Amplitude, Inc. (AMPL) Earnings Call Transcript & Summary

November 4, 2026

NASDAQ US Information Technology Software earnings 61 min

Earnings Call Speaker Segments

Spenser Skates

executive
#1

When it comes to quarterly and long-term growth, investments, and what lies ahead for our business overall, these forward-looking statements rest on information, assumptions, and expectations we hold today, and they carry risks and uncertainties—several of which we cannot control—that might make actual results differ materially from what those statements describe. More detail about the risks that could make actual results differ appears in the filings we make with the Securities and Exchange Commission. We caution you against placing undue reliance on these forward-looking statements, and we take on no obligation to update them after today's call, except as the law requires. Some financial measures discussed on today's call are presented on a non-GAAP basis. We use these non-GAAP financial measures internally so that analyzing our financial and business trends is easier, and for internal planning and forecasting purposes. These non-GAAP financial measures have limitations < > and shouldn't be relied on alone or as a replacement. Today I'll go over three things. First, our Q3 results. Second, how we've turned Amplitude into an AI company, and why every company I speak with wants to know how they can do the same. Third, a look at our product and a spotlight on our customers. Let me begin with the numbers. Q3 revenue came in at $111 million, which is up 25% year-over-year. Annual recurring revenue came in at $431 million for the year, marking 24% growth compared with a year earlier and a $21 million gain over the prior quarter. On a non-GAAP basis, operating income reached $5.7 million — the most profitable quarter of that kind we've ever posted. The count of customers spending above $100,000 in ARR rose to 855, which is 31% higher than a year ago. As enterprises pursue AI transformation, they are looking to us as a partner they can trust, and they are taking up a wider range of our products. We're expanding ARR faster than at any point since 2023, even as we deliver record profitability. Over the past two years, long-running agents have reshaped coding in a dramatic way. Output is surging, with the volume of code produced climbing 6x annually. The amount of work AI can take on is increasing sevenfold annually. We're riding an extraordinary curve of exponential growth. Still, each CIO I speak with shares one common desire. They tell me they're burning through enormous sums on tokens, yet they wonder whether that spending is producing the outcomes they actually want. That's precisely why Amplitude matters when companies bring AI into their operations. Making more software is no longer the constraint. What matters now is figuring out which things to build and determining if they deliver results. That sits at the heart of our business. We assist our customers in creating superior products. In my conversations with executives, every one of them seeks our support to make AI transformation succeed within their organization. Each of those individuals stands at a distinct stage along that path. Starting on the left side, you'll see companies relying on older, disconnected point solutions; over on the right are the most advanced firms, using coding agents and context agents such as Wave to power self-improving products. The majority of companies today remain on the left, within the first 2 categories. That gives us a major opportunity to help businesses replace legacy technology providers—companies that were assembled through mergers and acquisitions over 10 years ago. We meet customers wherever they are, help them move their operations forward to the next stage, and because of that, we're seeing wins across the full life cycle. Our platform is built to guide customers from whatever point they're currently at through an AI transformation. Amplitude provides customers with the most profound insight into how users engage with their product. Increasingly, our agents carry out that discovery on their behalf. Statsig offers