Alright, everybody. Welcome to the second day of KeyBank's 26th Annual Technology Leadership Forum. My name is Jackson Ader. I'm the Enterprise Software Analyst here at KeyBank. We are thrilled to have Andrew Casey, CFO of Amplitude, join us for this fireside chat. We're going to let Andrew introduce himself and the company for just a couple of minutes before getting into questions. I will come to you guys, the audience, a couple of times before the end. That way we don't, you know, get to the very end and try and squeeze some things in. Andrew, if you want to go ahead and introduce yourself and the company, that'd be great.
Thank you, Jackson. So, Andrew Casey, I'm the CFO of Amplitude, as Jackson mentioned. Amplitude is a digital analytics company that got its start by understanding that any company that was trying to digitally engage with their clients needed some fashion to instrument and understand what those digital engagements were. That could be through a website, through a mobile application, or any method really in which you're trying to understand how best to reach out to your customers via promotional campaigns, marketing campaigns, loyalty programs. At its foundation, it was focused on how to help other developers create great digital experiences through those applications. When we got started, most of the companies we were appealing to were startups, SMBs, digital native clients, and they had a high propensity for B2C related engagement.
Think about DoorDash or PayPal or others in that ilk that were trying to create digital experiences and just intermediate more classic enterprises. Amplitude became the instrumentation for it. I know you've got a question later, but as you start thinking about AI related applications, you can imagine that Amplitude is increasingly one of the core pieces of an AI technology stack because they're trying to do what? They're trying to create more digital experiences and applications, and they want to make sure that they're getting the right feedback on it. Now, fast forward, we went through IPO in 2021, one basic use case, which was product analytics for these types of companies. We started to see, at the beginning, there are a lot of other applications being created around experimentation, session replay, web analytics, marketing analytics use cases, guides and surveys.
There's this long tail of applications being created around analytics hubs. Our founders basically said, look, a lot of these capabilities, they shouldn't be adjacent to analytics. They should be a core part of our platform. We started on this journey of creating multiple different products. Started with experimentation, moved into CDP , which we call Activation now, around how you take cohorts of information and transfer those cohorts into other application environments like a Braze or a Klaviyo or other. We moved into web analytics to take on Google directly on websites because websites were, frankly, archaic technology, which was screen scrapes and sessions as opposed to deep event-based information. We moved into session replay, which gave us qualitative feedback. We acquired a company called Command AI that gave us Guides and Surveys.
Most recently, we acquired a company called Kraftful, which is voice of the customer, which moves, you know, when our minds move to the next stage of the evolution of how you do surveys and customer sentiment in a more logical feedback way as opposed to asking people for feedback. The platform build out was a really big inflection point for Amplitude. As we saw more macroeconomic conditions, and I'll say some self-inflicted wounds causing headwinds for us and our growth rate post IPO, it was this cross-sell, more platform-oriented focus that started to bring us back and started to drive a real cross-sell part of our business, which is increasingly driving acceleration in our business.
At about the same time when the platform discussion was happening, we also started investing in a go-to-market that was more focused on enterprise clients, classic enterprises that were increasingly trying to figure out how they take their more, I'll say, suboptimized or archaic methods of interacting with people and getting to a more digital way of interacting. This last quarter, we had First American Title, which was, you know, as you know, they're more in the real estate business, but they're trying to figure out how they digitize their interactions with clients, move away from a paper-based framework. They're using Amplitude to actually understand how best to architect a lot of that transition. Two major things that we invested in are now starting to show real progress.
That's the platform actually building out all these applications and giving classic enterprises a value for the money solution where you can consolidate somewhere between five and eight applications that they may be using, displace them, and show a much better integrated environment and drive optimizations, as well as a go-to-market focus, which was targeting enterprise clients as opposed to SMB mid-market. That shift from a very transactional model of selling to one that's more value-oriented. That's a little bit about Amplitude.
Yeah, that's great. A perfect overview of kind of where you started, where you are now. I think it's worth noting, you know, Andrew, you mentioned acceleration.
Yeah.
You are one of the few software companies out there that is actually accelerating your top line growth rate. Now, what you mentioned about the platform and going to enterprise and all these things, those seem like more long-term trends. What about the last three or four quarters has driven the acceleration in the top line of the business?
I think it was really about the idea of what needed to be done when I joined about a year ago. I would say there were gaps in the ability to execute.
