This Week in Startups - Anatomy of a Win-Win Acquisition: Synergy, Growth, AI, and the Future of SaaS | E2047

Episode Date: November 19, 2024

Todays show:Spenser Skates and James Evans join Alex to discuss Amplitude's acquisition of Command AI, shedding light on recent startup M&A trends and the strategic vision behind this move. The conversation delves into Command AI's user assistance technology and its integration into Amplitude, which aims to enhance user surveys and feedback. Spenser outlines Amplitude's product-led growth strategy and its approach to scaling through acquisitions while managing risks and ensuring cultural alignment. The discussion also covers the acquisition deal structure, regulatory concerns, and the decision-making process between raising funds and selling. Also, we hear Spenser’s unfiltered take on private equity and regulatory oversight!*Timestamps:(0:00) Spenser Skates of Amplitude & James Evans of CommandAI join Alex.(1:40) Startup M&A trends and Amplitude's acquisition of Command AI(4:30) Overview of Command AI's user assistance technology(7:05) Integration and benefits of Command AI within Amplitude(10:17) Squarespace - Use offer code TWIST to save 10% off your first purchase of a website or domain at https://www.Squarespace.com/TWIST(11:45) Enhancements in user surveys and feedback with Command AI(16:58) Amplitude's strategic vision behind acquiring Command AI(20:22) Gusto - Get three months free when you run your first payroll at http://gusto.com/twist(21:28) Product-led growth and enhancing user experience at Amplitude(26:08) Amplitude's approach to scaling through strategic acquisitions(30:01) DevSquad - Get an entire product team for the cost of one US developer plus 10% off at http://devsquad.com/twist(31:30) Managing acquisition risks, integration, and cultural alignment(35:19) Acquisition discussions and deal negotiation(39:48) Acquisition deal structure and stakeholder conviction(42:08) Addressing regulatory concerns and valuation mismatches in acquisitions(44:20) Decision-making: Raising funds vs. selling a company(46:00) Post-acquisition: Integrating Command AI's technology(49:37) Amplitude's future growth strategy and leadership(54:00) Transition from founder mode to distributed leadership(56:15) Role of AI in Amplitude's product development(59:44) AI model considerations and financial sustainability in B2B software(1:01:34) Market position and growth of Command AI*Subscribe to the TWiST500 newsletter: https://ticker.thisweekinstartups.comCheck out the TWIST500: https://www.twist500.comSubscribe to This Week in Startups on Apple: https://rb.gy/v19fcp*Mentioned on the show:Check out Amplitude: https://amplitude.comCheck out CommandAI: https://www.command.ai*Follow Spenser:X: https://x.com/spenserskatesLinkedIn: https://www.linkedin.com/in/spenserskates*Follow James:X: https://x.com/dazzeloidLinkedIn: https://www.linkedin.com/in/james-evans-7086b3126/*Follow Alex:X: https://x.com/alexLinkedIn: ⁠https://www.linkedin.com/in/alexwilhelm*Thank you to our partners:(10:17) Squarespace - Use offer code TWIST to save 10% off your first purchase of a website or domain at https://www.Squarespace.com/TWIST(20:22) Gusto - Get three months free when you run your first payroll at http://gusto.com/twist*Gusto pricing shown in ad is based on pricing prior to March 2025(30:01) DevSquad - Get an entire product team for the cost of one US developer plus 10% off at http://devsquad.com/twist*Great TWIST interviews: Will Guidara,Eoghan McCabe, Steve Huffman, Brian Chesky, Bob Moesta,Aaron Levie, Sophia Amoruso, Reid Hoffman, Frank Slootman, Billy McFarland*Check out Jason’s suite of newsletters: https://substack.com/@calacanis*Follow TWiST:Twitter: https://twitter.com/TWiStartupsYouTube: https://www.youtube.com/thisweekinInstagram: https://www.instagram.com/thisweekinstartupsTikTok: https://www.tiktok.com/@thisweekinstartupsSubstack: https://twistartups.substack.com*Subscribe to the Founder University Podcast: https://www.youtube.com/@founderuniversity1916

Dude, sorry, I was going to rant on private equity for a second. Let's do it. They make tremendous amounts of money, you know, one of the great players in modern capitalism. But one of my rules is never be on the other side of a transaction for private equity because they have a spreadsheet about how to take all of your surplus away during that transaction. Dude, if you want to talk about where the FTC should regulate, fucking regulate the hell out of private equity. Like, that would be my first order thing.

because like if you look at their playbooks, most of their playbooks are, hey, consolidate market share to create monopolies or pseudomonoplies and raise prices on consumers. And if you look at U.S. anti-trust legislation, the test of are you behaving anti-competitively is do prices go up for consumers or down for consumers? And like the private equity playbook 101. So Lena Kottnick, if you're listening to this or whoever the next person the Trump administration is, fucking take a hammer to private equity. I'm all for it. This week in startups is brought to you by Squarespace. Turn your idea into a new website. Go to Squarespace.com slash Twist for a free trial. When you're ready to launch, use offer code Twist to save 10% off your first purchase of a website or domain.

Gusto. Gusto is easy online payroll, benefits, and HR built for modern small businesses. Get three months free when you run your first payroll at gusto.com slash twist. And Dev Squad. Most dev agencies only offer developers. Why? Because product management is hard. Get an entire product team for the cost of one US developer plus 10% off at devsquad.com slash twist.

Hey everyone, it's Alex. Welcome back to this week in startups. Now, on the show quite a lot in the last couple of months and quarters, we have talked about startup M&A and more precisely a lack of startup M&A. If you look at this chart that we have here from the NVCAQ3 report, you can see that overall exit volume for startups in the U.S. has gone down dramatically. Essentially since 2021, we've seen the exit market for startups collapse. And what that means is we care quite a lot about deals that do get done.

Now, pick your poison, is it regulatory oversight? That's too much, too few IPOs, maybe valuations were too high. It doesn't matter. What does matter is that we just haven't seen as many startup exits lately as we would like to see. So when deals happen, I like to talk to the people behind them. So today on the show, we are not going to talk about Yelp and Repair Pal. Instead, we are going to talk about a deal that I think is much more interesting, which is Amplitude buying Command AI.

Now, if you're not familiar, Amplitude is a former startup that was public back in 2021. It is a public unicorn today. And Command AI is a startup that raised about $24 million in known capital while private, including a $19 million, roughly, Series A back in mid-200. 2022, which pitchbook pegs at a valuation of $123 million on a post money basis. I wanted to learn more about how that deal got done, what Amplitude saw in the startup, why Command AI decided to sell, and also what founders like yourself can learn from the deal.

So please welcome to the show. We have Spencer Skates, CEO and co-founder of Amplitude, and James Evans, CEO and co-founder of Command AI. Hi, guys. How you doing? Alex, fantastic to be here on this week at startups. I think, Spencer, what I do is I just changed jobs. I make you come talk to me at whatever job that I have.

Alex, I'm a huge fan of yours from TechCrunch to this to whatever you're doing next. It's always a pleasure. You are genuinely curious about this world in a way a lot of other journalists are. So it's always a pleasure to hang out with you. I only had to pay him $15 to say that. Also, James, welcome to the show. You and I just met, but I'm actually familiar with your company. Can you tell people just a little bit about your background?

