UTS Startups chats with Amplitude CEO Spenser Skates

Hello and welcome to a very exciting fireside chat. I'm your host Victoria Wei from UTS Startups. Joining us is the CEO and co-founder of Amplitude, Spenser Skates.

Very, very excited to be here and share my story, and it's an honor actually to be part of the opening of your guys' new space, so thank you for having me.

Oh, thank you so much for coming and joining us today. Uh, just for anybody who doesn't know, Amplitude is a leading digital analytics platform that empowers businesses to uncover clear insights about customer behavior and take faster action while using trusted data. Amplitude began its journey at Y Combinator and is now a $2 billion NASDAQ listed company. Uh, more than 4,000 customers around the world use Amplitude, including Atlassian, Envato, safety culture, Zip, Walmart, and more. So everybody, thank you so much, Spenser, for joining us. I really appreciate it. Just to kick things off, could you introduce yourself?

Yes, of course. So yeah, so I'm Spenser. I started Amplitude with my two co-founders back in 2012. So I guess that's 13 years at this point. We launched the business in 2014, have grown it quite a bit since then. Today we are a public company, about $300 million a year in revenue. We're still very, very early days. We do digital analytics, as you mentioned, so we help you understand what your users are doing in your app or your website so that you can create a better experience, understand where they're getting stuck, what they like, what they don't, what your best users have in common. And so very, very excited to be here and to share the story.

Yeah, thank you so much. I'm really excited to learn more about your story.

Story.

So I heard that your journey actually began, um, in high school as IT support, and then now, um, you have founded Amplitude. I'm sure that journey has led you through your fair share of pivots, painful funding rounds, a lot of big decisions to make. Can you walk us through the path you took from Sonalight, which is the startup that you had at Y Combinator, to Amplitude and where it is today?

Wow. That's funny that you include the high school bit. I didn't know that they were going to do that too. But I did— for those of you who don't know, I did a— the first thing I ended up doing was tech support. And so I kind of go around the neighborhood and help people fix their computers and stuff like that. And you could make good money as a high school student doing that way back in the day. Anyway, ended up going to college. After graduating, there were a lot of founders in the years ahead of me that had gone on to create big and successful companies. And I was like, "Wow, if they're doing that, maybe I can do that too." So the founders of Dropbox were a few years before me, and then there was a few other companies as well. So I'll tell you a little bit about the founding story. So I had studied startups for about a year. So just kind of really wanted to understand them deeply, what made them successful, and ultimately whether I was capable of building a company. And one of the things that I found was that if you were willing to work at it for at least 2 years, you would likely end up having some success. So every single company I'd read about out there had a period where they probably should have given up if they were rational, maybe a year, maybe a year and a half in, but for whatever reason, and they didn't, and then they went on to be successful. And then even if you didn't have outside success, you would generally find an opportunity that's worthwhile, whether your company being acquired by someone else, whether that leading to a new business, whether developing some skills that you wouldn't have otherwise. And so that was enough to convince me, even as a fresh grad out of college, that I was capable of doing this, even though I had no idea about what it meant to build a business or build a great product or really anything. So I got really, really passionate about startups my last year in college. And I wanted to go start something, and so I asked tons of friends of mine to come and build a company with me. And unfortunately, at the time, a lot of people said they were interested, but very few were willing to take the leap and say, hey, OK, I'm willing to go do this full-time and not work a regular job and not make any money and go through that incredibly risky path. The one exception out of everyone I asked was my now co-founder Curtis, who was willing to come start a company with me, but he said, OK, I'll do it, but I want to go work for a year first. And he ended up working at Google for a year. And I was like, oh, man. And if he goes to Google, he's not going to come back out of that. You get all those nice perks that they give you, and just the lifestyle, and the pay, and everything else. But that was the best thing that I had. So he ended up going off to Google. I did a detour for about 8 months working in finance. And then as we came up to that year mark, I called Curtis. I'm like, hey, we're going to go do this. "Right?" He's like, "Yeah, yeah, yeah." So I ended up— I left my job. I moved out to California and ended up moving into his apartment and living with him in his living room on a mattress for the first bit and just starting to work on stuff together. And up to that point, we had worked on kind of different projects on nights and weekends, but this is the first time it became real. Where they— we would, you know, where it's like, okay, hey, you know, you go from having a support system where you have a job, you have— you're getting paid, you have a boss, you know, you get a lot of feedback on whether you're doing well and you're going in the right direction to, you know, literally nothing overnight. And that was very, very disorienting for me. And so during this time, Curtis was still working at his job at Google. And he would kind of come home at around 5:00 PM, and we'd work on side projects together during the night. And it was actually, frankly, very tough, because most of the time during the day, I was just working completely by myself. And so I was just— I'm not a particularly disciplined person. If you talk to the Amplitude team, they'll let you know that. And so it was very, very hard for me to stay focused, and so I'd like— I feel embarrassed saying this, but I'll share it with you guys. I would play video games all day, and then 4:00 PM would come around, and I'd be like, "Man, Curtis is going to come home in a little bit. I better show him I've actually worked on stuff to progress what the business we're doing together." I'd work on stuff for an hour, and he'd come back, and I'd be like, "Look at how much I got done today. It's so amazing." Then we work on stuff for a few hours, and he'd go to sleep, and I'd just stay up late playing video games. It was really, really bad. Anyway, after a few months of that, to his credit, Curtis actually left his job at Google, and we started working on stuff together. We worked on all sorts of side projects before that. We had worked on a website for photographers. We had worked on a website for alumni from our school so that you could see where people went after they graduated, because they'd go all over the world. We worked on an outsourcing website, lots of stuff that frankly never went anywhere. We ended up taking a step back and just saying, "Okay, what do we think?" We didn't really know anything about what it meant to build a successful business or product, but what actually makes sense to work on together? One of the things that actually interested me was voice recognition, because it was a technology— this was back in 2000. 2011 that was on the cusp of, it seemed like it was not quite ready yet, but maybe over the next few years it would get there and be usable by everyday people. And this was before Siri and before Amazon Alexa and before voice recognition was a standard feature on our phones. So we ended up deciding, okay, yeah, let's work on an app called Sonolight that allows you to send and receive text messages from your phone. So we spent a few months on that, launched it, got— we actually got a few, like 10,000 downloads or something, so it's like, all right, somebody was using it that was not us. And got into Y Combinator off of the back of that. The program was phenomenal. You kind of got— you became peers immediately with 60 different companies. You kind of realize, wow, these people have their stuff together and we don't, and we better work really hard and stay focused on how do we build a really great product and a business around it. So we did that very intensely for about 6 months, or for the 3-month program, culminating in Demo Day. And we actually had this really cool demo at Demo Day where I had my phone and I would I'd put it in my jacket pocket, and then I'd have a conversation with my phone on stage. This was before any voice recognition was really standard, and it looked like magic. We actually got the most press coverage coming out of that demo day out of 60 companies, which was incredible. We got 12 different press articles about it, because it was just a really cool-looking demo. Fast forward a little bit to right after demo day, and one of the things that started to become clear to us was that the quality of the voice recognition was not good enough to get users to keep using it. The retention rates for the application were terrible. At that point, we probably had 300,000, 400,000 downloads. Tens of thousands of users using it on a daily basis. But the users would kind of keep churning out. You'd kind of get a new crop of users, but then they'd leave a few weeks later because the product was not good enough. And so we actually made the really hard decision at that point, even though we were kind of coming off this hype of Demo Day and a lot of coverage and a lot of excitement, that we knew that Okay, hey, there's— we could probably keep doing this, but it's going to be a real grind and we're probably not going to see outsized success. So anyway, so that's like the pre-Amplitude, Sonolight story and what had happened up to that point.