Is that the stuff where you said some self-inflicted? I'm sorry to interrupt.
No, it's all right.
We will go.
I think that's true. Part of it is, you know, how, so maybe I'll start with kind of my journey and then that'll probably give context. I've been in the Valley for 30 years. I started my career in the Valley, at least at Sun Microsystems, went through lots of different rotations at Sun, did investor relations, but probably the time I enjoyed the most was when I was interfacing with the go-to-market teams and helping better to make them successful and build out our growth. It certainly was at a great time when Sun was really getting noticed and we were the dot in .com and Java was really hot.
I got to a point where I wasn't getting the experiences I wanted and Sun wasn't transitioning the way I thought we should into a more software business versus a hardware business, which was a really hard thing to do. I went to Oracle, ran corporate finance, did 37 acquisitions in a two and a half year period. A lot of corporate finance, corporate development work. Didn't really appreciate the way Larry liked to run the business. It was a very difficult culture where I was younger in the roles, but I was making big progress and I felt like I should be getting more accolades for that and doing all the acquisitions we did. I decided to go to Symantec, which had just acquired Veritas. They didn't really know how to integrate Veritas effectively. I helped them with that, build out the enterprise business there.
We were about ready to go through a transformation from perpetual to SaaS-based, believe it or not. The board and the executive team didn't really have the stomach for making that transition. I tell Mark and Shantanu all the time that they beat us to the transition over at Adobe, but I got very frustrated with that and then decided after four years at Symantec that I was going to move over to another software business and try to make improvements there. It was at HP and they wanted to invest more in software. They hired me for that, but then they acquired Autonomy and kind of put my role on hold and they said, you know, we acquired EDS and we really don't know how to run that and integrate it effectively. It's losing money. Can you help us with that?
I thought about doing a transformation of services business, which I knew a little bit about, but not, you know, not a $30 billion IT outsourcing business. But believe it or not, one of the principals over at EDS and I took the reins and drove that to major improvements to a point where it was no longer losing money. It was actually making money, and we sold it off to CSC. I kind of looked at my career and said, I don't want to be Cathy Lesjak, who was the CFO of HP at the time. I wanted to go to do something that was a startup. I talked to a bunch of my VC friends and I pinned on one that they said, look, this one's already public, but they need an operational finance exec like you to help them scale in the enterprise.
What do you think about going to talking to them? I said, sure. Who is it? They said, ServiceNow. Never heard of them. You know, they're about $400 million at the time. I met with Dave Schneider, Mike Scarpelli, and Frank Slootman, and thought it was a great fit, but it was a business that was still very immature. I mean, didn't have a functioning comp plan structure, no territory focus, no enterprise partnership frameworks. There were a lot of pieces that needed to be developed. I joined in 2014 and, over a six-year period, we scaled that business. I'd probably still be there if they would have offered me the CFO job. When Mike and Frank left to Snowflake, John Donahoe, who was the CEO at the time, called me and asked me if I wanted to be considered for the CFO job. I said I did.
Went through a great process. John told me that the board really wanted somebody who was already a CFO. They chose Gina, and Gina's a great CFO. John asked me to stay to help Gina build out the business. I told him probably a week or two later, I really couldn't because it didn't matter whether Gina was going to retire in five years, I would still face the same headwinds and would be unqualified by their characteristics. I wasn't the CFO. That kind of serendipitously led me to WalkMe. I knew about WalkMe. We used them at ServiceNow and took them public in 2021. We went through some difficult times. I'd say that we, between myself and the CEO, had a number of disagreements on where the business should be going.
When Frank and Mike and Dave, principals that I worked for at ServiceNow, told me about Lacework and it needed to be turned around, I went there. I partnered with Jay Parikh , who's now head of Microsoft's AI group. We turned around Lacework, got in a great position. What happens when you turn around a company and get in a great position? You start attracting strategics. Despite our desire to continue to drive the changes, the board disagreed and had us go through a process, which ultimately led to the sale to Fortinet. At that point in time, I'd also started looking around and Thomas Hansen, who's our CEO and President, wanted a CFO partner to help him build out the go-to-market team and convinced me that I should join Amplitude along with Spenser. I joined Amplitude on August 1st, 2024.