I know you were at Bain back in the day. Yeah, totally. Thanks for having me. It's fun to have a place to talk about this. That's not a press release or LinkedIn. So I appreciate the opportunity. Yeah, I started my career at Bank Capital on the private equity team where, among other cool experiences, I was exposed to a lot of crappy software.

And that led in a circuitous path to our company, Command AI. So in a nutshell, what we do, we call it user assistance software. So the idea is make the internet, whether that's a piece of software you log into or an e-com store where you want to buy something. just make it easier people to use. The kind of motivating observation is that you just look at the stuff we log into on the internet or access on the internet.

It doesn't really look that different from today versus 20 years ago, even though there's been like a crazy proliferation of stuff we can do and the power behind it. We're still facing kind of like the same annoying usability challenges. Like, where do I do X? Do I really have to read this Help Center article

to figure out how to do it? So our company is basically technology to help solve that problem. And I know we're going to let you show it off here. But if I understand Command AI, there's other software companies out there that do nudges, kind of user assistance stuff. But your pitch has always been those are very annoying and users don't like them. And so I'm curious, how did you guys build something that people didn't hate? Because no offense to the world, when I get a software pop-up trying to help me,

I scream and go, leave me alone. Yeah, you just immediately close it. I guess it's like the dirty secret, like in our space. And when you ask like people who use sort of more legacy solutions, like, hey, just forget about your company. Like, as a user, when you encounter stuff like this, the stuff that you're like hurling at your users, what do you do?

And people invariably say what you said. They're just like, yeah, I close that. And then if they're closing it, it's probably not creating the impact you want to create. I think, like, it's kind of a visual product. So I'm glad we'll have an opportunity to show it. And like, the proof is kind of just in the pudding. Yeah, okay.

Let's do that now. And James, you have agreed to sportscast. And Spencer, I do want you to weigh in as we go through this, because of course, your perspectives on what they've built and what you've now bought are pretty paramount here. All right, James, so what am I looking at? Cool. Okay, so there's a few parts to our products.

So I'll just order you to the basic physics. There's the place where our customers' instrument experiences, whether that's a nudge or they want to tweak something in our chatbot, that's our dashboard. That's what you're looking at here. So this is where, you know, whether it's a marketer or product manager or a customer support person, like this is the interface where you're building these experiences. This feels very no code to me. Is that fair?

You nailed it. Yes, exactly. You know, there's a few scenarios where it makes sense to use our SDK. But for the most part, we really want to make it to that anyone who has like a hypothesis or idea about how to solve a user problem or move a metric can come in here and like instrument that as an experience without having to bother engineering. Because if you have to bother engineering, like way fewer of these actually have. So, and here, what I have here is a user flow in which I can essentially reward

a user for doing something, and in this case, you're showing them a GIF and a what else can I do button. Yeah, exactly. So this is the kind of thing that you might show to a user after they completed like some sort of a milestone to give them a bit of dopamine and also suggest like a next best action. Okay. Now, Spencer, one thing I'm actually not aware of is how no code friendly the user interface is over at Amplitude. Does this interface mesh reasonably well with how Amplitude looks today? So there's two parts to it. First, yes. Yes, this is the direction. We just did a big launch of Amplitude Made Easy a few months ago,

where the goal was to make it friendly for low-code and no-code people who wanted to use the Amplitude platform. The second answer, though, is no in that James, Vene and team have done such a phenomenal job of creating interfaces that feel native. That's actually not expertise that we have an amplitude and a big part of why we're excited to acquire the Command AIT. Okay, that makes a lot of sense because this looks pretty clean. But then again, if Command AI couldn't make clean, well-defined user software, they wouldn't be very good at helping other people do that. So I guess kind of table stakes for them to be good at this. The expectations are too damn high.

Yes, that is a no burden we have to bear. Okay. And what else can you guys do apart from nudges and diffs? I know there's a, I think the survey component, James? Yeah, well, actually, I want to just get in and show that kind of end user-facing part because when I said the proofs and putting, like, this is what I'm talking about. So you work at a company, you instrument experiences,

or your users. This right here is just like a sample app we built to demo Command AI. That is like a calendar to do list app. So I'm in this product. Maybe it's my first time or I'm just not an expert. And I'm clicking around. Like I'm doing the thing where I don't really know how it works. Maybe I'm trying to find something.

And I've kind of got these like spastic mouse movements. This is the kind of thing if you watch sessions, people notice all the time. It's very obvious when a user is confused. And you kind of just want to reach through the screen and be like, what are you trying to do? And so one of our most popular loops is actually instrumenting that. So if I'm like looking around here, we can actually throw a what we call a nudge to kind of intercept that confusion. What I'm doing here is in the top right of the screen, a little message popped up and says, hey, you seem like you're a bit lost.

Let's help get you on track. What do you want to do? and I'm presenting the user with some options. So here, we're showing things that we think based on what we know about the user. They might be interested in these could just be things that are common on morning action. But to be clear, though, if you were just using the service, this wouldn't have popped up. Instead, tracking mouse movements to see when you might be lost. So effectively, you've told it how to note when a user is just kind of like flipping around.

Yeah, exactly. It's really important to pick up on that because a key problem that many tools, in this space make is they show things to users at times when they don't care or are interrupting them. Often, it's actually just as basic as you hit a page and you immediately see a pop-up, which is actually, if you think about it, like, the worst time to show the pop-up because you just clicked on a page, you want to see what's on the page before maybe interacting with the pop-up. So we put a lot of emphasis on these triggers. We call it the when. And in this case, this is one of our most popular triggers. It's a confusion detection trigger.

I'm still looking for exactly those mouse movements that you're talking. Okay. Twist is brought to you by Squarespace. SquareSpace is the place to build a beautiful website. You know we're huge fans of Squarespace here at this week in startups because every startup means a gorgeous website. People will judge a book based on its cover, but you don't have a ton of time. You don't want to waste cycles on doing something repetitive like creating a website.

Oh, my Lord, your team is busy building your product or service. Why not use the best? and most powerful website builder ever created, and that's Squarespace. And they have a new AI-powered design intelligence tool that's going to help you build something great, unique, and fast. And they've, of course, got an e-commerce engine built-in. It's so simple to use with built-in payment technology. So if you're selling something online, you can start racking up that revenue right away. Squarespace will also give you all the tools you need, like an email campaign, donation, capabilities, and of course, domains, analytics, SEO, all built

in from day one. Squarespace has what your startup needs, and we can save you a few bucks. Squarespace.com slash twist for a free trial, and when you're ready to launch, go to Squarespace.com slash twist and get 10% off your first website or domain purchase. Again, Squarespace.com slash twist, an amazing product that I personally use and endorse, and the longest running supporter of this week in startups. Go ahead and tell them that you found out about Squarespace from this week at startups. So back to surveys, though, because I love knowing what people think. I want to see the survey element of this. Oh, yeah, sure. I can shoot in service. So yeah, so here what I'm doing is I just, I got rid of the GIF and I added a text input. So this could be a survey that

you show a user after the same milestone, but instead you're kind of looking to get their take on what just happened. And this is something we find companies actually kind of criminally underuse is they think of surveys as like these long things that you hit users with, like, infrequently. We actually encourage everyone to do a more micro-survey approach where you're constantly showing things or trying to get the pulse of what the user or visitor feels about what they're seeing. And the really cool thing is you can actually then use that same information that you're getting for targeting. So a really basic example would be, instead of a text input here, let me actually do a rating input, so a star rating. So here I can maybe rate, you know,

how that experience just went. Look, the fact, what you guys saw earlier with the confusion detection, that, when we saw that, that is fucking genius. Yeah. Because you are intercepting the user at the right time in the right context. It's shown that if you have a pop-up in the right context, that is five or ten times more engaging than just spamming you with something. So getting something that's personalized to your behavior, to what you might be experiencing, what might you want to do next?