Yeah, really fascinating having to figure out when to call it quits on something that you've been working on and decide to pursue something else. I think your story is very aspirational for a lot of people, a lot of fresh grads, a lot of fledgling founders. Are there any key lessons?

By the way, I'd love to understand the audience really quickly. Like, who here is a current student right now? You have any current students? Okay. Who here is already working full-time on a company? Okay. Sorry. Okay. Who's working in a company currently? Okay, got it. And then who here is, uh, starting a company of their own currently? Okay, so a few folks. We got folks from all different sort of backgrounds.

Yeah, diverse founders and some maybe aspirational entrepreneurs. Um, for those people who are thinking of starting a startup or has already started a startup, are there any key lessons that you've learned from your time that, um, you know, could help them maybe shorten the length between starting something and, you know, mistake, or—

There wasn't, you know, if I were to retrospect, so I think, let me say some of the things that I think we did well during that, the pre-Amplitude days, and some of the things we did poorly. One of the, I think the most important thing we did was being willing to continue going without success. One of my shocks, and so, you know, I had that statistic I quoted earlier where Most companies go through a phase where like a year, a year and a half in, they end up deciding to— they probably should give up, but for whatever reason they don't. They ended up deciding to see it through. That told us, okay, let's at least give this 2 years no matter what. And that was actually really important because when we got to the point where, hey, Songlight's not working out, it was like, okay, That's okay. It's fine. We're going to keep working together. We'll keep the team intact. We had learned a lot in terms of what it took to build a successful product and a successful business. And so there was no question. Whereas a lot of other companies— one of the craziest things to me is looking at a lot of the other Y Combinator companies is that so many of them would get 6 months, 9 months in and then say, hey, I give up. Let me go work a job, or let me go back to grad school. Or let me go do something else. And it's kind of like, this is already after you had been very, very successful. Or sorry, this is already after you had gotten some great validation just from, OK, here are people who are experts at startups that are saying, hey, I think you have the chance at building something reasonable. And so that was kind of the first thing that I'd say that was really important that we got right. The second thing I'd say we got right was we actually got along. Curtis and I got along really well. In that we both were looking to do something big and long-term. And that, you know, we're still working all these 13 years later, 14 years later, still working together. And that is actually, I don't know, I'd like to say I architected that, but frankly, I just got, we got lucky. I remember the first really big disagreement that we had was we were, It was about the user interface, of all things. Normally, you disagree about big deal stuff, like, oh, the strategy of the company, or the culture we're going to have. But it was literally about the user interface in the app that we had this giant— we almost were yelling at each other about it. And this is going to sound so stupid when I say it, but it goes to show how passionate we were about it. So we had a button that you could hit. That would allow the app to listen to your voice in the background while the phone was just sitting there. And we had this whole disagreement about whether the button should reflect the current state of whether it was listening, so it should say on or off, or should it reflect the action that it did when you hit it, which was crazy. And I don't remember which side of it I was arguing at the time. And the funny part is, actually, there's a bunch of UX research these days that you can just go look up, and there's actually best practices on when you should do each one. But I remember we would start to get really intense, and I'm like, wow, Curtis is not backing down on this one. I'm like, what the hell? And then my second thought was like, man, this is going to be really interesting to see if our partnership on this business can survive intense disagreements like this. And to both our credit, we ended up— I can't remember which one we ended up going with, but we tested it and it ended up working out great. But that was one of the core moments. And we were lucky in that we could go through those intense disagreements about the business and survive. So those were two of the things that I think we did well. One of the things that we didn't do well, frankly, on Sonarlight in retrospect, and probably would have spent up the entire thing for us is talk to users. So we were two engineers, and as an engineer, what do you know how to do? You know how to solve problems, right? See a problem, you solve it, and that's what you're trained to do through college and through school and through work. And it's like, okay, you're given a particular problem. But there's also, when you're building a company, a very, very important piece, which is like, what problems are worthwhile to solve? And you have no idea. And the only way to do that is talk to your users or talk to your customers. And so we actually spent probably 90+ percent of our time building and a very, very small amount of our time talking to customers. And so in retrospect, that was a huge mistake. We should have spent half. One of my pieces of advice for companies now is to spend half of your collective team's founding time talking to customers. Incredibly important. And because we didn't, we worked on all sorts of stuff that was a waste of time on the voice recognition app, and we could have gotten a lot further to, OK, let's find the group of customers that are really passionate about this app. Let's understand how they use it. Let's go deep on solving those problems so that they stick around. And then maybe we'd have either been more successful or realized more quickly that we weren't going to build something successful. There's this very famous story of Airbnb during the Y Combinator batch where they would literally travel to New York every week to visit customers in person, because that was how important it was, and that's where their customers were. And they learned all sorts of stuff and ended up being the most successful marketplace in, frankly, a very, very crowded marketplace for listings and hosting people. So anyway, that was probably the biggest mistake that we had made at Sonalight. Wow. OK.