The funny story I tell people is that I was signing a definitive agreement to sell Lacework to Fortinet at 12:00 P.M. on August 1st. I joined Amplitude two hours later.
Wow. If we zoom out on Amplitude, the macro, you know, view, do you see a difference between IT budgets and marketing budgets in terms of their demand for the Amplitude services? Is one of those driving more of the growth than the other?
Yeah, I definitely see that there's, I mean, IT is typically more, how can you drive more with less? How can you drive greater productivity across the business? There's certain aspects of Amplitude being a part of that for sure. I think that where you see a much larger investment in growth for businesses still comes from marketing or from sales-related investments. Where Amplitude is increasingly showing this very holistic view of how you're engaging with all your channels, not just your classic channels, but now your digital ones as well. You find more and more CMO and Chief Data Analysts or Chief Technology Officers looking for ways to have a more comprehensive view of what their investments are yielding. You can pick apart different types of investments: demand gen, sales-led demand, promotionals, overall marketing spend for awareness.
All those things come back to, hey, I'm expecting that to actually generate greater revenue. If you can instrument and understand that better based upon the interactions that your customers are having with those campaigns, you get more and more effective, better and better returns on it. I do see that as a big catalyst for Amplitude. I would tell you, we're just scratching the surface.
Does that bring you into competition with different people than who you would compete against just for product analytics?
Yes, for sure. I mean, we're now more head-to-head with Adobe and Google Analytics than we ever have in the past.
How are your win rates against those competitors?
I'd say increasing because we're increasingly seeing more of those conversions from Adobe, where forward CMOs are understanding that they want that comprehensive viewpoint.
If I think about the suite of products now, you know, that you have, you've expanded, we've talked about going just from IT into marketing, from product analytics into some marketing analytics. You mentioned CDP, all these different things on the platform. You have, I think, an exciting, there is an exciting AI story to tell that, you know, frankly, some other companies, either in and around our coverage, don't always have, like, there is no obvious reason why AI should be embedded into their products.
Yeah.
I think you're, you know, having an agent do the A/B testing, you know, for the digital product analytics makes all the sense in the world. Can you give us an update on the timeline of when you expect those agents to come in? Give us background on what it is and then what's the timeline for that to be kind of rolled out to general availability.
Yeah, and I think your postulate is right. One thing I'd ground it on too is one of the things that we see with more classic enterprises that don't really have a digital channel set up yet is they've struggled with how best to begin. Do they have the resources? Do they have the data scientists? Do they have the know-how to actually instrument this appropriately? Maybe they're not convinced yet that they need to go to a big IT outsourcer or a system integrator in order to do it for them because they don't understand the steps they need to take. That impediment is increasingly falling with Agentic. Now we'll get into why we're so well positioned for it because what happens is you start to reduce those barriers.
You don't need the same amount of resources in order to adopt Amplitude's platform because you can deploy agents to actually help you with that adoption and that implementation and that experimentation over and over. We've developed a set of products that are labeled to do experiment and activation, which is us taking cohorts of behavioral heuristics and saying, all right, this cohort of customers, you should target them with this email campaign. You should target them with this promotion. You think about last-mile marketing companies like Braze or Klaviyo or others. They're actually leveraging Amplitude with their customers' implementations to showcase their effectiveness at driving those end actions. One of the things that we start to go into these environments where there is, it's just white space.
We're starting to find that the notion of adding agents that help with tagging of information on the websites, auto-capture of that setup, making sure that they're taking what you do have potentially in digital assets and giving you recommendations how you can improve it, gets them on the path of quickly improving that digital engagement faster. Whereas before, to do experimentation, which if you don't know, it's like you take your website and you say, if I did a different font, I did a different dropdown, I did different coloring, if I showcase the promotion difference, is that going to increase how you're monetizing conversion rates? Think of it as like a retail or a fast casual restaurant or any others that are digitally engaging through mobile applications. They're doing these testing, but they're doing them in a very methodical way.
You know, it takes like, okay, the data scientist says this, we talk about the business application, we get some returns and voice of the feedback from customers, but you can deploy an agent to do that. They can do thousands of them. They can probably actually surface other experiments that you didn't even realize and give you better customer sentiment and feedback of the changes you need to make. By the way, it starts to change how you have your advertising campaigns to be less shotgun approach to a large cohort, start to drive more and more personalization of what those interactions should be, because those experiments can happen in real time.