Like, that's next level. And so I'm excited to live in a world software that does that. that was customized for me and customized for each one of us. I would actually frame that differently because when I get hit with a pop-up, to me, it's not like a zero-something. It makes me a much more negative experience. So if this can be a positive thing, it's not just going from zero to five, it's going from negative five to five, you know?

So glad you said that. Every other product in this space, you have to measure the rate of which users close your profits. Like, everyone does that. People frame it in the context of click the rate. So what percent of users who had an impression of this nudge? what percent of them actually kind of took the next action, click the big book.

We also measure something that I think no one else in the space would like dare to measure, and we call it a rage close. And this is if a user closes a pop-up quickly. If a user closes a pop-up quickly, it's so much worse than not clicking through. You're exactly right. It's annoying.

And it contributes to this sense of like, this product is hurling shit at me. It's not personal. So I'm just going to develop pop-up blindness and just never pay attention to these things. We actually see that if someone Rage closes three pop-ups in a row in a site or a product, you've basically lost them forever and you'll never be able to engage them with the farm factor again, which is like a terrible loss. And you will not be able to pick that up unless you differentiate between this state of like,

I didn't engage with it, but whatever, no harm of the foul, versus like, no, they actually hated it. I'll show you one more thing. We focused a lot on our Nudges product, which is the kind of proactive message. kind of getting in front of the user before they maybe know what they're trying to do. Yeah. We also have another product, which we call co-pilot, which is a AI-powered chatbot, basically. We call it a user assistant. I shouldn't call it a chatbot.

I thought the website said it's not a chat bot. It's the next level. It's better than a chat bot. So here, let me give you an example. So how do I create new? So here, we're utilizing a lot of information that we have about the products. Like, for example, a help center, internal wiki. But one of the things that we do that no other things that call themselves chatbots do

is we'll actually teach the user how to use the underlying product. I think a lot of people treat chat interfaces as like a fallback that's like, oh, I can't use the website. So I'm going to talk to someone through the chat bot. We think the right mental model is like, no, the chat interface is just a weary first way of helping the user and often the best thing to do is actually to teach them to fish instead of just like tell them how to fish. But here, instead of giving me like a long,

text answer. It's just saying like, okay, I can just show you. So I click here. Now it's throwing up a hotspot and pointing out in the interface where I can do the thing. And then it can take me to the next step as well, if it were a multi-step blow. And this is something we think is just takes the chatbot form factor or makes it like way more useful. Well, I think that it actually would be useful because the last thing I ever wanted to use a chat bot and then gets sent to a help center article that I then have to read and then relogged back into the application because I got logged out and then try to get back to where I was, and then I just give up and go back to using Google Docs.

Totally. Here's three FAQs that might answer your question, but probably don't is something we're trying to vanish from the internet. I've always read that as, we could hire humans to help you, but we're too cheap. So here's some text that we had AI right for you, and we hope you like it and too bad if you don't. Alex, the worst part about it is you hit on it, which is that humans crave that authenticity, and they can tell if they're talking to an AI or talking to an

another human. And so if you're just getting a long wall of text, you're not actually getting any value out of what it is you're interacting with. Whereas if you're being guided and shown the experience, like, that is a magical use of AI for me as an end user. And so the fact that James and Vene and team had this incredible vision for what this could be, that's what got us at Amplitude, very excited to work with them. Yeah. So let's pivot back to Amplitude now. When I think about your company, Spencer. I think of it as digital analytics with a product focus. Is that fair? Yes, that's right. Okay. And I was going through your last earnings call, which you guys talked about this transaction quite a lot through that. And you said a couple things that really stood out

to me because I wanted to get into why do this deal? And you said, customers are looking for consolidated solutions. And essentially, I read that as point solutions are becoming just harder to sell and so forth. Why not build something? Why buy in this case? Because I, I, love what James and team have built, but Amplitude has a lot of resources. You're a public company. You could do quite a lot. So why in this case, um, buy not built? James and Venean team have an incredible amount of expertise by working on this problem for the last four years. Four years, is that right? Yeah, four years. Four years. And so that's way beyond any expertise anyone of Amplitude has. And so we could spend three or four years to try to get to the same point that they've

already built this amazing product to, or we can say, hey, let's partner with you and let's bring you in as part of Amplitude and let's have our combined expertise, get this to market a lot quicker. We've actually set a very aggressive internal goal for this to get out in the first 90 days. So 90 days from when the transaction closed, we want to get this out to our customer base and get them using it. Because I think you guys announced this October 15th-ish, so it would probably be about a week and a half before that. It was the same day. It was the same day. We closed on the day we announced. Yes, that's right. That's right. A bit of a heroine experience. Yeah, there's no, there's no net under that tightrope. Okay. Well, shout out to you guys for being bold, I suppose.

The thing that I'm kind of surprised by your answer to that, Spencer, is when I think about the way some investors, really on the private equity side, think about software. They say things like software tastes like chicken. It's all kind of the same. But you're describing much more craft and almost like art than science when it comes to building software like this. Is that a standard element of building modern enterprise software that people kind of want to use? Because I don't think that user experience was as much of a priority 10 years ago. I think particularly for B2B, we're undergoing a renaissance now where the great products went out. If you look at companies like Datad or GitLab, they have done an amazing job crafting these wonderful

user experiences. And they are winning because largely because of that. Because there are lots of other pieces of software that do very similar things from prior generations to what they do. But they've won because they've made it simpler, because they've made it more usable, because they've just created these amazing end-to-end experiences. And so I'm in a belief that the next generation in B2B software is going to look like that. And so we've been intentional about that at Amplitude 2.

One of the big strategic pillars that we have is we call Win Simple, which is about how do you make it much easier to start? Analytics is a scary intimidating thing. When the average person hears analytics, they're like, oh, that sounds like Matt. I hate math. Like, what is this? And our thing is, hey, how can we make it accessible if you're not technical? If you don't know SQL, if you aren't an engineer that can instrument code, how can

we help you understand what your users are doing? And so that's why we've added things like session replay. That's why we did Amplea easy a few months ago. And that's why we're really excited to work with the command AIT. All right. You didn't start your company to run payroll. Did you? Of course not.