I guess that kind of leads to my next question. When you are in the early stages of building your startup, it's kind of easy to get caught up on different directions and different products that you could create. What's your advice for founders on how they can continue to drive product innovation as they still scale and, you know, balance that? So creating products that are user-friendly and but still be able to grow your company.

Yeah, so what we did— let me say what I did after that, and I'll answer the question. What we did really— so we shut down Sonalight, and then what we did after that was we had spent all this time building our own internal analytics systems because it was obvious to us that's how you should look at— that's how you should understand customers, and we wanted to know Okay, how much did the voice recognition accuracy matter? Or what was it that would help customers retain and keep them around? And so ended up building all this internal data analytics. And what was interesting was when we showed that to other startups in our batch, they're like, "Wow, this is amazing. I'd want to get these same sort of insights." And so that's what ended up leading us to Amplitude. And then in starting Amplitude, I was like, "All right, I don't want to repeat those same mistakes." on the Sonolight side. And so before we build anything, let me go talk to customers. And so we spent a few weeks, and I'd say, OK, look, we've got to talk to 30 companies that could be potential customers of Amplitude before we actually build anything. And so we ended up just talking to a lot of other startups and met other companies that they knew to try to figure out if there was a real market for what we were doing, and that was super freaking important, because if we hadn't done that, we would have spent a lot of time building the totally wrong thing. We added a third co-founder, Jeffrey, during the process as well. Yeah, and then we were kind of off to the races. I switched— one of the things that I then did was I switched full-time from the engineering and building side to the customer side.

Oh, interesting.

And so, you know, it's funny, it was more like, Curtis is twice as good of an engineer than I am, and then Jeffrey is twice as good as Curtis, so just my comparative value for the company was much, much, much lower in building the product. And so it partly was that, but it was also partly, we recognized that from Sonolight. A funny story on the code side, I remember there was one month where I was very determined to prove that I was kind of close to Curtis as an equal and an engineer.. And so I just focused really hard on, like, all right, look, I'm going to, like, ship lots of code and build lots of stuff and, like, work really hard and focus in. And we get to the end of that month and we run this. There's this, like, script that you can run to understand, like, the different stats of your GitHub repo. And so we under— so it looked at, like, lines of code shipped and, you know, percentage of codebase touched and, you know, number of commits and all this stuff. And on, like, every single metric, I had just like half of Curtis's score. It was like, oh, he had shipped 2,000 lines of code. I only shipped 1,000. And he had like 100 commits. I only had 50. And I'm just like, you know what? I give up at this point. It's OK. I'm not as good as he is as a programmer, and that's OK. So I ended up switching that time full-time to the customer side and just spending time doing support for customers, selling to customers, meeting prospective their prospective customers, and that was huge. So we grew from zero to, you know, and that felt very, very different from the years before, where all of a sudden, instead of, you know, having struggling to retain customers and, and not really knowing whether you're building the right thing, we grew from zero to a million in ARR in about 9 months. And that, and then it was just, um, you know, kind of exponential growth from there. And we grew from 1 to 4.5 second year and 4.5 to 13 in the third year and just keep compounding and fast forward today. And then it was just how can I keep up with that as much as possible so that I can start to hire a sales team and start to hire management and executives and you get into the real company scaling. But looking back, it was a real almost almost a little over 2 years that we were just not— Sonolight and then failed, and then Amplitude before getting real traction on the company. Sorry, I don't know if— Did I answer your question at all? I can't even figure it out. We don't even have to go by this. I want to be useful and interesting to you guys.

What I'm hearing is that you're balancing building with intention and speaking to your users versus scaling, which I'm sure is a very difficult thing to do. You kind of talked about taking, you know, a lot of the user insights. How can founders think about data and the type of data they should get to validate what they're doing and to help make them decisions? But also, you know, thinking about how the data leads, you know, the product development, go-to-market, all of those things.

But my biggest thing actually is just so, so Data, I consider kind of the second stage of maturity in this. The first stage is just talking to your customers. You have to do that. And as a founder, spending half of your collective team's time doing that is incredibly important, especially if you're engineers and especially if it feels uncomfortable. You don't have to be selling them or pitching them or anything, but just talking to them is incredibly valuable. And that will give you so much insight into what their problems are, what they care about, What to solve, and are you building worthwhile? It's kind of like you're navigating this very complex space, and you could do a lot. You can go in 100 different directions, and how do you know which to go in? Well, it's like the only way is if you're continually talking to customers to try to understand. I would say once you're doing that, and if you have a really, really good motion around that, that's when you add data on top of that.

And what were you looking for when you were talking to people to be able to actually validate, is this a direction that we should go into?