customers feature flagging and experimentation, constructed atop the globe's most sophisticated stats engine with an engineering-first perspective. Our customers are now able to test and release whichever features produce the best result. Wave represents the future of product development—self-improving products where we automatically suggest what customers ought to build next, drawing on signals from users. Although we're at an early stage here, I'm really eager to show you a demo today. All three of these products, when you put them together, close the loop on product development: you understand what's going on, you measure what you ship, and you ship what actually matters. That cycle is how AI native companies operate, and it's where every enterprise is trying to get its product development process to land. Let's start with Statsig. Six months in now with Statsig and with our customers, and things are going better than I expected. We've picked up the pace of development on the platform and continue to ship enhancements to what is already a best-in-class experimentation platform. In this new era of building, experimentation is important because it can demonstrate what is working. Paired with best-in-class analytics, it can close the loop on self-improving products. That's the strength of knowing the context behind what works and having a platform on which to test those outcomes; we're collaborating tightly with data warehouses to open up experimentation and analytics capabilities. I'm also very happy to have Statsig's founder and CEO, Vijaye Raji, join our Board. Vijaye started Statsig and guided it through its acquisition by OpenAI, and he now serves as CTO of Applications at OpenAI. The histories of Statsig and Amplitude are closely connected. Amplitude originally drew inspiration from Meta's growth team and their finding that a new user making 7 friends in 10 days strongly predicts long-term engagement. Back then, we felt certain that all companies would require the same kinds of insights, which is why we founded Amplitude in 2012. While at Meta, Vijaye witnessed directly how effective its internal tools were—Deltoid among them—in allowing teams to gauge how much their experiments moved the needle. The modern data science and experimentation stack that the industry now treats as standard practice was largely pioneered by Meta. Vijaye departed Meta in 2021, aiming to deliver those same capabilities worldwide via Statsig. For several years, the two of us were rivals. Each of us took an engineering-led approach, collaborating closely with customers and shaping our products around what they needed. Amplitude concentrated on analytics at its core, while Statsig concentrated on experimentation. Both of us were working to make sophisticated measurement tools accessible, and we challenged one another to improve. Earlier this year, Vijaye was looking for the ideal home for Statsig's product, brand, and customers. We came to the conclusion that teaming up with Amplitude was the best route for Statsig's journey to continue. I'm eager to work with someone who has thought so deeply about the same mission. It's a privilege to keep pursuing that work alongside Vijaye as he joins our Board of Directors. Next, I'd like to talk about a few of the more recent products we've launched. First up, agent analytics. Each company is attempting to figure out how to put AI agents to work serving their customers, yet they frequently find themselves stuck wondering whether those agents are functioning. The majority of observability tools can report if an agent is operating and how many tokens it's using up; what they cannot report is whether the responses are actually any good or whether they're enhancing customer outcomes; that is the reason we created agent analytics. Each agent conversation gets scored automatically. Did it finish the task? Was the response solid? Did the user end up frustrated? After that, we tie those scores to everything else we understand about the customer and Amplitude. You gauge your agents in the same manner you gauge the remainder of your product. Agent analytics is all about making agents better.

