Spenser Skates: How Perseverance Paid Off for Amplitude

0:00Almost every single company had gotten to a point where if you were to look at it from a rational standpoint, you're like, yeah, probably you should probably just quit now because you guys don't know what you're doing. but for whatever reason they didn't. The clear side to me was that if you wanted to be in that group that were successful, if you just stuck through that moment, your chances of being successful would go way up. That's Spencer Skates, founder and CEO of Amplitude. Amplitude has done more than anyone to make user behavior legible so product teams can stop guessing at product decisions.

0:35This is Mike Maples Jr. of Floodgate and it's go time with Spencer Skates. This is Mike Maples Jr. and welcome to the Pattern Breakers podcast, where we explore why some founders radically change the future and how they stand apart. Together, we'll learn about the counterintuitive mindsets and actions behind their remarkable success. Brace yourself for a world where chaos is welcome, naysayers are often a positive signal, and movements galvanize misfits who transform the impossible to the inevitable.

1:18When Amplitude launched, the analytics market looked crowded, and for over a year there was essentially no traction. By conventional thinking, the odds were bad. But Spencer prioritized truth-seeking over narratives. He studied past YC outcomes and saw a simple rule. Most companies die before they even learn enough to make something that people want. If you survive long enough, probability shifts. Two years was his threshold. So Spencer and his co-founders determined to stay alive past that milestone. They learned how to identify the customers who truly valued Amplitude's advantage. They developed the skills they needed one day at a time.

2:00And ultimately, their learn-it-all mindset and grit led not just to survival, but a publicly traded company valued in the billions. Let's talk to them.

2:19Spencer Skates, welcome to the show. Fantastic to be here. Good to see you again. Yeah, likewise. So before we jump into Amplitude, there was a company before that called Sonalite that maybe not everybody knows about. It kind of frames the origin story well in a lot of different ways. So I was hoping that you could just share what Sonalite was, what was up with that, why did you pivot, and how did it lead you to Amplitude? So before we did Amplitude, so we had Sonalite, which was a voice recognition app, and it was like a version of Siri for Android phones. before Siri came out. It was actually funny.

2:57We did the Y Combinator winner 2012 batch. And in the middle of the batch, Siri launched. And so during demo day, we called ourselves Siri on steroids, which really resonated with a lot of the press. We had this really cool demo where I took my phone and then I put it into my jacket pocket. And then I had a conversation with my phone while I was on stage. And it blew the audience away because it was like I was doing a magic trick, you know, in front of all these investors at YC Demo Day. But the funny part was that the product, it wasn't that great. Like it had a really cool demo. We had a little bit of traction.

3:37You know, there were something like a few hundred thousand downloads. So somebody was using it, which was cool. But we had a really hard time retaining users. And the quality of the voice recognition technology was just not there. And so even though we had this incredible demo day, you know, we weren't able to fundraise after. I was convinced there were three angels that put money into us at the time. And I remember all three of them. And I was convinced that the only reason that they invested in us was because they were investing in all the lame companies in the YC batch. And that was their selection criteria.

4:09So we did that. Demo day was in March. By the time we got to May, we were like, all right, we should probably shut this thing down because we could keep working on it. But it's going to be a long, hard, grindy road. and we're not particularly passionate about this technology and we don't know if this is going to be successful. So let's go try to find something else. Then that led us to, we had spent all this time building our own analytics in-house. We spent like a third of our time trying to understand what users were doing in the product because it was so obvious to us, this is the right way to look at and build products.

4:41And so that eventually led to Amplitude and we started that in June of 2012. Now I've heard varying rumors about the early days of Amplitude. So like one thing that I've heard attributed to you is that part of your goal was to survive long enough. Like there's kind of this heuristic that I've heard some founders suggest, which is if you survive long enough, you just keep surviving and eventually something happens. And like a lot of companies fail because they decide to quit rather than just keep persisting. And so to what degree is that rumor true? Oh, so that was completely how we thought about it.