We all know that. Gusto is here to help. Gusto is going to help you run your payroll and handle all your benefits onboarding and HR all in one place. The market agrees. 300,000 businesses trust Gusto today and you can too. As your startup skills, Gusto is going to grow with you.

You got state and federal taxes handled for your staff around the country. Gusto does all that. And, hey, maybe it's finally time for you to offer a 401k plan for your team, right? Gusto's got you on that. And you might need to get your compliance sorted, right? Well, three out of four employers say Gusto helps them be government compliant. And even better, Gusto is simple, easy to use software. So you can focus on what matters, building your startup.

So here's your quick call to action. Do you want all Gusto has to offer with no hidden fees? Well, how about a discount? Try Gusto and get three months free. Gusto.com slash twist. That's G-U-S-T-O-com slash twist. It's funny, though, to hear you talk about this, because when I think about having easier to use better software,

I think about product-led growth, you know? And I feel like product-like growth as a concept has lost some of its, I would say, favor in the market. But when you're describing helping people start using Amplitude, you guys talked about in your last earnings call, I think, like 40% of our customers are sending data to Amplitude, too kind of now. So clearly, you're seeing some momentum in this make it easier to use sign up and then hopefully expand your footprint. But that does seem countercyclical, or is the narrative

out in the market today about business software simply wrong? And people do care about crap. So I think there are tons of great examples of companies that have driven. You know, I named GitLab and Data Dogg. Upspots another phenomenal one that's really done a great job of cracking it. I think one of the things we've seen in the digital analytics world is we're in the middle of a transition. So if you go back to the previous generation of software and you look at what Omnature has done, they built out, they frankly, you know, and I love the Omnisher team. I love Josh and I love Brad and we have Catherine on our board and, you know, we're tight with a lot of the folks that built that company. But they frankly had a worse product. They had a

worse product than a lot of their competition. And that was okay because they were killer on the sales side. And that was kind of Generation 1.0 of SaaS. You don't have to build the best. crafted experience, you just have to build the best and most scaled sales force. I think what we're changing to is Generation 2.0 of SaaS, which is where the product experience matters a lot. Now, in our space for digital analytics specifically, we are a sales-led motion. So we're still, that part does matter. It's not like you can just put it out there and have all these folks come. But the way that they're choosing in these sales processes, it's about, okay, well, I need to enable thousands of people to self-serve on data, which

tool is going to let me be most successful at that, and I'm going to choose that one. Okay, so essentially the idea, the Vista Equity Partners approach of software tastes like chicken, big sales team, same playbook for every single company, doesn't entirely go away, but it has to be combined with great user experience and the things that we associate with product like growth. So the answer is not that one way of selling was the only way to go, but instead a blended approach, it seems, of high quality software and a high quality sales team. that fits under my impression that SaaS has gotten harder in the last couple of years as expectations have gone up and pricing pressures have also increased.

So that, that fits. Now, Spencer, I'm going to put you aside for a second and talk to James. James, Spencer just saying your praises about the team, the software, you guys raise money, I think up to a nine-figure evaluation, cool company. Why sell now? Great question. I think there's sort of like, there's an unhelpful. meme that exists in like startups, which is like, if you don't need to sell and you consider selling or you like even entertain like the concept of an acquisition, like you're, you're weak

and you're like not a committed founder. I actually think Zuck is to blame. I think the like Zuck, however old he was, like 21 year old Zuck walking into Yahoo and flip flops and like turning down the billion dollar acquisition. I think contributes to this meme a lot. So for us, it came down it a couple things. First of all, I'd say like, you're not going to consider selling the company if you don't need to, if you don't feel like the value is fair to the team and investors. And like, that's a, that's a decision you make with your board. So like, I don't think you can really have this conversation without talking about value. Okay, but to be clear, based on what you just said, Command AI was not out of cash. You were not, you know, ceiling,

burning down, oh my lord, we need a savior. I'm going to call Spencer and see if we can do a deal. No. No, definitely not. We were not thinking about selling the company. We had plenty of runway. Growth was good. Which I think banks the question of like, okay, then why entertain it? Yeah.

Say number one value. Number two, I think there's a way of approaching this, which is just like, forget all the trappings of an acquisition, like, you know, going to work at the company and all that. Can you achieve what you want to achieve as a company more, better, faster by being part of another company. And for us, my like sort of existential concern or fear with building a standalone company is I really think we've built like already a tremendous product that can create a shit ton of value for companies, but also just like make the internet a better place,

like banished these like annoying experiences and like help people get more value of the stuff they use. And that's where scale really comes into play because if you could have, you know, your same level of growth that you had before, it would take 10 times as long to reach the same number of people. And if you can plug into, and Spencer, back me up here. You said that you're now 3,500 paying customers, if I recall correctly. Yes, that's right. Yeah, I was salivate over the ability to get our stuff into their hands. And I think the fear is like you could build the great product, but like, that's not enough. And if you can't crack distribution and if you can't keep getting it right, like, you're not going to be able to achieve as much as you wanted

to. And so for me, it was just such a natural, like, strategic fit to, yeah, sell to those 3500 customers and all, you know, more customers will go get. And the last thing I'll say is, there's just a lot of like, the physics of our business align really naturally with the physics of Amplitude. Like, if you look at the demo I showed, you don't create something in command AI

unless you have found out something that's not working in your product or customer base or have like a hypothesis to something that could work better. And 95% of the time, those ideas are coming from a tool like amplitude. Like someone is using Amplitude, figuring out that like, oh, you know, users of this type aren't using this feature.

okay, great. The thing they have been able to do in Amplitude is, let me go fix that straight from the product. It's, okay, I'll go back to engineering and I'll instrument something. Now with our thing, they can just immediately fix it. And so just connecting, like, we're almost finishing the feedback loop of the thing that starts in Amplitude, and that logic is what ultimately led us to think it was a great tie-out. Okay, so to put this into relatively prosaic terms for everyone listening, people use products. Amplitude keeps track of what they do and how they do it. And then, with Command AI, you can then go, okay, here's a bottleneck we've discovered, and we're not going to go apply one of our tools to help unstick people or help them get more done or more value

out of it. So it's kind of a, you're on top of the stack, if you will. Exactly. Like, we kind of think of it as like, you've got a relatively fixed rigid layer, which, like, is the product. We're not trying to replace, like, a website or the app or whatever. But then you've got this kind of more liquid layer that you can experiment with and can be more personalized. And that's what we're attaching on to all of our customers. I kind of like that. But again, I'm kind of shocked, Spencer, you haven't had this before because you mentioned in your, in your notes that people were asking for this.

I'm kind of shocked that amplitude didn't build kind of a quicky version of it, just to have something in place before this deal came to be. Is that just you guys being focused? Yeah. So we've always been focused on being the best in class in digital analytics. It's only more recently in the last few years that we've started to branch out to experimentation, session replay, and other things. And so this was one of the things that we were looking at.

And it was really clear to us that we'd much rather partner with a talented team because that, again, it would take us a bunch of years to figure out how to do this ourselves. Like, there's a bunch of expertise that James just showed in the demo that we didn't have. How do you create an interface where a pop-up looks natural and in context of what's already on the screen? How do you think about the management interface for, you know, tens or hundreds of these nudges and guys? how do you think about, you know, creating all the right templates for surveys? So, like, so many of these interactions that we just didn't have expertise on. And so, you know, we're not, you know, as big as we may seem as a company, we're not that big.

