to whatever extent that I can? Open source is a strategy to solve that problem. PLG is a strategy to solve that problem. Ecosystem building is a way of solving that problem. And if you can reduce your cost of customer acquisition or just keep it relatively constant, it's a huge competitive advantage.

Definitely. And you look at thousands of pitches, right? What stage of companies you guys invest? It seems like all the way from like a seed stage to growth is you guys go full spectrum, right?

Yeah, we'll invest anywhere from 1 to $50 million in companies all over the world. And we do a lot in SaaS and infrastructure companies and then also consumer marketplaces and then also fintech. We've been very successful with so some of our portfolio companies, like we mentioned Snowflake and Looker, but we're investors in Stripe and Twilio and hashicorp, Root Insurance, Nubank. We're fortunate to have partnered with 27 unicorns in the last five years and generated more than 120 billion in market cap during that time.

That's fantastic. And you said, you said 1 to 50 or 1555 050. Awesome. So based on all the pitch decks you're seeing, of course moats and the SaaS data network or SaaS architecture, you're having a lens on it. Where do you see like most of the companies that are exciting and you think are going to be strong going forward? Where do you see their moats being?

Yeah, it's, I think it comes down to cost of customer acquisition. I had this amazing marketing professor in grad school who said innovation equals invention plus distribution. And I think right now we're At a place where it's becoming easier and easier to build incredible software off the shelf because of all the tools that are available through the massive cloud vendors. And so distribution actually becomes a really important thing. You look at Monte Carlo and what they're doing when it comes to category creation around the data observability space, I think that's super important and just a great example of. For that people are asking, what's plg? We should clarify. Go ahead, Lloyd. I'm sure you find it really well. No, no, please.

Okay. I was just going to say it's product led growth. You're leveraging your product as a sales motion, but like dive, dive deeper into the dynamics of it. I think in the. If you look at the last decade, the biggest IPOs and most successful companies of the last decade have been product led companies from Slack to Dropbox to Zoom. What does the data tell us?

Yeah, for sure. So I was looking, I like it must have been nine months ago. If you look at the top five companies, publicly traded software companies by multiple, forward multiple. So three of the five are PLG.

And basically broadly what they're what. And some of this you were describing in the SaaS architecture, where the feedback loop comes to sales is you use your product as a way for people to use and the more value they get, you funnel those leads to sales or based on their interactions with your product. And traditionally what happens is you hire a sales team, they cold call, cold email a bunch of people, they run around in the community, they network, they build relationships. And the reverse on PLG is you provide a great product and you have a free or freemium or trial offering or maybe something more interesting. People use your product and based on how they're experiencing your product, you surface those people as warm leads for sales to contact. Did I summarize it?

Okay, yeah, that's exactly right. I think a lot of the credit for this motion. Atlassian was clearly an innovator here. And then there's a guy named Kenny Van Zant who was the architect of the go to market strategy at SolarWinds, which was a publicly traded company, and then also the architect of the go to market for Asana. And the idea behind PLG is get users to use the product, use the data to figure out which ones are going to be big customers, and then focus all the energy of the sales team on only talking to the people who've come in through the product. One of the challenges historically with PLG companies is you could divide the company in two where you had the product and the marketing people focused on the self serve leads and then the salespeople focused on leads that they were sourcing themselves by doing outbound calling. And that's tough on a company because it splits the marketing message in two. It splits the company in two, it splits the product roadmap in two. And so Kenny's insight was if the sales teams are only calling the people who are coming in through the product, then everything's aligned. And so that's what we're starting to see. And the benefit is like you have these companies now that are getting to 10, 15, 20, 25, 30 million in ARR. Like yours, Lloyd. Right. With very few people in sales, very few people in marketing. And the product is driving all of that. And it's a hugely efficient model.

Yeah. And we got a long ways to go on self serve. But the key here is we did a great session on reverse freemium with with Calendly and Calendly's chief product officer and Michael Litt from video. It's on our YouTube channel. Check it out. Reverse freemium traction comp. But incidentally, Kenny Van Zandt was my board member at Speakeasy, the previous company.

Oh, no way. Oh my gosh. Okay. Yeah, he's incredible.

He is incredible. But when you look at this, the core of this is you need product first mentality, company wide. Right. And that's why we brought on so many execs as both here. You need a self service mentality. Right. Like companies that are sales motion traditionally, which is SDRs, AES, that needs to be disruptive.