If you make them autonomous, then the experience that you might have through a mobile application versus one that I might have would be catered to us, not to our cohorts where we have alignment, but more personalized to it. This area of Agentic is enabling customers who haven't made traditional investments in data scientists or digital channels of engagement to reduce the barriers because they can just deploy the agent and not need that army of digital scientists to go actually do this work.
Is that the reason why, was it a mortgage company you mentioned at the first, the real estate guy? Is that the reason why they are able to kind of go digital now? Or are agents still on the come for them?
I would say they were less forward thinking on, agents are still on more of the come for them as the answer. The inspiration for them was more about how do they better digitally engage with their clients through a website. They discovered us through web analytics. They started to understand, wait a minute, we can do more with this even in the mobile applications associated with our third parties as they're used. They use DocuSign associated with signing real estate agreements. Now they can gather that information as well through the DocuSign implementation within their application environments to understand, wait a minute, is this the best way to represent documents to be signed, for instance? I've just given you one use case.
That whole journey, then they started mapping out, wait a minute, you can help us with qualitative feedback as well, so we can see it through session replay. Oh, wait a minute, you can actually pipe in voice of the customer information so we can get not just the application environment, but get feedback on what that application environment is. Suddenly they look at this as a really more comprehensive journey for them into digital engagement and less about the product and what the specific feeds are, but rather how do they achieve their outcomes?
How are you going to price AI?
I think the first thing you need to understand about Amplitude is we're not like a SaaS seat-based model. We have two mechanisms through which.
Seat-based is a four-letter word in SaaS these days.
It is.
That's smart of you.
Yeah, it was back to this notion that we were disintermediating more web analytics companies that were doing these screen scrapes or sessions that are very rudimentary in how to get feedback on digital interactions. We went down to an event level and started to measure the volume of events that are being ingested in the platform. One of our primary meters is the number of million events that happen ingested in the platform. I think the more channels that customers open up, the more data they're actually downloading into the platform. That's the way we price. We price per million events. Over time, the amount of data that customers are ingesting is increasing dramatically. It's because people are spending more time on the platform, they're spending more engagement, there are more channels.
For us, that was just a natural way of using a meter as a proxy for value for what customers are getting out of the platform. That's the first thing. The second thing is, as we've added more products and we've made them work cohesively together, our platform cross-sell story has been very powerful. That gets augmented when you add an agent because now the agent shows you how those workflows are optimized when they're all together. Not only will agents, as we implement them, drive increases in data ingestion, it should also, no pun intended, amplify our cross-sell capabilities. We're not thinking in terms of putting pricing barriers on the adoption of AI. We think we've already got the monetization mechanisms built within our platform. As customers get more and more comfortable with using AI across our evolving pricing and packaging, they'll see greater value.
As that greater value gets realized, we'll accrue it as well.
We got to follow up on, you know, what website traffic is going to look like in the world of AI search. I think Matt had a question.
Yeah, go ahead.
I was going to ask you to define what the event was on the meter.
How do you define an event when you talk about the millions of events?
It's any type of interaction that occurs with a person or a system that is actually engaging within your environment, your mobile application, your kiosk, your wearable, your website, any interaction you might do. Everything down to a cursor movement, to a reaction, to the time you spent. It's very, very detailed. For instance, The Economist , when they went from print to digital, one of the key things they really wanted to know was the difference between browsing and consuming the article. If they could tell the difference between browsing versus consuming, then they could actually cater and curate what articles they're going to use, as well as drive a good, better, best pricing mechanism on their advertising. It gets in a very, very detailed level.
We've had many customers who've got lots of different channels and ingesting, and some customers are actually at trillions of events that they're ingesting within the platform.
Question over here.
Yeah, you're going to be increasingly bumping up against Adobe and Salesforce in this Agentic mission. How do you think about doing that? How do you think about, like, does that change anything internally when we go to market or?
Absolutely. I think it's, we wouldn't be able to do that without the transformation of our go-to-market team to be more enterprise-focused. There's a lot in that. There's a lot in that. Everything from moving through a transactional model to an enterprise model, you got to start with new people. You got to re-architect your sales process, top of funnel, what the stages are, what the criteria are, how you fulfill that, what the value propositions are, how you represent that to customers, how you move from a speeds and feeds discussion to an outcome-based discussion. I would say that journey sometimes takes years. Fortunately, Thomas preceded me and started that process. When I joined, I would say I gave it the operational pixie dust, if you will.