Spenser Skates

executive
#2

With product teams, you can set up evals, pinpoint the signals that indicate whether experiences are good or bad, and identify where agents reliably succeeded or failed. Those building AI products get the same sort of feedback loop that AI engineers rely on to evaluate and enhance agents, tied both to how users behave and to business results. Let's go ahead and look at how this works. Here is a list of every agent session for SF Bay Auto, a local car dealership. That agent is the chatbot on their homepage. I can click into any individual session and view standard AI observability features, including cost, token usage and duration. Agent analytics additionally ships with prebuilt evaluations called signals, which assess things like whether a task was completed, how helpful a response was, and whether a session was safe. Customers can build custom evals suited to their own business requirements. Beyond that, agent analytics lets you connect agent behavior with business results like conversion and retention. In this view, we're looking at the conversion rate for visitors who came to the page, engaged with an agent, and later submitted a price inquiry on a vehicle. We can put those who engaged with an agent side by side with those who didn't, and observe how that engagement influenced inquiry rates. We can also examine how the choice of different models changes the rates at which people submit an inquiry. That type of insight is essential when it comes to putting agents into service, keeping them running well, and making them better, since they will serve as the primary entry point for customers going forward. Agent analytics became accessible to every customer in August. Within barely more than four weeks, over 125 customers had adopted it. We assisted HYBRD, an AI-driven fitness coach, in linking its agent brain with the remainder of its customer journey data, enabling it to determine whether that brain agent was of benefit to their athletes. HYBRD created an evaluator using Amplitude's global agent to identify conversations where a user had to request a tool run a second time, and to gain a clearer picture of where the agent was falling short; this revealed that during the previous week, frustrated users were the outcome in 25% of conversations. With that insight, the HYBRD team adjusted the prompts for the Brain agent and delivered a fix inside one day, and the rate fell substantially; they further found that athletes interacting with the Brain agent retained at 4x the level of those who did not. This came to light by linking agent conversations to retention and adoption metrics within Amplitude. Their current priority is putting resources toward improved onboarding for the Brain agent and expanding capabilities that keep athletes engaged. Only Amplitude can deliver this, thanks to the combination of our platform and agent understanding. Agent analytics is now being deployed by us at scale. Millions of agent conversations are being run through it right now by one of our biggest customers, who use it to refine their agents and boost conversion funnels at scale. We also apply it internally. An open weight model was tested in our own global agent using agent analytics. We found the failure point, fixed it, and today achieve quality on par with the leading frontier model at about one-third the cost. Moving on to the product highlight for today. I'm truly thrilled to present Wave. Wave is our agentic product team. It works continuously, reviewing all the information Amplitude holds about product usage — analytics, session replays, experiments, feedback, agent conversation, and more. It spots opportunities that merit action, lays out the supporting evidence, and suggests a modification. It then passes that modification along to a person or a coding agent for deployment, and later it assesses whether the modification delivered results. Wave informs you about what is worth developing and whether it affected the business, since it possesses complete context from the remainder of the user journey. This elevates what Amplitude has always provided: assisting teams in grasping user actions and choosing what to develop next. Wave turns that entire workflow proactive—drawing on the behavioral data you've gathered plus product best practices, it automatically surfaces the right opportunities and actions for product teams. Software factories are becoming an ever larger piece of how software gets developed. Coding agent fleets continuously write and ship code, yet they're missing a context layer telling them what should be built. That product behavior context layer is Wave. Here's a concrete example of how we applied Wave to enhance Amplitude itself. Wave is always seeking ways to improve every part of Amplitude, and that includes our blog. In this case, it spotted an issue. Every week our blog draws thousands upon thousands of views, yet the rate at which readers convert to signing up for Amplitude after finishing a post is artificially low. It recognizes that a topic-matched call to action should be placed within the blog post's text, prompting the user to sign up. For instance, when a post covers churn, it says, sign up with Amplitude to do a churn analysis now. Wave worked this out entirely on its own, with no human involvement. It generated a visual representation of the issue, suggested a revamped approach, and then it tells us exactly what to do in practice: incorporate calls to action tied to the subject of the article and offer readers a helpful follow-up step. At launch, it also assisted us in drafting the PR and setting up the experiment. Once it went live, we recorded a rise of 45 additional user sign-ups each week. For us, that is enormous. Picture hundreds of such enhancements occurring throughout the company's product and website continuously. I'd like to give you an illustration involving one of our customers, Mutant. Mutant operates an AI image generation platform. Wave was brought in, and it found that people who post their art are far likelier to remain active—yet posting itself was purely a matter of self-discovery. Nobody was tracking which users published and which didn't. Wave proposed a prompt with tool tips to encourage publishing, plus event instrumentation to see who posted versus who didn't. That simple change was shipped by the team. Engagement with the feature went up 76%, the actual publish rate rose 10%, and downstream day 1 retention then gained 3%—a huge impact from a very small tweak. Next, a few customer highlights. We had a solid quarter in both new lands and expansions. Right now, 855 customers are paying us over $100,000. During this quarter, we either brought on board or expanded ties with names like T-Mobile, Cy, Fireworks AI, a top British airline, F5 Networks, SoFi, Hilton Resorts, Coursera, and Adidas. We continue to replace standalone tools and win business away from older established players. Amplitude is the choice of today's most advanced companies; we've reworked our organization, our offerings, and our operations to be AI native, and we're shaping what data will be able to do going forward. Wave, our offering for self-improving products, arrives next week. The pace of our innovation continues to accelerate, and we're building in a way that scales with leverage. I couldn't be more excited about the road ahead. With that, I'll pass it to Andrew to take you through the financial results.