And one of the superpowers that we've honed over time is how to stay focused. And so it made a ton more sense to partner with the command AI team. The other thing I want to say on that is that acquisitions are incredibly risky. There's a stat where 90% of acquisitions fail. And so this is something, you know, it's not like that was the obvious pat either for amplitude or for command AI. That's funny because it feels from where I'm sitting, based on what you've told me, it does feel relatively natural because people want, you know, platforms, not point solutions. Your customers were asking for this specific capability. All right.

Take a second and picture the ultimate all-star team for your startup. Okay, you got that mental image? Maybe the Avengers, maybe even X-Men, Wolverine, Cyclops. pick your favorite superhero team. And let's be realistic. Until you've raised your Series A or your Series B, you might need some help building that dream team and finding top tier talent,

managing all the timelines you've got and maintaining quality, it's a big lift for all founders. I mean, scouting for that elusive 10X developer, that's going to eat all your time. And that's where Dev Squad comes in. Think of them as like Professor X, you know, they know where to find those exceptional mutant developers

who are 10x devs, and DevSquod offers a complete product team from Latin America. You're going to get 2 to 6 full-stack developers, a technical product manager, and specialists in case you need them like product strategy or UI and UX design, DevOps, and of course, quality assurance, all for 75% less than a U.S.-based team, all for 75% less than a U.S.-based team, while being on your same time zone, which is really important. There's no long-term commitment. It's just seamless collaboration. So if you want a GSD, if you want to get stuff done,

and you're ready to squad up, head to devsquad.com slash twist for 10% off your first engagement. That's devsquad.com slash twist and transform your startup with the all-star team it deserves. I know you guys had, if not customer overlap, some information that I know, commandee, I was sitting customers to amplitude and vice versa. So to me, it actually, it doesn't seem that. risky on the surface sensor. So tell me, you know, what is the risk that you're going to be managing now that this deal is done? Just overpaying or lack of integration? No, I mean, that like, you know, that, I think you have all this overhead where it costs a bunch of work and effort

and, you know, change on both sides. And so you're, you're kind of, it's a high barrier to say, okay, hey, it makes sense to do this. But the incredible benefit is that you have this combined vision. So we strongly believe in a world where products no longer are frustrating. Everyone has had a frustrating experience in their life with technology. And it's because you have to adapt your brain to how the technology works. And we believe with data, you can actually make technology adapt to your brain much, much better than it does today. And so we want to partner with other great teams that have that vision as well. And so it wasn't just, okay, the command AI team had this fantastic product and expertise. That was, you know, that was obviously kind of a part of it, and that was a

great proof point of their capabilities. But it really was, hey, we have the same view of how great software should be built. And we want to allow every single company out there to use that. So you had vision alignment, essentially. Yeah. And the other thing I'll say on that is I think the biggest mistake a lot of companies make in these is they're not clear about what the end state you're going towards. And so we were really clear from the start that, hey, it's going be one platform together, and that's what we're going towards. A lot of times companies will leave ambiguous and they'll be like, well, we'll let you run standalone, but then we'll kind of integrate you. And that's like the worst of all worlds. You either leave something fully alone and let it run,

or if the point is, hey, we have a combined vision to create this product that's better together, then you want to rebuild it as part of one platform. And so one of the things we're very deliberate in talking with James and Benet from the start was, hey, it's got to be one company and one platform and one product and one brand at the end. Okay, so, James, I want to ask this to you because you're the company being purchased. And I have known Spencer for years now. He seems very affable. I've always enjoyed talking to him.

But I don't see him behind closed doors with something goes wrong. So how did you know that joining his company and his culture was going to be the right fit for your team and its own unique culture and so forth? Yeah. The first thing I'll say is like, it's always a risk, right? You can only do so much of that diligence like outside in. and you're never going to get to 100% comfort. And that is, like, one of the things that definitely wait on me and Vinay was, you know,

today we're in total control of decisions, but also culture. And like, we know we're not going to, that's not going to be the case as we're acquired. And so, honestly, I don't have a great answer for you other than Spencer and I just like hung out a good amount before the deal. And I just got comfortable that from a product and like strategy perspective, we had alignment. and from a culture perspective, like, actually, it's funny. There was, I think Amphid has three values. We have five values and there's overlap in two of them.

And so we felt like there was, you know, we were going to be part of like a team with relatively similar, you know. And just everyone knows that the company with more stated values is more virtuous by definition. We were, we are diluting our virtuousness by entering a company with fewer values than us, yes. What were the two that overlooked? Ownership and growth mindset. We had slightly different words for them. Yeah, yeah, but the same concepts. That's very interesting.

So I want to go back to the start of this going from partnership to transaction. Which of you was the first to say, hey, what if we just did this as one company? Some folks on the product side had a tremendous respect for what the command AI team had built, and we had reached out to them, and James and Monet were incredibly friendly and were willing to share what they had built. I think the thing that both the product experience and the fact that we had a lot of overlap in customers, that was kind of the initial signal that it would be worthwhile to have a conversation as we got more serious about considering buying a company that that was able to do nudges and guides and surveys.

Yeah. And then James, how fast was the company growing before the deal came to be? So, I don't know, like how fast did you guys grow year over year in the first half of 24? We were growing about 3x year on year. So that's real quick. I mean, even for a series A startup, that's kind of the triple triple double, double, double, everyone talks about. So given your growth and the fact that you raise a good amount of venture capital, how much work

was it to get to your investors and get them comfortable with the transaction terms and taking an exit for an asset that had potentially more room to run? Yeah, I mean, totally. That's like the fundamental thing that is what kept us up at night, frankly, was determining you'd be really confident that this was actually the outcome we wanted, starting with that kind of LinkedIn DM, that Spencer mission that kicked off the whole process. But it was a LinkedIn DM.

And that is the most NBA thing I've ever heard. You got to do this a Twitter DM. You've got to be casual. Come on, Spencer. Thank God I saw it. Yeah, no, I was about to joke that I never actually read my LinkedIn. No, me neither. Now I wonder what's in there.

Yeah, maybe someone's trying to require you. Oh, God, please not. But anyway, so Spencer reaches out and then you have a decision to make on your end with your board and your backers. Yeah. I've heard like four stories where, you know, founders have one view and investors have another view. And like, that just wasn't the case with us. Like, our investors were, we're super supportive. Like, helped us reason through, you know, how could this fail? What's like, what path? What's the path to glory standalone? But like, ultimately, their, their vibe was like, we're going to help you make this decision. But like, it's your decision. to make. And if you think this is the right for the company, it's probably the right fit for us as investors. So we, it was a very high stakes, but, uh, but, you know, we're all reasoned and kind of non-emotional conversation. I got to give James tremendous credit for how he navigated and handled it. It took us a little bit longer to get there. We were dealing with a few

things on our end, including a CFO transition. Um, but, uh, you know, he saw the vision of what could be, we saw the vision of what could be combined. And they were actually, command AI was in the middle of fundraising process. And, you know, they were good to go. That explains a lot. Okay. Yeah. So they, they were in good shape to build a very successful standalone SaaS business and continue that path alone. And so, you know, it took us quite a bunch of back and forth to figure out, okay, I have conviction that the equity of these two companies combined is much more valuable together than separately. So how can we create a construct? And so, you know, we negotiate quite a bit.