5:23So before even starting Sonalite, I didn't even know, like I was just some college kid. We graduated from MIT in 2010, me and my co-founder Curtis, and I wanted to know whether I was capable of starting a company. And so I did this research where I looked at the Y Combinator batch for two years before. And I was looking at like, okay, two years later, how many are successful? And it was like, I think a third kind of had gone on to some clear success. A third had gone on to shut down and fail. And then a third were kind of, you couldn't tell. They were still out there, but it wasn't clear of how well they were doing.

6:02And so I started to get, wanting to get really serious about studying this. And so I just read all these stories. I read Jessica Livingston's founders at work. I don't know if you've come across it before, but just incredible set of stories from the early days of entrepreneurs. I read another one that was called Startups Open Source. And one of the really clear themes that stuck out, almost every single company had gotten to a point where from a rational standpoint, they should have quit. And this was normally six months, nine months, a year into the business. they weren't making progress they didn't have traction if you were to look at it from a rational standpoint you're like yeah probably you should probably just quit now because you guys don't know what you're doing but for whatever reason they did it the clear side to me was that if you wanted to be in that group that were successful if you just stuck through that moment your chances of being successful would go way up it was so crazy how our experience bore that out because watching our Y Combinator batch, after not getting traction in the batch, they'd kind of go back to their day jobs.

7:08Like a lot of them, or the co-founders would fight and break up. Whereas for me and Curtis, it was like, we get, you know, we do YC, we have this awesome, cool demo, we get all this press, but then, you know, we don't get any investors and we see that we can't really retain our users. And it was no question to us that we were going to stick it out and work on something else. It was just a question of what. So we spent that May to June period in 2012 just looking at lots of different ideas and figuring out what we want to work on and eventually coming on the analytics thing. Some ups and downs, but the whole thing ended up working out.

7:43But the important part was that there was no question in our minds that we were going to stick after this problem. And even if the first thing we don't work on with Sonalite doesn't turn out, we will be more formidable and we'll go after it again. How did the Sonalite idea lead you to the idea for Amplitude? I mean, it sounds like that your experience in measuring the success of the product somewhat led you there, but were there other factors that led you to the idea for Amplitude? Yeah, the biggest thing was it was so obvious to us, this is how you should build product. You should look at the data of how people are using it because there's nothing more powerful, right?

8:20People say, you know, talk is cheap, but if you want to know what someone really cares about, look at what they do. We had this one question, which was, how important was the accuracy of the voice recognition to the long-term retention? So if you came into the product, it asked you, hey, who do you want to text? And then you say a name. And then do you have a successful match against your contact book on that first prompt? Or does it fail and you have to redo it? We were like, all right, what do we do? Let's look at the data. So we ended up spending all this time collecting data and building a system that we could understand and visualize this question.

8:58It turns out the accuracy of the voice recognition is incredibly important. If you had a successful first recognition event, your chance of retaining was double than if you didn't. So massively impactful. So that means, hey, if you make the recognition 1 % more accurate, you're going to have 1 % more retention. Or if it's 10 % more accurate, you're going to have 10%. So it's a lot of, it actually matters a lot if you put energy into this thing. You know, that ended up leading us to realize that, okay, we can only improve the accuracy of this thing so much. And we can't really get it to the point where it is going to be good enough retention.

9:31It's going to be a usable product. And so, you know, funnily enough, it was kind of this existential question. And we answered, okay, this is actually not a viable product and a viable business. But we realized a lot of other people are going to want to know the same thing. Huh. And I guess there were other tools at the time, right? You had Google Analytics and you had Flurry and you had Mixpanel, all these guys. What was it that they weren't able to do that you thought was necessary to do? Was it literally this product instrumentation relative to retention? So they weren't able to do that, but it was even more crowded.

10:05It was incredibly crowded. So I'll list off the ones we tried. So Google Analytics, there was Adobe, which was the enterprise guys. There was Flurry, which was big on mobile. There was AppSolar, which was mobile attribution. There was Localytics. There was Mixpanel. There was Kissmetrics. There was Keen, which was focused on developers. The list just goes on and on and on. There's like 30 something odd competitors. And you're like, all right, do you really think you're going to be smarter than all these guys? Well, the one thing we saw was that being able to look at the user on the basis of behavior and say, is this a good behavior?