Andrew Casey

executive
#3

In our fiscal third quarter, we continued to execute our strategy: selling into the enterprise, extending our platform's reach, and introducing new products that provide exceptional value to our customers. Revenue rose 25%, exceeding the top end of our guidance, and we achieved a 5% non-GAAP operating margin — the highest level we have ever posted. That, by definition, is growth with leverage. We continue investing ahead of our growth, while simultaneously observing improving trends in platform adoption. An increasing number of our customers access our platform via MCP and headless as well as through our traditional UI. Ingestion of data is climbing, and our utilization rates are at record highs. We think we're well positioned to fuel expansion and enlarge our customer base. Today that manifests as a higher cost to serve, since demand shows up before potential monetization. We're dedicated to expanding while also improving our operating leverage. Over the last 2 years, we've concentrated on trimming costs, boosting efficiency levels, and fostering greater accountability across the organization. We accomplished this by flattening management layers and bringing our teams into line with the value proposition we deliver to customers. The outcome was reduced costs as a percentage of revenue even while we continued investing in innovation. As we look ahead, there are chances to support both existing and prospective customers throughout their lifecycle, as their experience turns AI native and as they attend to their own clientele. We can grow our enterprise client roster and broaden multi-product adoption among customers we already have. Beyond that, we're adding new offerings that we can generate revenue from, including agent analytics and Wave, which bring in fresh budgets and distribution channels. I think we're well positioned to continue executing our growth with leverage strategy through 2027. Turning now to our Q3 results. Just a reminder: apart from revenue, all the financial figures I'll cover are non-GAAP. You can find our GAAP results, along with a reconciliation of GAAP to non-GAAP figures, in the earnings press release and supplemental financials posted on the Investor Relations section of our website. Revenue for the third quarter was $111 million, which represents 25% growth year-over-year and 10% growth quarter-over-quarter. As we closed the third quarter, total ARR reached $431 million, up 24% year-over-year and $21 million on a sequential basis. We estimate net new ARR of $21 million was made up of $15 million from Amplitude and $6 million from Statsig; this breakdown will get harder to track as we combine customer solutions and build derivative products from the Statsig intellectual property. Total remaining performance obligations climbed 27% year-over-year, reaching $497 million. Current RPO rose 28% from a year earlier, while long-term RPO increased 24% year-over-year. Enterprise new and expansion deals both had a strong quarter, and platform sales were especially robust once again. Half of our customers now own more than one product, and 81% of our ARR is generated by that group. Accounts using at least 5 products now account for over 29% of our ARR. On a pro forma basis, in-period net dollar retention was 105%, fueled by cross-sell expansions across our customer base; that pro forma number reflects both Statsig and Amplitude customers. During Q3, gross margin landed at 70%, which is about six points below the same period a year ago and one point down from the prior quarter. Several factors drove that: sustained expansion in data ingestion across our platform, inference expenses tied to greater customer adoption of our AI tools, and bringing Statsig into the business along with its hosted environment. Ingestion volumes remain robust, currently exceeding 90% relative to contract entitlement, which positions us well for expansion and upsell opportunities in the quarters ahead. Sales and marketing spend represented 34% of revenue, a drop from 43% — nine points — in last year's third quarter, reflecting our continued efforts to improve efficiency in sales productivity. G&A came in at 11% of revenue, which is 2 points below where it was in the third quarter of last year. R&D represented 20% of revenue, roughly 1 point higher than the third quarter of last year. Our total operating expenses reached $71 million, equal to 65% of revenue, versus 75% in the third quarter of last year. Operating income totaled $5.7 million, or 5% of revenue — the highest quarterly non-GAAP profitability the company has ever recorded. Net income per share came to $0.04 on 138.4 million diluted shares, compared with $0.02 a year ago. Free cash flow for the quarter was $5.8 million, or 5% of revenue, against $3.4 million, or 4% of revenue, in the same period last year. At the end of the quarter, we had about $155 million in cash and investments. We believe strongly in the long-term value of our platform, and we've used our cash—and intend to keep using it—to reduce dilution's effects when it makes sense. Our balance sheet is still in a strong position, which lets us take a more aggressive approach to M&A so we can speed up our R&D road map when opportunities come along. Now let's move to our outlook. As a reminder, the way we approach setting guidance is through the lens of execution. We're happy with how we're doing at consolidating point solutions onto our core platform, and with the uptake of our various AI technologies. Our business and selling process now have instrumentation in place, which makes it simpler for customers to consume more of our platform. We perceive an opportunity to keep generating growth in 2027, and we are putting investment behind that opportunity ahead of time while staying committed to growth with leverage. Looking at 2026's fourth quarter, our revenue expectation falls in the range of $110.8 million to $113.2 million, which works out to a 23% annual growth rate at the midpoint. Non-GAAP operating income should land between $7 million and $9 million. Non-GAAP net income per share is expected to come in between $0.05 and $0.06, based on a fully diluted weighted average share count of roughly 142.5 million. When it comes to the full year of 2026, we're increasing our full-year revenue outlook — reflecting how the third quarter went — to a range of $415.8 million to $418.2 million, which works out to 21% annual growth at the midpoint. We're likewise lifting our forecast for full-year non-GAAP operating income, thanks to the third-quarter performance along with the steps we took during the first half, to somewhere between $8.9 million and $10.9 million. Non-GAAP net income per share should land in the range of $0.07 to $0.09, on the assumption of roughly 139 million weighted average shares outstanding on a fully diluted basis. To wrap up, the pace at which we innovate is speeding up, and the value we're able to deliver to customers keeps expanding. We feel confident in our capacity to scale a business that is both durable and growing while at the same time achieving non-GAAP profitability. With that said, we'll turn things over for Q&A. John, over to you.