And that was the Salesforce motion. They wrote that playsbook. There's a book called Predictable Revenue. Yeah, that's right. And it was the encoding of the Salesforce model of how to build that. And now we're moving to this new model which has lots of different names like growth or PLG and was reimagining it. And as again, it goes back. I think it goes back to that point that the cost of customer acquisition is increasing and this is a competitive advantage. If you can qualify a customer, convert a customer, or move them down the customer lifecycle journey using only the product. That's the leverage that you get from technology because you're not using people to do it.

And then comes in there like things like usage based pricing or like cutting a sliver of your product and offering it for free. Where have you seen people become successful here? Because there's multiple ways to skin the product LED growth cat here. Right. It's like you offer Your whole product for free, for trial or is it like freemium? Or is it like hey, I make some pretty important but feature available for free. Where are you seeing people getting most of their growth?

We see in lots of different places is you have a bottoms up motion that's complemented by a top down motion. And so what you want to do is you want to get the one player experience to be really great so that people sign up within a company, they get really excited about a product, it spreads and then you can use that product usage data to inform an account executive calling a decision maker to buy for the team, for the department, for the company. And so you really have to split. And like open source does this really well. Right. Like I'm a software engineer, I need to solve a problem, I go to GitHub, I find elastic or whatever cockroachdb to solve my problem, I start using it and then at some point the company needs for it to be compliant. And so the VPN decides to buy the software for $500,000. That I think is like a really powerful motion. There's also the sort of let's just grow with a company over time and be patient motion. And like Atlassian is probably the best example of that where there's no sort of calling or there wasn't at the beginning. And both of those will work.

And Mark Kuo said double down on PLG machine to generate more PQLs. And should we not worry about outbound sourcing of leads? I guess if one thing is working, don't throw the baby out with the bathwater, right?

Yeah, exactly.

I think it's much stronger if you have an outbound sales machine that's working to not throw it out but complement it with PLG and then see where you get more lift over time versus like completely abandoning and refocusing the company when you really have no data to support.

Yeah, if you think like the most sophisticated go to market teams look at their lead generation and sourcing efforts as a portfolio. Just the way that you would look at your portfolio stocks, some of them are going to be in a quarter, some stocks are going to go up and some stocks are going to go down in a customer acquisition portfolio. Organic events, outbound, plg, all of those things are different parts of it and you're going to want to have a bunch of different levers to pull in order to hit your numbers. Oh, here Mark Kuo says we're 95% PLG today and have started to experiment with outbound. Yeah, it's very Natural because you want to see which parts of the portfolio can perform.

Yeah. And you recently wrote an article also, we're talking a lot about selling and customer acquisitions. And it seems like your definition of moats has changed a little bit. Like you're still saying network effects is good, you're still saying ecosystem creation is good, but you're saying maybe if you don't have a system of intelligence on top of the data, like data in itself is not value. Leverage that into a PLG motion and your PLG is probably the new suggestion for a moat. So speaking of that, you had a session on selling in the post Covid world. What are you seeing there?

We had this great session with a guy named Jim Benton, who's the CEO of Chorus. Chorus analyzed something like 15 million phone calls through Covid. We saw a couple of different things going on. The first is it takes one more call to close an account than it used to. The second is there are more decision makers actually coming into those meetings. And that was really around budget control. And so those are two of the things I remember off the top of my head.

Definitely. Awesome. And then it's funny. Interestingly, you also wrote on UiPath, which went public very recently at a 38 billion market cap. And on Tuesday we interviewed UiPaths former chief people officer who recently left and joined our lead investor Radiant Capital as partner.

Amazing.

And she was talking about people and culture. But what were the learnings there from UiPaths like the key learnings that startups can leverage? Do you think they were leveraging this sort of SaaS architecture? What were their big moats, you think? What led them to that big 38 billion market cap? Because this is an old company that hadn't raised for a while, operating like a services company at the start.