10:39Is this a bad? None of these other guys were able to do that. And so we said, okay, we at least know we can build that and that's different enough. And we know that something our batchmates valued. I mean, we would show, you know, when we talk to our batchmates and Y Combinator, we'd show them, you know, these analyses we did on voice recognition accuracy. and they're like, holy shit, this is like way more advanced than anything we're doing. Like, can we get that? You know, your voice recognition looks kind of crappy, but can we get that analytics? We were also a bunch of algorithms guys. Like Curtis and I won MIT's largest programming competition, Battlecode.

11:15And one of the key things is, okay, can you write, create a distributed system and write algorithms at scale? And analytics, analytics was so much more of an attractive problem to us than voice recognition because in analytics, there's like a clear right answer. And it's just, can you do it efficiently and fast? And, um, whereas like probabilistic problems, it's like, God help you. So one of the things I'm wondering about so far in this Spencer is, um, in your first product in Sonalite, even though it didn't work out, you had a demo that got instant attention by a whole lot of people quickly. And it sounds like that rhymes a little bit with the early days of amplitude.

11:57You could show somebody something and very quickly they would have the, you know, your words, right? Holy shit moment and say, wow, you know, I need that. Can you give that to me? Is that a fair assessment that, you know, when you've done these products that you've tried to be able to demonstrate the advantage really quickly so that people could instantly see what you're talking about and why it matters and why they should care? That's a good, I've never been asked that before, Mike. That's a really good question. I mean, for sure, for sure, Sonalite had that wow factor to the demo. With Amplitude, like this is like a data intensive nerdy product.

12:41And so the average person would be like, okay, you're showing a bunch of graphs, who cares? But for people who were really into data, they got it when you showed them. I remember one of our early groups of customers was a bunch of ex-Zynga product managers that had gone on to start their own companies. And they were so used to having this sort of stuff, and they had believed in the same way of building product that we did. And so when they saw it, they were like, oh, okay, yes, great, sign me up for it. One of the mistakes we actually made in the early days of Amplitude, it was a good year and a half before we got real traction.

13:15And so we didn't have that wow moment early on. And it was a real grind to get some of the basics together. And the mistake we made, frankly, was we tried to convince a lot of companies that were not as passionate about data as we were to use what we were doing. And we were basically begging them to use it for free. And it turns out out of the first 20 or so companies that were trying it out, none of them ended up paying or becoming customers of ours long term. It took us a little time to figure out the right group of people that did have that wow moment with Amplitude. So, and what was it about the early adopter successful believers that was different?

13:56Yeah. So, so this, the, the first one is this guy named Brett Terrell, who's the, who is the CTO and co-founder of this company called Super Lucky Casino. So they're, they're a mobile slots game. And he had worked at Zynga as a product leader before, and he had left to start his own mobile app company. And he just had this big gap where he was expecting, he knew this was the right way to build a product with data, but he didn't have anything that could. He was trying to use Flurry and Google Analytics and trying to cobble it together himself. And so I remember this was It was July of 2013, I think, where we came in, we do the demo to him.

14:36He asked some questions. We get to the end of the session. And at the end of the session, he's like, how much does this cost? And I was shocked to hear that question because I'm like, what? How much does this cost? Oh, because in my head, I was like, oh, I was going to beg you to use this for free. But I was like, all right, let's test this guy out. Let's see how much he could possibly pay. And so I was like, all right, what does SaaS cost? And I'm like, well,$50 a month. But I'm like, you know what? They always say charge more, ask for more money. So I'm like, let me think of a bigger number than 50.

15:09What's bigger than 50? Well, if you double that, 100. 100 is bigger than 50, right, Mike? I'm like, then right before I say 100, I'm like, no, no, no, no. I got to put an extra zero. I'm going to say$1 ,000. So I was like,$1 ,000 a month, which is much confidence as I could say to him. And to my absolute surprise, he was like, wow, that's so cheap. I really want this. Let's do it. And I'm like, oh my fucking God, this guy actually wants to pay me for software? Like my mind is blown in this moment. What's happening? Yeah, it was the first year of Amplitude was not like that at all. That took a while to get to that point and finally find someone like Brett.