John Streppa

executive
#4

We're moving into the Q&A portion now. Taking our initial question from Scott Berg at Needham, and after that, Mark Cash.

Scott Berg

analyst
#5

Truly a strong quarter. There are many things to cover, but let's begin with... Spenser, all this innovation — the speed has just been incredible. You've got a number of new offerings rolling out. You brought up Wave, agent analytics was just replaced; but how — I suppose, when you talk to customers, how quickly are they able to drink from this fire hose, this stream of product innovation you've got? Can they actually move through a product evaluation process in a sensible amount of time, given how much is coming at them, and then truly adopt it and drive the right value? It's remarkable how fast you're moving; and I just want to know whether your customers feel somewhat overwhelmed because there's so much new stuff to discuss.

Spenser Skates

executive
#6

Scott, that is actually such a really — such an excellent question, because you've put your finger on the key bottleneck. As you say, we can build and launch lots of things very aggressively, and that's fantastic. We're essentially extending our lead versus point solutions and a bunch of the legacy incumbents essentially constantly, so it's great. That gap only keeps widening whenever someone compares us with some of the rest of the market. The bottleneck, though, isn't that innovation anymore. The bottleneck is how quickly a customer can adopt. Now I think there's a common belief, from the executives I speak with, that the exciting part is they want to move as fast as they can. And the big question on their mind is who is the right partner that will guide me through that? Because they can get — there's tons of AI innovation emerging in a lot of different places. And the question they ask themselves is, okay, well, it's great that there are hundreds of companies that can sell me some AI software factory development platform. Who is it — who can I partner with that can take my organization and go through that process successfully? And so that is what I regard as a critical bottleneck to the adoption of AI in the economy; and so for us, we're spending a lot of time on how we help customers get through that maturity curve. That's why I showed the maturity curve, because it's a process. It's not like you go from Stage 1 to 5 overnight. That takes many years at the very least, if not longer. I think the good thing is that no matter where you're at, there's always a next step to be done. So we had a large telecom last quarter that came off of a big legacy MarTech solution and was like, hey, we need to move off this year, like help us do that." — so we sold them on the vision of like, hey, we built Wave, we built self-improving products. We’ve developed agent analytics, and the first step is to migrate analytics and experimentation over. By contrast, some customers we’ve worked with for many years already use a large portion of the platform, and they are ready now for Wave or ready now for agent analytics. So yes, I agree; and I think the companies that figure out how to help the broader economy get through that bottleneck will be the ones that win in the future. Maybe just one last small point I’ll add. Agent analytics has been going very well. I think if someone already has an agent product experience on their website, that plugs in very, very easily and they can get started right away. Whereas something like Wave is a very advanced motion. So that takes a lot more hands-on time and help; but I think these customers want to move fast. And their question is who can I trust most to do that. So we’re talking about something on the order of months or quarters, not years here.