Yeah, it's an incredible business and we're not lucky enough to be investors. So I only know the story from the outside and what's public, but I think Looking through the S1, this is a company that went 169 to 351 to 580 million in ARR and pretty sensational growth. What you did see in that business is there was a, a significant increase in the acv. So they actually started moving up market and they increased their contract value, annual contract value from about 18,000 to $73,000 in the span of three years. And so my understanding of the business is they were going to market either selling direct or also through channel, helping people automate. Going back to that idea of manual data entry for the accountant Automating a lot of those different processes and then system integrators or partners would come in and actually customize those processes and optimize them. So that's the way they went to market. But what they found is if you go to the largest companies, they get way more benefit because a 5 or a 10% improvement in a process, if you're a tiny company, it's not that meaningful. But if you're a massive business, it's hugely meaningful. To give you a sense, like if you're the comcast or a DirecTV, you're probably operating a call center of two that's costing you two to $4 billion a year and each call costs like 15 to $30 each time somebody calls. If you can improve that by 10 or 15%, you're talking savings in the tens of millions of dollars. And so that's really the value proposition UiPath offers you. And that's my hypothesis for like why they've been so successful moving up into the enterprise.

Shelley asks here in government and with international NGOs, interoperability is so important that they think the answer is open source. But are the cloud data warehouse and cloud prem architecture the possible solutions to them? Because I feel like government is so ingrained and it'll never change. Like how do you. I'm not entirely sure I understand her question, but what, what she's saying is it seems like you need interoperability, but they're ingrained in their ways. And is this the possible solution to them?

Yeah, it's a really interesting question. So we have seen in the last five years, startups are selling the government way more than we used to. There are a handful of different government agencies that are super active in buying startups. And so we've got a machine translation company that's doing great in the government, a bunch of security companies that are selling their mattermost. We talked about it's got a great federal business or global public sector business. So I think that's going to be an increasing area for startups to sell into, particularly the ones focused on the enterprise or bigger accounts. In terms of her question around like open source and interoperability, Dremio is an open source cloud data warehouse. And so you store all of your files in this open source format that's wildly successful called Apache Arrow that many of the leading companies in the world store their data in. And then any application writes to Arrow or reads from Arrow can use it. And so I do think that open standards, particularly for data storage, are super, super important. And then we'll give you that interoperability. And so if you can take all that data, put it into Arrow and then run dremio on top, then you get the interoperability that Shelley's talking about.

Definitely, that's, that's some good advice here. So just to close out the discussion on SaaS architecture, one last question is how do you think startups leverage this new architecture to challenge the incumbents? Like how do you summarize all of this?

Yeah, I think the big advantage is if you're a startup and you want to disrupt an existing category, you're going to go to a customer and all the data is going to be there. And so you're not going to differentiate on aggregating the data, you're going to differentiate on making the data more useful and providing that feedback loop. And eventually what's going to end up happening is the data is going to move out of the systems of record into the cloud data warehouses and the entire ecosystem will benefit. So if you're a startup, if you're a startup founder just starting a company now, I would encourage you to bet on this architecture because it's the way that we're seeing many of the most sophisticated companies go.

Do you think the incumbents will go away with time meaning or basically adapt or die? Look at Salesforce, it's still an empty house. CRM. Look at accounting systems like QuickBooks and NextSuite, they're still empty house. Do you think they're going to be disrupted to a point where they wake up and do something or.

No, data doesn't decay, it only grows. And so if you look at Oracle was a business that was founded in the early 80s and it's still a massive company. It's because moving data out of one system into another is really tough. Right. Siebel is still around and Salesforce will be around for a long time. It's just that the next generation companies are going to be architected this way and the really forward thinking large companies are going to be architected this way. If you think about Geoffrey Moore's technology adoption curve, you've got, you know, the early adopters, the early majority of the late majority and then the laggards, the late majority and the laggards aren't going to move to this architecture for another 10 years. Salesforce can be a great business for a long time.

I think what I heard under there is bad software can also be a moat. If you're first to market in your, your software has poor user experience just By a virtue of being sticky, I have no other options. That's also a moat for a long time, maybe.

Yeah, it's a great annuity business, like an insurance company. Software companies are annuity businesses where very predictable revenue, even if you've got a terrible product.

Awesome. You have a few minutes left here. I want to close out by asking around this blog post you wrote. Never raising, always raising. So I'm facing this as well. We closed the Series A in December, and I'm getting hit up a lot for the next round when we haven't even started spending the first round. And I'm looking around me, all friends, colleagues, et cetera, they're raising lots of money. And it just seems like this cycle used to be 18 months. You need Runway for 18 months, raise your seat, get to an 18 months. 18 months and it feels like people are raising every three to six months. I have another really good friend that raised about 1880 million between a seed, a series A and a series B in six months.