15:45So it's interesting. One of the things I've often seen with the successful startups is they'll have an insight. So let's say that your insight had something to do with not just understanding growth accounting, but understanding how product interaction impacts retention and the success of the product. That's more than just cohort analysis and churn and upsell and resurrection ARR and all the stuff everybody talks about. But there's a set of people who are prepared to see your new way of seeing, and then there's kind of everybody else. yeah and um you know so i like to say in product market fit we try to ask what can we uniquely offer that people are desperate for and quite often the insight proceeds finding the desperate like you you had a new point of view about what analytics should do but a whole bunch of people are like yeah whatever but there's a few people are like where have you been all my life oh my gosh a thousand bucks that's so cheap and so what did you learn about like apart from this guy who's doing Vegas gambling stuff.

16:52Who were the early desperate customers and why do you think that was? Yeah. So there was a lot of these Zynga product managers, like Brett was one, Siki Chen was another at HeyGen or HeyLabs then. There was another one, Shane, at this company called The Hunt. To us, it was so obvious this was the right way to build. You look at what users do and then you make a better product on the basis of that. And that tells you so much more than you could from any other way of building a product. And so it was just about finding that other group tribe of people that had that same outlook. I think the biggest surprise to me was how everyone would give lip service to data.

17:33They'd say, yeah, data is incredibly important. Of course, we're data driven. Of course, we want analytics and to understand this. But the problem was just too hard to deal with. So they weren't willing to spend money. They weren't willing to spend time trying to figure this out. They weren't willing to use a crappy tool that they'd never heard of that just started where they were the first customer. We just had to keep looking because we knew the future was going to be this way. And eventually we did find our tribe of people that had this crazy belief and so much so that they were willing to bet on us.

18:08One of my mistakes was just spending a lot, too much time with people. If you know Crossing the chasm, there's the innovators and then there's early adopters. And then there's the early majority. I spent a lot of time with like early majority people when really I should have just spent time in the really, really is with the innovators. And if only one out of a thousand companies is the right fit or even one out of 10 ,000, fine. So be it. Just don't listen to those other 9 ,999. Like just folk, just figure out how you can find that one that cares about it. And then eventually, you know, you start going through that curve.

18:44Like, I just couldn't imagine a future, which it wasn't true that people would look at data for their product. And it's still kind of crazy to me today, like, vast majority of companies, they pay lip service to it, but they're not doing it. And so from those early group of ex-Zynga product managers to starting to become a more mainstream thing in technology companies to now we're starting to see traditional enterprises do it. I guess we're 13 years in at this point. We're still going through. We're still in the early days for the adoption of what we're doing. But that's okay. That's totally okay.

19:18Much better to have a group of people, a tiny group of people who like would die for your product than, you know, to be lukewarm to everyone. That's which is, you know, is, you know, it's just like you can you'll you'll fade out. And so did you get a palpable sense ever? Things are really starting to work now. You know, was there was there kind of a defining moment in your mind or was it more just like one day at a time? just trying to figure things out until one day you come up for air and you're like, hey, this is starting to work. So there was about a year and a half at Amplitude where it was, we're lost in the desert and didn't have any paying customers, didn't know how to sell, didn't know how to ask for money.

20:03And so we just build stuff and beg people to use it. And then at the start of 2014, two things changed. One is that we added our third co-founder, Jeffrey Wang, who joined us. And our ability to ship product went way up. I always have this thing where Curtis is twice as good of an engineer as I am. And then Jeffrey is like twice as good as Curtis. So it's like 4X on me. And then I switched full time. I was splitting my time between building the product and selling it. And I switched full time to the selling of it. And I got a sales coach to help me because I didn't know which way was up. And he was like, oh, Spencer, you're wasting tons of time with all these people that are never going to buy.