Scott Berg

analyst
#7

Well, you're definitely moving quickly—that much is clear. And Andrew, as a follow-up, you and I have talked several times over the past 6 months about the shift in the company's go-to-market strategy and how you believe the company can drive better efficiencies. My model shows your sales and marketing spend as a percentage of revenue is down—it looks like 12 points across 2 quarters. That's a fairly remarkable change to happen so fast. The question isn't whether you can reach the 30% target, which I think you and I have discussed. It's more whether you have the right capacity to actually meet the needs for all the new products you have out, against a backdrop where demand seems to be accelerating right now.

Andrew Casey

executive
#8

Yes. And it's quite, quite opportune, because right now we're in the middle of our FY '27 planning; and as you suggested, sales productivity has shown great resiliency as we've made big changes in structure and focus on enterprise clients; we've changed territory designs; we've changed pricing and packaging. and it's all working in the right direction. So this is the first time since I've been at Amplitude that, for planning purposes, we're seriously looking at adding capacity in key markets where we see that the current investment just isn't meeting the level of demand; and that's a good problem to have given the changes we've made. So there's still—I still say there's more to go. There are still things we can do better, and we're constantly looking at changing how we can make productivity higher. But the classic way I look at our sales productivity and making investments is, are we getting closer and closer to that magic number? And if so, where are we seeing that? And is pipeline outstripping the ability of the salespeople to actually address it? And if those 2 things are true, then that starts giving you indications you should be making capacity investments.

John Streppa

executive
#9

The next person to ask a question is Mark Cash of Raymond James, and after that we'll hear from YC Wong.

Mark Cash

analyst
#10

Yes. So Spenser, I want to ask — with Wave possibly shifting Amplitude earlier into the development life cycle alongside Teams, sort of deciding what to build. And I realize it's pretty early, with the launch arriving next week, as you just mentioned. But what milestones, I guess, would indicate it's ready for wider adoption? Anything you've observed from the beta program up to now? And might this become a meaningful commercial opportunity in '27? I think perhaps it goes beyond that, since organizations need to sort of rethink things and need to trust this Wave product, right? It's sort of a new concept. I'd appreciate your thoughts on that.

Spenser Skates

executive
#11

Yes. Naturally it's still early days, so I'd like to avoid being too specific about our expectations for it. I'm genuinely enthusiastic, since it takes a task customers currently have to perform themselves and manages much of that work on their behalf. It continuously watches their data, examines their website, and generates opportunities; and just to be clear, it's not completely automated at this point. Quite a bit of effort is still required to set it up. With Wave, we're really pursuing a somewhat different approach. Instead of simply pushing it out to each and every customer, our aim is to go deep and large with a handful where we can be highly hands-on. Regarding Scott's earlier question on AI adoption: the real obstacle isn't some magical new software or technology. What's hard is integrating it into an organization alongside existing processes, workflows, and systems, and training the staff. Since Wave can be very powerful — rather than acting merely as a data pipeline that produces analytics charts or helps run experiments, we can now say: if you take these steps, conversion will go up, retention will go up, and sign-ups will go up. What people will pay for that is roughly ten times what they'd pay for mere data, plumbing, charts, and dashboards. That has always been the goal for Amplitude. Nine years ago I authored a white paper on self-improving products that laid out this vision. So the thrilling thing is that it now exists; yet, as you suggest, adopting it represents a major shift for a company. And the constraint isn't really whether the software performs well enough. The real question is how you fold that into the customer's entire build process and the way they work; so with Wave we're pursuing a somewhat different strategy, going big and deep with a handful of customers. And I hope — I'm excited that in the coming quarters we'll have some exciting results to share about it.