And James is a, I'll give them a lot of credit, very good negotiating. It helps. I think it comes back to like, do you, are you like, have you decided yourself in your company? And that's just like a totally different style of conversation and like thought process versus you're excited about the company. You're, you know, you're going to make it. If the deal falls apart overnight, it's fine. You just keep going. And that's very much how we approached the conversation was like, hey, guys.

We're happy. We don't have to do this deal, but we're really excited about the potential of doing it. And here's what we care about. And it just made it really clear. We described the deal that was interesting to us. And we were like, hey, no worries. Like, if this isn't the deal that's interesting to you, like you mentioned earlier, Amplitude could try to build the functionality.

It wasn't like, you know, we had to make it work. I think that made it much more of like a conversation versus like, you know, a really intense, you know, zero sum style. discussion. Okay, so from the first LinkedIn DM to the deal being closed, what was the timeline like? And I ask because we have a lot of founders that watch the show, and I think the timelines of emanated transactions are pretty opaque. So I'm just kind of, I want to get as much detail as I can to help those folks understand how these things actually happen. The first conversation was sort of like late spring, early summer, like I think, I think April or May. But I wouldn't say it was like,

we weren't like negotiating an acquisition since April or May. It was more of like an exploratory, like, hey, maybe it makes sense to partner, like, maybe it's a go-to-market partnership. We already had a product integration. Maybe it's a deeper product integration. So I don't think we really started like focusing on the, you know, hey, let's try to make an acquisition happen until like middle of the summer. Okay. Now, Spencer, I know you have a new CFO and I was reading their early commentary. And they said, I believe this is a direct quote. First, we are not a growth at all cost company. We are investing appropriately to drive accelerating growth while generating positive free cash flow. And then, of course, when it comes to leverage, you'll make decisions. How hard was it to get your C-suite and board on board with purchasing an asset that just given what I've learned about the company probably wasn't cheap? And I say that not as a diss, but more of a, it was growing fast. No, look, I mean, James had this past, you know, he was already growing very fast. Command A.I. was already growing very fast. They were

very successful. They could definitely be incredibly successful as an independent company. And so we needed to make it worthwhile for them to do so. I was very lucky to have an incredible board of directors that's seen the good, bad, and ugly on transactions like this and to get their guidance through this process. And so we had done a few smaller ones earlier, a few years back, and in both clear brain as well as iteratively. And so we had a little bit of experience. This was by far the the largest one that we had ever done. And it came back to buy conviction and the conviction of the team that this was a really valuable combination for amplitude and that this was something our customers were asking for and that we could sell a lot of. And so we put that case together

and they were very supportive of what we were doing. So I want to ask about the public company perspective because one thing I've heard from some founders and I'll say some VCs is that the government has so heavily slammed the door on acquisitions because of essentially antitrust concerns that no one can do a deal, transactions are kaput, and so forth. But then I'm sitting here talking to you guys and not once this far and, I don't know, talking for half an hour or so have you guys mentioned concerns about regulatory oversight? So did that come into concerns, conversations at all? Now, this sort of, this sort of acquisition is small enough that, you know, it's not something that's on the radar of the FTC. So,

But they are definitely cracking down for much larger companies and much larger transactions. Right. But I mean, not everyone is meta, right? And so to me, like hearing you guys talk about the deal, this, I'm almost surprised I don't see more of this stuff happening because there are so many cool companies out there building neat software and not all of them are going to go public. Maybe, you know, James and the command of AI team could have, but I guess the complaint that I hear, and I'm mostly talking to myself right now, but the complaint that I hear about

it's all Lena Collins fault doesn't match the reality of talking to you guys about. this transaction because it doesn't sound like you were terrified of the FCC kicking the door and saying absolutely not. For stuff of this scale, you know, this is, ultimately this is better for amplitude customers. It's just not a, there's not, it's not like the market that we're in digital analytics, nudges, assistance. It's so early, like nobody knows what's going to happen. And so, you know, we're way, this is many, many years before anyone has any regulatory concerns

about this market. I will say the reason that this doesn't happen and that you see this, this go down is because I think there's a mismatch in valuations. So if you're a public company or recently public company, valuation multiples are actually quite low. If you look at, you know, revenue or A or R multiples, you know, a lot of times you'll get like two, three, four X AR, whereas private markets are still actually very high. You know, a lot of times, I mean, you see some of the AI companies, they'll get like 200 X AARR. But even for midstage stuff, you know, you often be 10, 20, you know, or more AR. And so,

So what that means is those companies can run independently for a long period of time. And that means a lot of the natural buyers, larger companies, aren't going to be able to pay that much because they don't have, you know, the stock isn't worth that much. And so you actually, it's a much higher hurdle to clear. You have to have conviction. There's a huge amount of value combined. You can't just say, okay, these companies are 10% more valuable together because then, you know, the private company might as well continue running because its valuation is already

pretty high. You need to be able to say, hey, we can create five or ten times more value together than we would separately. Yeah, it's funny you mentioned the AI companies. I had a question in here that I didn't have to get to because we talked about that command AI could have raised more, but I was going to ask, given AI excitement to Command AI's copiled, I would think you could have raised more. And the answer is, yeah, because everyone wants to pour money into AI. Totally. I think people simplify or I think, at least speaking from the founder perspective, I think people simplify the decision as, oh, well, if we can raise at a better valuation, say you're deciding between raising at, you know, evaluation and sell your company for less than that

valuation. Oh, well, of course, it would take a higher valuation. In some cases, that's totally correct. You know, like if you're growing 10x year and year and, you know, your, your path to IPO is like two years and by all means, like, I'm not suggesting that everyone sell their company or sell it for less than it's worth, but you can't make that comparison isn't as straightforward as I think a lot of founders think it is. And I think a lot of founders don't entertain the idea that just think of a simple thought experiment. What's more likely that we get sold for 3x what we got sold for or that we can 3x the value of the combined entity? That's predicated on being purchased in equity versus cash. Because if you could purchase an equity and that appreciates, then you have even more upside. I meant to

asked this earlier, but if you guys shared what the deal was done in, if it was, you know, dollars, shares, monopoly money, small shells. We did a small amount in cash for some of the existing investors, but for James and for the whole existing team, it's all amplitude equity, because we want them to share in the combined upside as well. Right. And they were owners for day one. Right?

Like, we wanted to participate in that in the upside. Well, no, it makes a lot of sense to me. There's one, I want to get to AI and also a little bit about the enterprise in a second, But there's a tension between what you guys are telling me that I want to unpack a little bit. Spencer, big on offering a platform, attach rates, you know, multi-product customers. I think you said something like 23% of current customers have more than one product today. And then, you know, I have James and Command AI, which was more pointy solution versus broad. And they were going at 3x.