Wow.

In six months. So how should founders be thinking about fundraising right now in this current environment? Valuations are bonkers. Inbound is crazy. How do you think about it?

I remember when I joined Redpoint and I started asking some of the partners who'd been there during the.com era about how did these valuation. They were talking to me about rounds of $100 million or $200 million for companies that were really early on. It's like, how do you get there? And now I feel like we're living that moment again. And that's basically being driven by the fact that there's 23% more US dollars in existence this year than there were last year. So we've increased the money supply, and over the last hundred years, the dollar has devalued by 98% was the number I read. So there's just a lot of money, and the scarcest resource is just access. Lloyd, I'm really excited about your company. There are a number of investors who are also really excited about your company. The capital is undifferentiated. So the only thing that really differentiates is actually my relationship with you, our expertise in a particular category, that kind of stuff. And so the result of this is that venture capitalists and hedge funds and growth stage investors are all competing to buy shares in these companies. And like you said, companies would go to market once every 12 to 18 months based on milestones. And now what we're saying is we believe in you. And since it's so difficult to access, we're willing to give you credit for forward execution. So execution over the next two, three, four years, because we think the company is going to be hugely valuable. And so it's the VCs that are catalyzing the rounds, as opposed to the founder saying, hey, I'm starting a fundraising process. So if you're a founder, I mean, I've been writing this line for 10 years, there's never been a better time to be a founder because the amount of capital that's available to use is just massive. So if you have a big idea, it will, it'll get funded. The downside of these really large valuations? The first is that there are three. The first is that the expectations on the execution of the business are very high because you've been given a very large round and people want to see the money being put to work and increased. The second is it becomes harder to recruit executives because your 409 valuation, the value of your options goes up a lot even if your revenue isn't increasing. So that's definitely a question. And then the third is the preference stack. So if you raise $500 million and you sell for $500 million, all that money is going to go to the investors and the employees aren't going to get anything. And so you really want to make sure you're balancing those with the capital needs of your business.

Definitely. And you're seeing all these players coming in to the VC space, reaching out to startups like, I never even heard of them, but they were traditionally hedge funds. And even the hedge fund that gave us 100 million debt facility they wanted to start with first acquiring, we said no. Then wanting to come in the aid, said no, and then gave us the big massive debt facility there. And it just seems bonkers. But what I'm seeing also is companies with 1, 2, 3, 1, let's call it 1 to 3 million in revenue raising massive Series B rounds. Is that feasible or is that the anomaly? Because I'm just quoting from my friends, or is that the norm?

So series A valuations have doubled in a year, so it used to be that they were around like 50. Like I'm looking at the top decile companies here, but like a series A valuation last year was around like 50 posts and this year it's closer to 110. 120 Series Bs have done something similar. And it's just there's a lot of money. There's this quote that I used from Stanley Druckenmiller in that post, which is it's liquidity that moves markets. And like I said, there are 30% more dollars everywhere. And so if you're an investor, like you got 30% more and you look at the value of some of these businesses that are trading in the public markets, the rewards are way bigger than Most people anticipated. UiPath of 35 billion. Hard to predict at the Series A or the Series B. People are like, well, I want to get in there. And there's an awesome book, axe feel, on the history of hedge funds. The whole idea behind the history of hedge funds is they're brilliant people who find market anomalies and then basically leverage them in order to generate great returns. The book is called More Money Than God by Sebastian Malaby, and it's a fascinating read to just see. In the beginning it was all around convertible warrant arbitrage and how it's evolved to what it is today, which is you look at the largest hedge funds, they're basically building index funds on top of growth stage venture capital. And that didn't exist before. If you were an lp, there was no product that you could buy an index and late stage private capital. But now it exists, and you said.

It rightfully so relationships transcend companies and the way to differentiate then is the relationship and. And the value add, because money eventually ends up being cheap. One last question to close out this always raising, never raising. What are the revenues that you're seeing at Series A and Series B right now that are interesting?