Mark Cash

analyst
#12

Great; and Andrew, may I just build on that point? I'm considering Wave and the headless momentum you're seeing, such as with MCP active users, and I'm curious how you weigh the path of product costs against spending on innovation to ensure customers are aware of these new products and take them up—what different levers do you have for finding efficiencies?

Andrew Casey

executive
#13

Absolutely. Several exist. To begin with, it's quite standard; we need our sales team trained and the marketing message landing with the right customers—the ones we believe are most likely to purchase. It's really about beginning to demonstrate the value customers receive once they've implemented. Customers get quite excited when they see the product working. Only this morning, I spoke with a customer regarding broader adoption of our Statsig product. That was a terrific conversation. Yet as soon as we shifted to discussing Wave, their eyes lit up over the possible gains they can foresee across their entire software development experimentation process, where Wave surfaces the things they can do to improve their applications. So we believe this will truly set our value proposition apart, and we're working to be very clear on what we expect the rollout to look like and to do our utmost to model the expenses and requirements needed to really spread the word.

Spenser Skates

executive
#14

Apologies — we switched rooms for this earnings call, so Andrew and I are at somewhat different heights. That's fine. I'm on my feet, he's seated. No problem at all. We'll have it fixed by next time. Good to see you, YC. Please go ahead.

Yitchuin Wong

analyst
#15

Yeah, maybe just to begin with, some of the security incidents we've seen — definitely a fair number over roughly the last month, with PHP and Node.js as well, including one that Ample discussed last week. Are you noticing any greater scrutiny, like around the way your customers store their data or their usage on the platform, or even Amplitude's buying behavior?

Spenser Skates

executive
#16

Absolutely, it's a question that comes up more often. In this era, customers want to be sure they're dealing with a company they can trust to be secure, one that doesn't leak their data and that operates using all the best practices. I think we're also seeing more third-party attacks, where someone goes through other libraries or other indirect routes to hack a company rather than going direct. So I think it's a wake-up call for all of us now that these models are — now that these models are sophisticated enough that a single individual can run a bunch of rogue forms of agents to attack a company. We've been extraordinarily lucky in that we've never been hit in a big way. And so I think what we're doing now is being very aggressive and proactive with our own stack, and then also communicating that to customers. That way they're much more confident that working with us — someone on the bleeding edge, not only on the bleeding edge of AI but also on the bleeding edge of AI security — that they can trust us and all of that; so — but for sure, it has come up as a question, and I think customers are looking to us for education on it.

Yitchuin Wong

analyst
#17

Understood. No, that's clear. It's certainly a significant worry. Perhaps just a single quick follow-up for Andrew. Regarding the Statsig acquisition, we know last quarter we discussed the ARR breakdown or even what impact it had. Could you update us on whether Statsig outperformed? And how should we think about that outperformance and what Statsig's impact was in the quarter?

Andrew Casey

executive
#18

Yes; so we referenced the $21 million in net new ARR that arrived, split between what you'd typically consider $15 million originating from Amplitude and $6 million originating from the Statsig business. I also noted that continuing to provide those breakouts will be very difficult for us, because we're already seeing new solutions built on our road map by the combined teams, and we're taking those products and beginning to create new products within the Statsig experimentation space. But we're extremely pleased with how strongly the Statsig business has integrated into Amplitude and how we've managed to serve clients. We've managed to renew contracts — and we keep seeing tremendous, tremendous interest from both new clients and existing clients in the Statsig business; so we're quite optimistic about the future. We're making substantial investments to keep driving growth and looking forward to the next phases.

John Streppa

executive
#19

Nice work on that sharp pickup once more. Billy Fitzsimmons of Piper Sandler will ask the next question, and after him comes Rob Oliver from R.W. Baird.

Billy Fitzsimmons

analyst
#20

There are plenty of thrilling developments in the business — Statsig, Wave, agent analytics — but what has somewhat caught us off guard across the last two quarters is simply core demand, core utilization, and improved monetization of the base. Where does that come from — better selling? Would you credit that strength to improved sales blocking and tackling that you implemented over roughly the last year, and now you're seeing the payoff on the other side? Or would you instead credit it to feature velocity and shipping more products, and being where customers want or need you to be?