So there's a few different ways. One of the questions I actually like to ask, particularly when you're doing kind of when you're really early on and doing market development, is how painful is this on a scale of 1 to 10? And so if it's 5 or less, then it's like, "Okay, clearly this is..." They might talk to you because they're polite, but it's not a big problem. Whereas if something is an 8, 9, or 10, it's like, "That's a really good signal. Okay, then let's spend time there." That means something is painful enough that they're willing to go through a bunch of work and to learn your crappy little product that has no users, brand and no, it's very early and has a lot of issues with it. They're willing to overcome that barrier in order to solve their pain because it's an 8, 9, or 10 in terms of pain. That's a really good one. Another one, if you're in B2B, is really digging for business pain. Business pain comes in 3 flavors. It's either you're increasing revenue, reducing costs, or reducing risk. You should be able to map whatever it is you're doing back to one of those three things. If you're not, if you can't do that, then you need to keep asking, "Okay, well, why is this important? Why do you need to do this?" I actually just had this conversation with a prospective customer yesterday. We were talking to the Australian Football League in Melbourne.. And they were asking us for help on this because they're like, OK, yeah, if I get more users, but so what? And it's like, OK, well, look, let's work together to make the business case to say, OK, here's why more people downloading your app leads to more revenue overall for your business and a larger, stronger business. And so that applies no matter what level and what scale you're at. One of the other traps that I think I fell into in the early days was just not really appreciating how deep you have to go on that pain. And so I'd say, okay, I had this coach and sales mentor that I'd get help from, and he would always ask me, hey, what's the pain of this customer? I'd be like, well, they want some dashboards, they want some charts, they want to use analytics in this way. And he's like, you know, Spenser, that's not a business pain. Go find me a business pain. And so I'd go back and he'd beat this into me every week. And so I would finally remember to show up in the meeting and be like, okay, let me keep asking, keep digging for pain. Because it turns out, if someone's in pain, they love talking about that pain. Like, you know, they love talking about, you know, sharing, sharing whatever frustrations and pain they have. And, you know, if there's even a chance you can help and solve it, that's great. So, you know, really, really focusing in on, on that.

That is, uh, is key. And was there any times within your business journey, um, creating Amplitude that you thought maybe, um, you were close to maybe calling it quits?

Oh, so many times. I mean, I, I just— well, I, I— there's— I think there's every 3 years there got to be some point where it was so painful and I'm like, man, I don't know if I can do this sustainably. And so I guess, you know, I'm on year 13 now, so that's 4 rounds. And so probably, I don't know, I'm due for it next year, I think.

Now the existential 3-year mark.

Yeah. So, so the first time was actually, um, trying to, to fund— believe it or not, it was actually talking to users and having them not care. That was like, you know, I knew that was part of the startup journey, that was fine. Like, so I was like, you kind of get immune to that after a while. But I remember the The first time that actually was quite painful was when we were raising our seed round for Amplitude. We didn't really have anything at that point in terms of customer traction. We were just pitching these VCs and we get a bunch of nos, and then even worse, a lot of maybes that turned into nos after some time. That process took around 8 or 9 months end to end, and I remember just It was brutal, because I was like, man, this is so obvious that this will be a really big business, because you're at the intersect. There's a bunch of different trends. Mobile was just taking off then, and there was a new— I knew that there was a new data analytics infrastructure, because Google Analytics and Adobe were too outdated. And I was telling the story, but it was just not— I don't know. I was not very good at pitching the story. I think it's just like it took a while. So it ended up being about 9 months end to end to raise— that we were trying to raise $1 million. And yeah, that was the first time where I was like, man, this is really, really tough and I don't want to do it. The second time was actually, I remember I had a very, very toxic employee, and that was rough because it was the first time I realized, okay, I have someone who's not a fit for the team and who doesn't agree with the mission we're going on and the direction we're going and all of that. And I realized I had to make a call to, one, to hold them accountable on that, and then two, to eventually, when they didn't want to go in the direction that I wanted, move them out of the business. And so I remember many late nights where I'd like kind of dread going into work the next day because I'd have to have conversations with this person and like hold them accountable for something I knew they didn't really want to do. And so, and it was, and that was the first time. Yeah, that was kind of the second time. And then next time after that, I had to like turn over the executive team and the business had gone through a slowdown and that was tough. And then I recently did the same thing where I turned over these executive team again a few years ago. And I look at— and that also had its difficulties. But such is the nature of these things. And part of it for me was knowing that, okay, if you just continually stick at it for long enough, what you will do will— markets are really big and the time you have is long. And so the compounding, if you continue to do the right things day in and day out, the compounding in whatever market you're in will eventually take care of itself.

Yeah, and if you just kind of keep on persevering, as you said, you've had so much experience where you've been through a tough period and then been able to get through it. Probably that shows you validation that those—

and eventually— one of the psychology things that's really interesting is that as humans, we're wired for loss aversion. So, if you were to lose lose, let's say you were to lose $100, that is twice as psychologically devastating as gaining $100. And so what that means is, let's say you were just to like fluctuate around a point, like, oh, today you did well, tomorrow you did poorly, today, you know, next day you did well, you know, next day after that you did bad. The losses are twice as psychologically impactful as the good parts. And so what that means is like, if you kind of zoom in on any particular day, you're going to have something go bad. You lose a key customer, your product breaks, you're not able to close a key deal, someone important to the team leaves, and it feels bad. It feels, frankly, really, really bad. I actually had an employee earlier today that I saw this morning where it's like they had sent me a note being like, hey, I'm leaving Amplitude after doing work here for 3 years. And I was like, oh my God, this is brutal. But you recognize that in part of it, part of the skill in the persevering is just kind of taking that pain as it comes. And then eventually the goods outweigh the bad, even though psychologically they feel a lot worse for you in the day to day. And so I'm not telling you you guys anything new. You guys probably heard this, you know, from anyone who's done anything worthwhile, that, that, um, you know, the, the grit and perseverance is, is like 50% of, of the predictor of success. But, um, I think just recognizing that as humans, that the loss and the bad is like really programmed to hurt us a lot more than, than the upside, is important because that helps you manage your own psychology.