20:42Let's focus on the few that really care about this product. So that happened at the start of 2014. We grew from zero to a million in ARR in nine months, and we're kind of off to the races, and then raised our Series A, and then just things started compounding and growing like crazy from there. And do you think you had natural sales skills, and you just needed a little bit of coaching to unlock them? Or do you think that your success is, maybe it's a story about how more people can be good at sales than they think? I think a lot more people can be good at sales than they think. I think the impression of salespeople is like, it's a negative derogatory term in a lot of circles.

21:25But you look at the best salespeople today, and they are leading companies, and they are the leaders of the future. You look at, say, Sam Altman, just incredible salesperson. He's a technologist for sure, but his real superpower is as a salesperson. You look at CEOs of any of the Fortune 500, they're all salespeople through and through. So you're selling your vision. You're selling your customers. You're selling people working at the company. You're recruiting more people. You're selling investors. And so the same skill in rallying people towards a vision and leadership is kind of the same. It has a lot of the same elements of skill.

22:00And a lot of people can do it if they put their minds to it. They just need a little bit of coaching. actually, Mike, one of the biggest mistakes I see a lot of great engineers get into is they think they learn it from a book because that's how a lot of them have learned how to be incredible engineers and programmers. They'll be like, well, I learned Python by going through the manual and testing stuff out. And they try to learn sales the same way. You're going to learn by doing it and then getting a little bit of coaching from an expert. That's the best part. There's not some mental model. Yeah, yeah.

22:32There's a lot of fucking mental models. Like there's, you know, every sales, you know, there's sales methodologies, med pick, you know, solution selling, spin selling, you know, an endless thing of these. But they're not how you learn how to be great at it. You learn how to be great by doing it and getting a little bit of coaching from an expert. How did you get to know the guys over at Benchmark and how did they, you know, what did they see in Amplitude, right? Obviously, they're a pretty selective firm, and they don't just write checks for anybody. What do you think was the thing that captivated them?

23:05We had started the Series A process. We had already gotten a bunch of term sheets from other firms. And I didn't even know who Benchmark was, candidly, at the time. I mean, for those in the know, yeah, it's like one of the most legendary venture capital firms of all time. So I finally get an intro to Mitch Lasky at the time. He passes me to Eric Vishria, their newest partner. And at that point, I had already gotten a term sheet, multiple term sheets, and we were like close to deciding on them. So I show up to the benchmark offices, and this is when they're still on Market Street. And I came in 20 minutes late to my meeting with Eric, and I didn't even have my computer, and so I didn't have a deck to go through.

23:51And for whatever reason, I don't know why. I still don't know why this to this day, but Eric really loved what he saw in that meeting with me. Great. Let's get you in front of the partners. I'm like, okay. But hey, this time, just make sure to show up on time and bring a deck and everything else. And in retrospect, I think Eric has a very particular type of entrepreneur he loves. I had a lunch with a bunch of his portfolio companies. A bunch of us were meeting up for lunch a few years back. And it was clear, Here it was like we were all the same person, like Uber nerds that somehow had like started to try to grok the business side and really cared and passionately were competitive and wanted to win at building a successful business.

24:35So, you know, for whatever reason, he, you know, he had just invested in Confluent, you know, who was like fucking, OK, these guys are open source legends and I'm just some like kid at MIT who doesn't know shit. But for whatever reason, he resonated with me in that moment. And so the determination factor, you know, it sounds like he could kind of read into that. And then, you know, you've talked about your early days pivoting and saying, I'm going to outlast this for a few years at least. Is there anything from your childhood or just earlier in your life that would have suggested that you're a highly determined person?

25:11Honestly, I, you know, I think high tolerance for failure is probably the only thing, you know, I didn't mind failing. And then you look at anyone who's achieved anything worthwhile, and the story's the same, which is 50 % of it is persistence, like just showing up and not quitting. And so you're like, okay, all right, well, I can do that. Just show up and not quit. And then I was like, all right, let me think about how I do that to the extreme. I did this exercise where it's like, okay, what's the ultimate founder look like? They'd be willing to go at it for a long period of time. They wouldn't have other commitments in their life, like whether family or debt or something or hobbies that made up their life.