So we have a really hard time. I used to be able to answer this really succinctly and now I have a hard time because what a series A is, A series A can be a $5 million investment and it can be a $30 million investment. And so internally we're starting to talk about rounds as by the dollar size as opposed to Series A versus Series B. So what we see for like a company raising 5 to 15 million as their round, they're probably somewhere in the 500k to 750k range. And then a company raising 25 to 35 is probably in like 1 to 3 million in ARR range. The other thing that's happened is it used to be like a normal distribution like this, where you had a lot of companies clustered at one particular valuation and dollar size, and now it's just being pulled apart. So even the word average or typical or median doesn't really tell you the story of what's going on in the market. Everything's being pulled to the right.

And then probably at 50 million, you're seeing at the probably 8 to 10 million ARR or somewhere in that range.

Yeah. And the crazy part is like a $50 million round can go down at 300, but it can also go down at 6, 7, 800. Wow.

This is bonkers. Awesome. A few more questions here. Charles asks, what are your thoughts on enterprise SaaS enabled marketplaces and data moats with this new architecture model that you mentioned?

Yeah. So I wrote a blog post a while ago talking about SaaS enabled marketplace which was imagine you've got somebody who wants to sell software. You've got Salesforce wants to sell software to Tesla and SaaS enabled marketplace is a free SaaS enabled marketplaces. I give software away to Salesforce and I give software away to Tesla and then what I do is I create a marketplace in between and I charge a rake like an ebay or something like that. And the idea was that in order to reduce the cost of customer acquisition, let me give the software away for free and then I'll make my money in the middle. There was a push to create these companies. What they found out was that the margins within the marketplace are actually significantly lower for most businesses than they are in software. So they were talking like 5 to 30% rakes their gross margins compared to 70 to 90%. And so a lot of companies have stopped that, go to market and instead just charge for the software. That was wrong, dead wrong on that idea.

Honeybook just raised a lot of money. Honeybook in that realm there they have a sort of marketplace, but then they have a SaaS tool and then they have a network effect going on.

Yeah. So there are definitely some companies that have been able to get it to work. Honeybook is a fantastic example. But yeah, I thought it would be a sort of a bigger trend.

Yeah. But I follow NFX is another blog I follow a lot. And they said the next 10 years will be this SaaS enabled marketplaces or they call it Market Network.

They're really sharp. So I hope they're right.

They invested in Honeybook. Awesome. And then one last question here was where do you see partnerships or how do you see partnerships fitting into this whole model here?

Yeah, or bd. There's three different kinds of partnerships. There's technology alliances, which is, hey, I'm gonna integrate into somebody else's product within the ecosystem and do it in a really nice way. Then there's go to market partnerships which is, hey, let's two account executives will go and sell a combined solution and then the third is resell. So let me rent somebody else's salesforce. And I think technology alliances are super important. People should be doing that at the A or the B. The sort of combined selling. You typically only see that when you can offer somebody else something like 5 to 10 million in ARR per year. Most companies will sign up at that point and then the reselling stuff happens much later. But we're investors in a company that actually helps businesses with this and it is called Crossbeam. And so if you're interested in developing partnerships programmatically and scaling your distribution through those partnerships, you should totally talk to Bob, who's the CEO at Crossbeam. He experienced this problem firsthand and is now building the product to solve it. A really great company.

Tom, you have a great blog. It's tomtungooz.com folks. Read it religiously. You'll be a better founder and more informed technologist. If you read Tom's blog every week when he publishes it and you've got like thousands of posts there, I think you're over a thousand at least.

Yeah.

One last question to close it out. When do you think the easy money is going to end?

My sense is it ends in whenever the Fed increases rates like the first rate increase. And what the Fed has said is sometime between 2022 and 2023. It's all tied to the management of inflation. And so that's why you've seen a sell off in the big public markets. It's because of Yellen's comments on Tuesday talking about when they're going to raise rates.

Yeah. So it's not a gold rush. And don't trample over each other to go and raise money. Now wait a quarter. So the funny thing was we all, all these execs, there's a lot of inbound B and I'm like, oh, maybe we should go. Don't kill your strategy. Just make it flow naturally. But if the money comes to you, maybe take it. Awesome, Tom, take care. Thanks for joining us. And folks, tune into Tom's blog here tomtungoose.com I need some traction. Thank you for listening and we hope.

You enjoyed this week's episode of the Traction podcast.

If you enjoyed the show, please leave.

Us a five star review and you can find more information and all the resources mentioned in today's episode at Boast AI that's B O a S T a I blog.