So those seem to be statements that go in different directions. I need to be more platform, but also points doing well. So what am I not understanding there to harmonize those two different data points? Alice, I just want to make sure I understand your question. You're saying that the command AI independently is growing at this great rate, but does it make sense to have as part of this platform versus run independently? Well, no, it's that you're emphasizing the importance at amplitude of having a suite of products that work together so you can offer a non-point solution.

But the point solution that that Command AI was selling in the market was doing well. And so I'm just confused between why you're telling me this, but I'm also seeing this over here. And I presume that I'm just missing something in between. So the whole reason that we do this is because it's that 3x can be a lot faster with Amplitude's customer base. Because we have thousands of customers, whereas Command AIs much, much earlier. So by getting the product into the Amplitude customer base, instead of growing 3X a year, you're more like growing 20x or 30x or even more in terms of the growth rate because it's part of that platform that so many other people already using.

And how many of your existing customers do you think we'll be using command AI software via amplitude in, let's say, 12 months post-closing? Is it half? Is it 25%? That's a good question. We actually haven't, we haven't looked at it from that lens. I think we want to be really aggressive in getting it out early next year and then getting You know, probably, it'll probably be, you know, 10-ish percent in the first year is a good rate. But then, yeah, eventually you want to get it to 50 and then 100. James, 10% in the first year, it feels like crap numbers to me. Can you get 25%?

I don't mean I'm allowed to make forward-looking statements is one of the things. Oh, right. I've been schooled on. But I think, just to go back to your question around like Point Solution versus Platform, I think there's sort of like a misconception in the space that like, if you have a platform and if you have a bundle, your products, the bundled products are like kind of shit, but like collectively they make sense because platform or because one buyer and because like discounted pricing, it's not like, we're taking our

product that was selling well, independently differentiated, and like making it worse so that it can fit into a bundle. Like, it's the same product. But now when you create a cohort in amplitude, you just immediately get to use it in command of, or you can view the sessions where users have impressions of your assistant stuff in amplitude. So it's like, getting better, but it's not like it's losing the ability to stand alone compete against other products. In fact, I guess, one of the most exciting things about this deal. And one of the reasons why I think it can make sense for companies to buy instead of build, because you're getting that like standalone differentiation out of the box. Okay. So given the endorsement there,

Spencer, of buying versus build, I'm just curious. And I know public company and I know you just did deals. This is a slightly brat question. But, you know, how attractive is doing more acquisitions in the next pick a time frame for Amplitude. And I'm asking this because I want to understand what other public company CEOs might be thinking about, you know, the market, what's available in prices. So first on our end, we're absolutely going to be doing more. I think there's lots of great technology in the space out there

with people who are just as smart as the folks that we have here at Amplitude that we would love to partner with. So we're going to be doing more. We're also going to be building more internally. It's going to be both. I think that the great companies are able to do both. One of the things I talk a lot about is that we are so early in our space that product, to James's point, product, it's not like the innovation's done.

It's like, we're at the actually at the really early days of innovation. What you guys saw, like, that is the bleeding edge from the command AI demo of what great user assistance and great nudges look like. And there's so many other similar adjacent spaces that I think that make a lot of sense with the digital analytics and with the platform. and so we're absolutely going to be talking to more teams and going to be doing more acquisitions in the future. The one other thing I did want to say is part of what made this so exciting from my standpoint as Amplitude CEOs, the thing we're always behind on as we grow and scale is great leaders.

So just great leadership. You know, as you grow as a company, the ability, it's like you're always at a deficit for, you know, if you grow 10% in a year, you need 10% more leadership. You know, you need 30% more leadership. And so we're always behind on that. You're always trying to like cash up in a technology organization. And so being able to work with like James and Bonae are really important part. They're not just, oh, the command AI founders.

They're actually, I think a lot about how do we get great product and engineering leaders here at Amplitude to own bigger parts of the stack and of the platform and of the future things that we're building. That's very interesting because there was a, a, a meme, and I said that very politely, in the startup world, the last six months about, you know, founder mode. And instead of federating out responsibility to individual leaders at a company who were subsidiary to the CEO, the founder, instead kind of constraining that authority and decision making to some degree inside the singular executive. But it sounds, Spencer, like you're comfortable with having some distributed leadership and perhaps more than the founder mode meme said was correct?

I want to give a little more new. I was actually at Brian's talk, and it was crazy. You were at the YC thing? Yeah, he was basically had just come out the other side of a religious experience. And to hear that directly, the level of passion and intensity and just like, yeah, religiousness that he had about what it meant to run a great organization. He had this great thing where it's like, yeah, it takes 12 years to make a good CEO, you know, because he felt he was a bad one for the first 12 years.

of Airbnb. The thing that, the thing that I think the, it's fun, and he actually wasn't the one that came up with founder mode. That was Paul Graham, who's just an incredible marketer. The thing that, the point, it's actually not either of the scenarios you talked about, Alex. It's not delegation, and it's not, hey, like, I'm deciding all this stuff myself. It's actually, the point Brian made was that you are in the details of what is critical path for your company. And so for Airbnb, that's the great product experience. For Amplitude, that is also the great product as well as how do you sell this product. And you can't delegate those. And what I'm looking for is fantastic leaders who I can partner with in those details. And so as an example, one of the things I've had for the last year is

I do a two hour a week product review of how do we make the platform easier to use. what I'm looking for in that meeting is who are folks that I can go back and forth with as peers to say, hey, what does a great experience look like? And it's not just me saying, hey, do this, do this, do this, do this, do this. It's saying, okay, hey, here's what I think. And then they'll say, hey, here's what I think. And it's that back and forth that leads to that combined great outcome. So it's not a delegation, but it's also not a micromanagement. It's about, I want to get people at amplitude that I can have that interaction with. And that's, what we're always at a deficit for. Okay, so turning this into a World War I analogy, because

that's where my brain went first. You're not back at headquarters 20 miles behind the front line, nor are you the only person aiming down a rifle. You're instead in the trenches and very involved with the strategy. Okay, that makes a lot of sense to me. The key point on the founder mode is there is you have to be in the details for whatever is important in your company. You cannot delegate that. That's not to say you don't work with other people. You absolutely do. And that's not to say you micromanage them, but you're working with peers in the trenches with them. Yeah, it's the opposite of what Boeing did, which was abstract leadership way. Totally.

Disempower the engineers and turn it into, I'm just going to keep making private equity jokes because James is here. Turn it into kind of like the private equity, you know, platonic ideal of a management team. Dude, I, sorry, I was going to rant on private equity for a second. I love that. They make so much fun. They make tremendous amounts of money, you know, one of the, one of the great players. in modern capitalism. But one of my rules is never be on the other side of a transaction for private equity

because they have a spreadsheet about how to take all of your surplus away during that transaction. So, you know, you can either do a transaction and get no surplus or not do a transaction. So you might as well not do a transaction. But I love them. I love them. They're great. They're great capitalists. So I have tremendous respect.