I started pushing harder on some of the things that I've learned in the past about how to re-architect, to align the sales incentive to the outcomes you want, how to drive better accountability for the sales rep at the territory level, how to make sure your territory planning is set up so that reps that are first joining actually have an existing customer to talk to, so they learn what the value propositions are and can have potential expansions within their territory. I'll give you a host of others if you like, but some of these things take a while. The piece, I wasn't curious on what we needed to do. I knew what we needed to go do. What I didn't know was the rate of absorption. I don't know how fast we can go sometimes.
Sometimes you have to test that a little bit and figure it out. I would tell you some of the things that are showing progress, cross-sell, for instance, contract duration, which results in RPO growth, both long-term and current, because we're moving more to annual payments in advance. All those things are instrumented changes that we put in place. When I joined, those showed up in contract structures, and they showed up in the comp plan structures. I love reps. I love enterprise reps because they're coin-operated, and I know what to expect from them. If I put the comp plan in the right way, I'll drive the right behaviors, which drives the right outcomes. That's an easy way of saying you got to put more rigor in the process. I'd say that rigor is increasing, and we're doing better and better.
We still have a little ways to go to my level of satisfaction.
Won't events come down if Google becomes an answer engine instead of a search engine? Won't less people go to The Economist ?
I don't think so. I think the content rules. I think data rules. I think that increasingly, we've even built our product where marketing analytics analysts can look at the progress of their campaigns through LLM searches. I think that the mechanisms, you know, augment a little bit, but I still think that you're going to find that people will use some mechanism to find the proprietor of the service or the product. Our application at its foundations is that how do we help customers improve the engagements with their customers? I think that evolves in a positive way for Amplitude, even with the Agentic implementations over time.
All right, I'll get you out of here on this. The margin, you know, I think right now you're accelerating the top line. That's kind of the focus. I totally understand that. Where should we be thinking about that incremental margin? Are you building in sustainable practices to be a much higher margin-generating company of the future?
Yeah, both gross and operating, I would say. Both gross and operating. First and foremost, you need to know on the gross margin side, the biggest cost we have is for hosting our own applications within cloud providers. We use AWS. Over time, we'll probably get to a multi-cloud provider. How well we're optimized for our application in those environments is a cost, and the engineering teams are constantly working on that. We'll drive our marginal income on cost of data ingestion down. We'll share some of that with our clients so that they don't have a linear progression in costs as they ingest more data. At the same time, I think that'll be one mechanism driving better gross margins. The other thing is the analytics-related first land is the highest cost for us.
After that, as customers add more of our modules, they don't have necessarily marginal incremental data ingestion. They're paying more, but there's not more data. That means that you can see yourself, wait a minute, Guides and Surveys, that's almost 100% gross margin. The more we have customers on multi-product, the better our gross margin is going to be. The other thing I would tell you is that we're going to build a services business, which is a little bit of the headwind we have right now. It's not because I want to build a services business. It's because I want to go build a partner ecosystem whereby more and more customers are building, and frankly, GSIs ultimately are building use cases and applications upon our platform. In order to do that, you have to prime the pump a little bit.
You got to get to services being generated and giving people insights on how they can make money. We've made some investments on that front. That was a little bit of a headwind for Q2. Over time, the resources we have, they'll start generating billable utilization, which will help to offset some of those costs. Now, the last thing I'll say on operating is.
Quickly on the operating margin. We only have two minutes left.
Okay, on operating, I promote a growth with leverage framework for our budgeting and our focus. We're going to grow revenues, but we're going to grow expenses at a lower rate. That means that I'm putting sales efficiency targets on our team. As a percentage of revenue, we're, I think, around 43% now on sales in market. That's got to be like 10 points lower, if not, you know, low 30s. G&A, when I first joined, was 17%. We're down to 13%. We're going to try and drive that down below 10%. On R&D, I'm going to try and keep it right around that 18% - 20% range. We'll make investments, but we're going to do it at a rate which is going to enable us to do that growth with leverage.
Awesome. All right. Thank you, Andrew.
Sure.
Thanks, everybody.
This is great.
Thank you very much.
Yep.
Great job.
Yeah, thank you.
Awesome.