Yeah, absolutely. Like, day to day you might feel like you're in the trenches, but when you zoom out a bit, you see how much you've climbed. And I think that's really important for any founder, anybody trying to start something new, to recognize, um, because yeah, we do have a negative negativity bias, so all of those do feel much worse. Um, a little bit of a change in, um, direction. One of the biggest conversations where most companies and, um, that are happening right— having right now is about AI. Yeah, I did want to get your thoughts on how do you you see the role of AI for startups, and what are some ways that startups could be leveraging AI?

Okay, so this is where, for anyone who's a founder or with a small company, you have an incredible opportunity now to kick larger companies' ass, including Amplitude. Like, we're 700 people now, and we are so incredibly slow-moving. It drives me up the wall. We try to be fast, we try to be agile, but it's It's so hard to do it. I think because you have the ability to automate a whole bunch of pieces of knowledge work, you can actually build an organization and a process very, very differently than if you're large. Okay? Because Amplitude is a large company, there's certain constraints. It's like, hey, I got to deliver the quarterly number to that we've planned around, hey, we have these initiatives that are longer running. Hey, we have people in these specific roles that we've hired and you can't just all of a sudden totally change them on them the next day. And so we have very real constraints. And so the way we're doing it is like, okay, we're having individuals and teams who see the potential from an AI standpoint to totally change how they're operating. Now, if you're a small company or a founder, you don't have any of those constraints at all. It's actually freaking phenomenal. And so you are able to build and scale the different ways of operating your business in a fundamentally different pattern. And so that's actually, from my standpoint, if you're an early-stage founder, a huge, huge opportunity. I'm trying to give you a very, very simple example. The lead generation process in Amplitude is really screwed up, and we have a lot of work to do to get it to be great. Where it's like right now, the way it works, it's like if you come to our website, you fill out a form, and you get a notification back saying, "Okay, cool. Thanks for filling out the form. We'll contact you." That's crazy because you should be able to, if you want to talk to someone at Amplitude to get you should be able to book an appointment right on the website and like go do it like right now. But you can't do it because it's like, hey, we've created this particular pipeline so that all of our sellers can like take advantage of that, and it's like a whole bunch of work to change it. Whereas if you're a new company starting out, you don't have any of those constraints at all. You can say, hey, what is the best customers, what is the best experience for my customers to have? And if they want to talk to someone, they can book an appointment right away and just go and and have a conversation with us directly. And so that's kind of just a really small example of incredible opportunities you have. Now, you may not have as many resources as a larger company as well, but that kind of doesn't matter. I think it's the technology has the opportunity to be very disruptive. The second thing I'd say about it is don't be intimidated by it. It's actually the technical barrier to entry on AI has been going down very rapidly. So it used to be like if you look back 20 years, you had to have a specialized AI team that understood the different algorithms for creating and were capable of creating their own custom models and everything else. And then a few years ago, it used to be, okay, hey, you're going to use these foundational models from companies like OpenAI, Anthropic, and then you tweak them a little bit with fine-tuning and other methods. Now it's like fine-tuning actually doesn't even matter that much, and it's like prompting. So just literally feeding these things the right text, that's how you get the great results. And anyone in this room is capable of doing that. And so I think now that the technical barrier to entry has basically gone to almost none, nothing, it's actually much, much easier to create products and create great experiences for customers in leveraging AI. The other thing is that if you look at any larger company, it's just taking them so much longer. A lot of them will have a chatbot, but it's like if you use the average B2B, like I look at MySpace and data analytics, frankly, no one's done anything innovative. The best people, The typical thing people have done is like, oh, okay, we can generate SQL queries for you. But to be able to actually use natural language and use that to interface, and if you look at kind of the average B2B company, what they'll do is they'll have a chatbot and it'll be like, you know, it'll like be connected to their help center and I'll feed you back stuff from their help center, but it won't actually do anything in the product for you. And so, you know, if you're innovating, say, in that space, like you're going to be able to go much, much faster than any of of the existing incumbents. So I think it's absolutely a tremendous opportunity for, if you're founding a company or you're an early-stage company, to both build a better customer experience as well as to build a much more efficient and effective organization. And frankly, I think a lot of the way— you have the opportunity to invent how every future organization can operate. So I think, yeah. I think it's absolutely incredible.

It's a really fascinating time right now. Is there any way that Amplitude is using AI that you'd like to share, or any interesting ways that you guys have been integrating it?

There's a few different places. One is we have an Ask Amplitude chatbot that turns your chats into charts and funnels and retention charts and a whole bunch else. We've had that for about 2 years now. We're just coming out this year with an AI agent that will automatically uncover trends in the data for you. One of the really challenging parts about product data is it's very large and complex. The average product has thousands of touchpoints and you have millions of users doing all sorts of things. No human can look at all of those data points and see what's going on. We're excited to automate the process of, OK, well, here are interesting trends in your data, and here are interesting insights out of it for you automatically. The last one— so that's quite big. The last one I'll share is actually we just developed this— we're doing a lot internally. We just developed this bot called Moda that pulls from a lot of different sources of data within Amplitude. So it looks at transcripts of Zoom calls. It looks for Salesforce notes. It looks at product board requests. It looks at Jira and Atlassian, and it summarizes anything you want from that information. So what customers, say, want a particular feature, or what was the context of the last conversation. And so that's actually pretty cool to see. The usage internally has taken off in the last few weeks. So anyway, that's a long way of saying I think it's still very, very early in terms of seeing how it's going to be deployed, particularly at larger companies. But, but I'm excited. I'm very, very excited by it.