Just don't do transactions. I think you can be a capitalist as I am and have one or two disagreements with standard private equity approaches to personnel and staffing. Dude, if you want to talk about where the FTC should regulate, fucking regulate the hell out of private equity. Like, that would be my first order thing. Because like, if you look at their playbooks, most of their playbooks are, hey, consolidate market share to create monopolies or pseudo monopolies and raise prices on consumers. And if you look at U.S. anti-trust legislation, actually the test of are you behaving anti-competitively is do prices go up for consumers or down for

down for consumers? And like the private equity playbook 101. So Lena Kotick, you're listening to this or whoever the next person the Trump administration is, fucking take a hammer to private equity. I'm all for it. You know,

I really think that if you were hoping for private equity, you'd take a whacking, we have elected the wrong next president because I don't suspect that's going to be a 47 goal. All right, we've gotten a little bit off the rails here.

I want to bring it back to Spencer, what you're building at amplitude. You said something very interesting about your near-term goals and your near-term focus. And you said that there's two things you're working on right now. The first is automated insights, which you can use to surface two customers practically without human intervention, and then automated actions, where Amplitude can, quote, recommend and take actions that we know will improve the end customer experience.

Now, the first one seems to be digital analytics and also, I would say what Command AI does. The second smells quite a lot like AI agents. and people have been making a lot of noise about AI agents, Brett Taylor over at Sierra, Open AI and so forth. And so I'm just kind of curious, is that the direction that you're pointing at here with that comment? And if so, what does that look like in an amplitude contact?

Yes, absolutely. That's the exciting part. So you saw that great interaction that James just demoed, where you're asking the chatbot, hey, how do I create a new project? And then instead of just giving you a list, it actually shows you. It says, hey, here's, I'm going to create a tutorial just for you and I'm going to walk you through this process.

That is the future of software. And so one of the things I'm really excited about is how can we automatically create? Right now, it's a very manual process to create those experiences. How can you automatically create them? So we can, every single person can get a custom guide to the software. So the guide to software is custom for me or for you, Alex, or for James or for anyone else. But more than just guides, though, because you're talking about where Anthony, you can take, recommend and take actions that will improve the end experience. That's more than

just nudges. That's more than just. So, so nudges is, is kind of a very, you know, like, that's like, the first step towards, towards doing a bunch of this. I think there's a whole bunch else. Like, one of the areas we're really excited about is starting to recommend different text or images on your website based on what we think will convert better or resonate with a particular user or even change the layout of a website or break down and create functionality. That's, That's where this is ultimately going. And that's what, when I think about the end vision, it's, we want to create a world where products are always improving, they're self-improving all the time.

Okay. And then this will all rest on an ocean of data. And one thing that you said that one of the quote, really exciting things about amplitude is that you have this enormous repository of data, a customer behavior, one of the largest in the world. So does that mean that there will eventually be, I don't know, like an in-house amplitude, LLM or similar that you guys then deploy against your data set for your customers? Yes, absolutely, absolutely.

So are you going to raise another billion dollars to train it? We haven't gotten to that point yet. I think we're doing a bunch of different things with AI internally. I think what's interesting is that you have so many dollars going to Google and meta and Microsoft and Open AI where they're spending crazy amounts of money to create these phenomenal models. And so our thing is not to get into the model building business ourselves. Our thing is to leverage those.

And then, okay, what we're experts is putting them to work for this particular application. So how do we know how to train the model to recognize the right types of text that is going to lead to an engaging great user experience or create the right type of nudges or what have you. Okay. So then the question then becomes, and again, thinking about founders that are watching this, just helping them kind of learn how you think. There's two approaches to using off-the-shelf AI technology. there is closed source models and open source models.

Open AI, of course, closed source, meta's llama family medium open source, if you will. So where does the amplitude direction lie? Which way are you going? Oh, we're using a lot of the open AI models, but there's no reason you can't use the open source models with amplitude as well. Okay, so you're not, you're agnostic, but currently giving Sam Altman money.

Okay. That's right. That helps. He's losing money on every transaction. That's the crazy part. Open AI is losing money on every time you, you submit a query to it, which is bananas. Well, that's why the best way to take big AI down is to use lots of big AI.

That's why I use GPTO1 or whatever the hell it's called all the time, because that's the, I'm joking. Sam Aldman, I've only met him once. But ghost margin negative is a weird place to be for any business. It is. It is a weird, it is a weird spy. It's going to be interesting to see how these start. The other problem is a lot of the AI businesses have much worse retention characteristics. And so, you know, it's not sustainable.

at current level. So it'll be interesting to see where it ends up a few years from now. Well, I'm just glad that Amalthitude is not going to get involved in the great CapEx race. We're in SaaS, which I think back in 2021, a few people figured out that this was the best business model of all time. And it still is the best business model of all time. So we're, we're just happy way creating a great product and continuing to sell it to more and more companies. And we're going to be building a great business that way. I want to say that that is a contrarian statement in late 2024, that SaaS is still the best business model out there because it's the best business model of all time. It's taking lumps,

but that's also a beautiful segue to my last question for you, Spencer. You guys talked about how you thought that the majority of overbought optimization contracts has kind of gone through, but the macro environment is still choppy, quote, buyer scrutiny remains high. So I'm just kind of curious from where you sit, what is the state of the B2B software buyer today and how quickly is it improving or getting more healthy or not? So I think you had a lot of of software buying back in 2020 and 2021, that frankly just didn't get any scrutiny at all. So as long as you could justify it, you could spend crazy amounts of dollars on it. I think what is changed is that people now want ROI for their software.

They want to make sure it's addressing their top priorities and they're not going to spend anything on, you know, they're not going to spend anything on whatever it is that's out there. And so I think that's going to continue. Like, people want to make sure that they're getting, yeah, a real, return for what they're doing. Now, the great thing is data analytics and understanding the user journey, that is a top priority of almost every single company we talk to out there. And so whether times have been good or whether times have been bad or wherever in between, there's going to be a lot of demand for what it is that we do. And so we're going to be out there selling

it. I just realized something, though, about what you just said, which is that in a higher scrutiny software buying environment, the fact that Command AI was growing as fast as it was is actually more impressive because they're not growing that fast in 2021 when everyone was just taking bricks of cash and going, whee, sass. Let me tell you one of the most impressive things. Actually, so Command AI was the number one rated on G2 against much, much larger companies because that experience you saw, you can't get that experience if you're using the top guides product or the top service product or someone else out there.

So it's just, yeah, what they have done is incredibly impressive and we're excited to work with them. All right. James, how much money did you guys have to pay G2 to get that top

of the pyramid ranking? It's a beautiful grid. I look at it every day and I hope the whole software

market does too. And that's what we call a very polite dodged by your friends. Anyways, Spencer, James,

thank you so much for coming on. And Spencer, uh, just drop for people Amplitude's, um, website,

socials. Amplitude. com or follow me on Twitter at Spencer skates. Critically, Spencer with an S. Yes.

If you try to see, you will not find him because Twitter search is garbage. But guys, thank you so much. Congrats on the deal. And Spencer, I will be tuned in to all your earnings reports as always. And I'll talk to you in early January. Sounds great. Alex. Thank you so much. Awesome. Appreciate you all. Bye, everybody. Bye.