Yeah, I'm excited to hear more about that. Um, and we are at a university. I did want to ask you before we wrap up, um, the formal questions, if you were to go back, um, to yourself maybe at the age that you went to university, or even when you were, you know, working doing support in high schools, what's some advice that you would give to your younger self, or any words?

I mean, I think the one of, one of the things that it took me a while to appreciate was that you did not have to go on any prescribed path that people set out for you, and you could be any, you could be anyone. And what I mean by that is like If you were willing to make the life choices and the trade-offs, you could be a founder, or you could be an artist, or you could be very, very successful in the corporate world, or you could go travel. You could be any character in any novel or any film if you were willing to make the right life trade-offs. I think where I saw people people get into trouble is they try to do kind of everything all at once. But if you're willing to make the choices, you can choose to be everything. You can't just be it all at once. And so that would be— it took me a while to really appreciate that. I remember I saw this talk by Brian Chesky, the Airbnb founder, and he said this thing, which is like, one of the things that convinced him he could go start a company was he saw a founder, and literally the only difference between that founder and him was that they had made the choice to build a company and kind of do the hard work associated with that, and he hadn't, and that was it. And so that would be— yeah, I'd definitely say that to myself, because I think you spend— in the early days, frankly, you spend a lot of your life being told what to do. By adults, and you have a very specific set of rails you have to go on, and a very specific set of things you have to learn, and places you have to be, and how you spend your time. But actually, once you get past that stage of life, it's like your freedom is actually incredible in terms of what you have the opportunity to do.

Amazing. Thank you so much, Spenser, for joining us and answering those questions. What I'll do now is we'll move to the Q&A portion of this session. But before we do that, could we all have a round of applause for Spenser Skates? Thank you so much. So I just have Alex walking around with a mic.

Perfect.

If you have any questions for Spenser, Please feel free to raise your hands and Alex will come over.

Thanks very much. Thanks very much, Spenser. Your point around bigger companies versus, you know, or startups making a move to deep learning AI stuff, I guess kind of resonated with me. So I've got a tech startup from Sydney, one of the founders on the panel there, happy, happy few customers. I wanted to understand if you have seen Amplitude, any ways that your team are kind of getting around some of those things where it can be slow moving to get going on AI, where, you know, risk— risk tolerance towards trying things with AI? You mean internally within Amplitude? Yeah, well, yeah, what you're hearing at Amplitude or, or elsewhere? Um, I, I think you have I just wanna make sure I understand your question. You're saying, you know, how have I thought about driving change so that we actually are able to leverage AI with an amplitude? Yeah, it's hard. I mean, first thing I'll say is it's harder as a larger company. I think there's a lot of thinking of like, okay, well, so first there's, there's just a, OK, here's how our current process works, and a lot of people are dependent on that, rightly so. And so breaking that and changing that, you've got to coordinate a lot of times with a large group of people. So sometimes that might be 10, sometimes that might be hundreds. And so the question is, how do you do that successfully? So one of the things I think I've been focusing on within amplitude is just getting individuals or small teams just to go and do things and kind of show what's possible. Like this Motobot I was talking about that pulls from a lot of different data sources, luckily it wasn't dependent on anything internally. You could just create it and a lot of people would start using it. And what's happened since, a lot of people have seen other people using it and they're like, oh, okay, I can actually ask it questions and leverage it that way. It's interesting, it's actually, it's much easier to create something new that hasn't existed before and leverage that with AI than kind of change an internal process that a lot of people are dependent on. And so that's kind of one example on it. Another is like, I've basically just, I'm constantly pushing executives at Amplitude on this to be like, hey, how can we go faster? How can we get things approved faster? Like there is one, thing that happened. There's one use of AI internally within Amplitude where it's like, hey, in order to get permission for this, we've got to get approval from IT and legal and security and I don't know, some one other function. It was like 4, and it's like, okay, let's meet with those 4 functions. And I'm like, guys, this is like way, way too long. Okay, I need one person who can approve all of this. And then like, look, if it needs to be reviewed all that by those groups later, fine. But like one person should be able to approve the access and usage of AI within the company, because we got to be able to go and we cannot block. We cannot wait for, you know, 4 different VPs to have a meeting in order to approve every single new use case, because if a startup was doing this, they wouldn't even— like there'd be zero people that would need to approve it. I'm constantly trying to push the level of speed in terms of adoption internally. Frankly, it's hard. There's not any magic to it. I think we have a $300 million in revenue business and 700 people that are built around that, and it's like that has a momentum of its own, and rightfully so. You want to keep it going. You want to keep that working. Compounding that. And so I'm thinking about, okay, I'm always kind of balancing that tension of how aggressive can we go without disrupting that. I also will say we're having a good set of executives in place who understand this change and are embracing it and are leaning forward and understand how to translate that effectively at scale. But there's no good answer to it. I'll give you another example that drove me up the wall recently, which is we were trying to— we're setting up an AI week where the goal is to train all of the product development organization on AI tooling. So like first principles, how's the technology work? What's it capable of? How do you use the models? Different use cases like coding, product requirements, docs, things like that. And to get a week that a majority of the company and the executives could be there. So I wanted to do, I think we started talking about this in, I wanna say March or so, and it was like June till we can get it scheduled. And I was like, what the heck, guys? Like, I really wanted it in April. And then like we chose a date and then turns out the designers were out for some conference that week, and so it had to get pushed back and it got, you know, and then the following week I was out and it was like, Okay, we're not doing it till June. And it was like, ah, it's driving me nuts. And so we've been trying to do things internally to train the team on AI beforehand, but this is one of the times where one of my executives kind of pulled me back and my head of engineering said, look, okay, I want to do this right, and so let's get, we need to get the different key leaders within the team excited about this and starting to do these use cases, and then that will maximize the success of the week. So anyway, I don't know if that answered your question, but hopefully that's helpful.

Thank you, Spenser. My name is Lorenzo, and today the B2B space and the SaaS visually, I can see that the lines between platforms are getting blurrier. So like you have Amplitude for analytics, but you have the marketing automation tools now doing some analytics, and you have AppsFlyer, a marketing measurement platform, doing analytics. Yes. And it's becoming tougher and tougher. So how do you believe that Amplitude is able to keep on differentiating itself and being the top of the market?

Yeah, there's no— it's a very crowded space. And it was very crowded when we started, and it's even more crowded today. So there's no magic to that other than— the thing I've thought a lot about is that it's less about building some killer feature or some silver bullet that solves everything, and it's more about building an engine that outbuilds the rest of the market. And so what I've thought a lot about is, that loop I talked, that thing that you're constantly talking to customers, you're constantly getting feedback from them on what they want and what they value and what they're looking to you to build. And so one of the things I'm paranoid about, particularly as we scale to a larger organization, is how do you do that effectively? Kind of has as tight of a loop throughout the organization. And so one of the things, as an example, is like we have product managers and engineers. It's a big deal to constantly talk to customers and so that they're seeing kind of firsthand what our customers are asking for and needing or wanting, and then taking that and reflecting that back in the product. And so, you know, and I'm doing the same thing, you know, when I'm out here in Sydney, you know, I talk to, you know, a whole bunch of, you know, I talk to safety culture this this morning, they had a whole bunch of asks on how do we make the instrumentation and taxonomy better and cleaner. And it's just that kind of day in, day out over a very, very long period of time. And if you do that over a long period of time, you're going to keep compounding all the set of capabilities that you have, and that is going to outstrip kind of anyone else that is kind of competing with you in the space. For example, digital analytics, it's very much the case like every single— there's a million companies out there that do the basics very well. And every actually large tech company like Google obviously has Google Analytics, Amazon has their analytics, Microsoft has Microsoft Clarity, Apple has their own version of analytics, Facebook has had analytics at various points in time, and so you can kind of go down the list. List, and a lot of companies would have something basic. And so us recognizing that our value was kind of going deeper and further and better than anyone else, and just outbuilding the competition, and having that mentality, particularly as we continue to scale. So yeah, I don't know that there wasn't really any magic to it in terms of saying, okay, hey, we focused on this strategy and nobody else does, but it was more about, all right, how do we We create a bigger— making sure we're more responsive and more aggressive in terms of getting stuff that our customers want to them more than others. Hi, Spencer.

So you emphasize the importance of talking to customers.

Yes.

Quote by Henry Ford, which says, if you asked a man in a wire, they would have said, a faster horse. So how do you balance listening to customers but also being innovative?

So I think that's a great— that's a fantastic question. The key is not to just blindly take what your customers want and go build build that, but instead to use those data points to understand that pain that I talked about earlier. And then from there, you will build lots of stuff. You're going to build stuff that directly solves it and is directly what they asked for. You might build stuff that solves it in a different way that nobody expected. There's a lot of ways to be very innovative on that. The biggest mistake, though, that I see product teams, both early-stage founders and, frankly, product teams at large companies, companies do is they don't spend enough time on this. If you're a product manager at a large company, you should be spending half your time talking to customers. That's what I tell the PMs at Amplitude, and that's what I expect generally. And I'll talk to product managers all the time and be like, "How often do you talk to customers?" It's like, "Once a week." I'm like, "What are you doing?" That is a gold mine of— Understanding that deeply is what you should be doing. Your job as a product leader is not to come up with requirements, docs, and specs, and run a process, and all that. I don't care about that. Those are just things. What you should be having is a very, very deep understanding of your customer and the pain. And if you do that, you will build something much, much more successful than those who don't. I think, frankly, it's easy not to do that. It's very, very easy not to do that. It's very easy to build. Very easy to focus on stuff, internal processes and systems and things in front of you. So that's why I'm religious about it. It's not so much to say, okay, you should just build exactly what they tell you, although that's not a bad strategy. And frankly, I think a lot of companies would benefit more from just doing that. But that information is so valuable to understand out of the thousands of things you could do, which one is the the most important. So it's just that, like, you're not going to go— like, honestly, a lot more product leaders would be better served if they literally just kind of spent way more time doing that and building exactly what their customers asked them for.

SaaS B2B management third-party contracts. So I work with different types of big companies such as DHL, Universal, and stuff like that. It always comes like this difficulty where you have so many stakeholders, so many users asking for a lot of things, customization, stuff like that. And how do MFT can help in the future? Like companies, besides enterprises, have more tailored experiences as what they need due to data-driven analysis, and much less, I think, of what they think they want, but actually how they use the product. And if you think some kind of solution with AI for, for this kind of departments of Less of how the financial predictions, how the web works, but how do you connect with the behavior of the user and do you actually tailor-made product solution for them?

Yeah, so it is actually very interesting in that I think— so the technology is getting better very, very quickly and having a deep It's not obvious what it's capable of and what it's not. There are some tasks where it's really bad at. I think, for example, one thing I've actually found it to be not particularly good is inspecting quality. It's not a good judge of quality in general, and so that's something that I find humans to be much, at least for now, to be much stronger at. And so this is actually one of those places where I think the key part part is, so one, spending a lot of time talking to customers, but then two, understanding the technology, because it's not just a question. I think in most, so in most B2B SaaS, what you're doing is like, it's actually a very straightforward formula. You talk to the customers, you ask them what they want, you build that for them, you get to sell to more customers, you ask them what they want, and you kind of repeat the loop again and again and again. Whereas I think AI is actually quite different because I think most or almost all customers do not understand the limits of the capability. And so this is where becoming an expert in the technology and what it's newly capable of, that is actually really important. And a lot of times you'll come out with stuff that customers didn't even know they wanted or expected. And so I'm trying to think of some good examples. Examples of that. Like, I'll actually— let me pick out Cursor. I think has done an incredible job on that. I think one of the things they actually did very, very well is— by the way, are folks here familiar with Cursor? Who's a Cursor user? Okay. Okay. Good amount of the crowd, right? One of the things they actually did very, very well in it is like, one, so they bootstrapped off of VS Code, an existing IDE. And then two, they came up with a bunch of places where the functionality was very tailored to what AI was good at. So for example, autocomplete. Now I'm almost certain, it's not like they interviewed lots of coders and asked them, hey, do you want autocomplete that looks like this? No, they just built it because they understood that technology and what would be useful. And that's a very different approach than say, like a lot of the ones that try to do, generate a whole program in one thing. Iterating over the codebase. That was, I think, a good example of deeply understanding, okay, what makes things more useful as a coder with, okay, what's the technology and AI capable of? Actually doing really, really good autocomplete that gets the right context from other documents and the rest of the file and then makes reasonable size suggestions and makes it easy to change and edit those. And cascade those and look at those. That's where I think they did a really good job of both understanding what developers want to do with what is the technology capable of. Whereas I think if you see a lot of B2B companies that are trying to do these coding agents, they actually kind of hit the problem at the wrong level of abstraction. So they'll try to create an entire thing, entire code base from scratch and Often that actually won't be the right level of granularity. So I think AI is actually one of the places where being very familiar with the technology allows you to do something that customers never even think of.

We might have space for one more question.

To you. All right, go ahead. We'll keep going. We'll keep going. So as a founder, my fear is not failure, it's success. So I'd be interested to know your insights on evolving from startup to scale up or scale up to enterprise. Yeah, very happy to talk about that. That's a whole conversation in itself. The hardest transition actually was from founder to large company executive. Most founders do not make that transition successfully, and here's why. As a founder, your job is to lead from the front, so you're always diving in on the hardest problems. If there's a really hard product problem or a really tough customer or like a marketing launch, it's like your job is to lead and go do the work. As a large company executive, there's too many things to do, and so you cannot go do anything yourself. Your job instead is to review the work of others and to recruit people who can do it. And that's a very, very difficult mental switch. And most, even very successful successful founders I've seen. I remember I tracked a bunch of SaaS companies that were a little bit ahead of me, and even the successful companies that grew to $100 million plus in revenue, those founders did not successfully make the shift, and they ended up stepping down and leaving it to others. And so that was actually a really— you had to unlearn a lot of what made you good as a founder and learn what makes for a good large company executive. Now, I I think in retrospect, I think it's actually easier to be a large company executive. You have a bunch of resources, and you have leverage, and you're able to affect change on— and stuff takes longer, but you're able to affect change on longer timelines versus having to drive everything yourself. But it's a very hard mental switch. That took me about 3 years to really appreciate and understand. So yeah, you're right to be scared of— of success in the sense that it's great, you've achieved your dreams, you're reaching broader and bigger impact, you're generating revenue, you've built a successful business around what you're doing, but you have to rethink who you are and kind of reinvent who you are and actually change a lot of what made you great as a founder. And it absolutely is a kind of key inflection point for a lot of people. Can we do one more? No, we got to cut it off now. All right. One more question. I like that question. Mine is relatively simple to that, but I was thinking about more around what do you have to consider when it comes to building winning teams? What are the key considerations? That's so hard. That's so hard to say. I was very, very lucky, frankly, with my co-founders. I'll start off with that, in that I had a group of folks that was kind of similar mentality of what does a successful business look like and what's the process and the inputs to to get you there. And so one of the things I'd say on the co-founder front is I think a lot of people partner and get married too quickly on that in the sense of getting that right. I had to wait a year for Curtis, for example, but that was very worth it. I look 13 years later, and so getting that right is paramount. Because most startups die from conflicts between co-founders. Not— yeah, you might hit bumps, and you might have to restart and change directions, but that's just part of the game. Whereas founders splitting up, that's deadly. And so one of the surprises to me, actually, during Y Combinator was about a third of companies during the batch had major issues between the co-founders. Which was crazy to me because it's like, wow, you made it. It's like you kind of have this platform and you have the shot. Just shut up and make it work. But somehow they would have major issues between co-founders during the batches. And so getting that right, I think, is very, very underrated. So finding people who have a similar value set and have a similar goal of where you're trying to go. You can disagree about the strategy and disagree on how to get there, but it's similar. Hey, this is what we're trying to do. That's very, very important. The other thing I'd say, and then it changes. Then you start building up the early team, and then frankly, you're just kind of begging anyone to get in to the boat and trying to glue everyone together. And you're like a bunch of misfits trying to take on this really ambitious mission. And then you get to large scale where you're trying to get executives, and that's a whole different thing in itself. And on that one, the one thing I'd say is that there's this spectrum of what's called, I'd say, missionary versus mercenary. There are folks that are in it for the mission and are trying to build particular things, and there are folks that are more mercenary. And there's not like everyone's not one or another. It's more like you have kind of elements of both. Everyone needs to earn to make a living, and everyone needs to, but also everyone cares about the particular mission you're on. And so making sure you, one of the big things that I would have advised myself in retrospect is focus a lot on that in the hiring process and executives. A lot of the skill sets, a lot of the abilities you can learn and you can change, but the why you're in it, that kind of runs deep and that sets the tone for the rest of the organization. So thinking about getting, okay, hey, is this someone I would work for? Is this someone that can also recruit great talent that's also in it for the right reasons is incredibly important.

So cool. Cool. Amazing. Thank you so much.

You do have— Keep it up with Spenser.