Spenser Skates

executive
#21

Yes. I think during 2022 and 2023 we discussed fairly often the overselling that happened in the 2021 COVID era; so utilization — one of the figures you can observe is utilization rates, where I believe we were in the 70s across a good chunk of that period; and I have to give the go-to-market team an enormous amount of credit, by far. So Nate, our Chief Commercial Officer, Jenna leading post sales, plus many of the individuals and leaders across go-to-market. They have put in a tremendous amount of effort ensuring customers are enabled and onboarded and using everything they're entitled to, and we're clearing all the various blockers — whether it's, hey, we lack engineering resources, so we've gotten delayed, which is why we've instrumented, or hey, we began with a small use case but never graduated the big one. It was simply a lot of blocking and tackling across a number of years to lift those utilization rates. So I have to give them enormous credit. We've also — the engineering, product, and design team has likewise done considerable work making Amplitude easier to use and onboard with. So it's been a company focus for the last few years, so that we got beyond that overselling and reached a new regime where we ensure all of our customers, from our largest to our smallest, are fully using Amplitude.

Billy Fitzsimmons

analyst
#22

And then, Spenser, there's a bigger-picture question I'd like your take on. Even since the last quarter, so much has shifted around consumer agents, correct? Right now we've got Meta Muse, we've got Dots, and individuals are interacting with brands in different ways — and in many cases with many of your customers. Obviously, your product velocity has sped up. But can you just lay out for a lot of investors where Amplitude fits into that conversation and what the potential opportunity looks like for you there?

Spenser Skates

executive
#23

The big thing we're seeing is that a lot of our customers are actually launching their own first-party agents. Regardless of which sector you're in — telecom, financial services, or even e-commerce — you're going to want an agent on your website that assists your customers, whether that means supporting them, helping them locate the best deal, or searching for that one item that's hard to find. And none of those customers happen to be experts at building agents. That's why they turn to us, asking how to optimize this and get it working really well within the broader customer journey. So that's a pretty huge opportunity; and as for certain agents the labs have released, such as Muse, Grok Bot, or Dots, it's still quite early for them. Regarding adoption, my wife is actually an investor — a consumer investor. She follows this area very closely, and we always joke that most people don't want extra work on their plate; these tools are ways to get more work done efficiently. So it looks cool to tech folks like us, but for the ordinary person, adoption still has a long way to go. I think we have a chance to sell into a lot of those companies, and a few of them are already our customers. That's exciting for us as well. When you think about Amplitude in the context of the macro picture, don't fixate on any single form factor—whether that's a website, an app, VR, agents, or anything else. Those distinctions matter less. We cover Internet of Things, so we have all those various surface areas, with custom SDKs and many different ways of tracking. As a company, Amplitude represents a wager that interactions between people and technology will keep increasing over time. So, provided people keep devoting extra hours—whether through an agent, an app, voice interaction, or some new kind of device, and nobody can predict what things will resemble in 10 or 20 years—but provided that engagement keeps rising, the creators of those experiences will want to improve them. And nobody on earth does that better than we do.

John Streppa

executive
#24

Rob Oliver of R.W. Baird is our next questioner, and Koji Ikeda will follow him.

Rob Oliver

analyst
#25

So my first question is really about end markets, Spenser — I saw you at the DigitalOcean event. You folks are obviously performing extremely well with AI natives and with AI companies. Tying back to what Billy asked, you did mention that you guys had kind of dealt with some of that overselling back in the old days. So Andrew, I know part of your strategy has involved diversifying and really getting better at moving into other verticals; so I'd love an update and some color on non-tech verticals, what traction looks like there, and how the pipeline shapes up into '27? And then I have a quick follow-up too.

Andrew Casey

executive
#26

Sure. What I'd say is that we continue to see a pretty wide range of customers stepping forward from all sorts of industries that are deploying Amplitude; and Center did a really nice job covering some of the customers who are early in their journey toward being AI native, and they might simply be swapping out some legacy technology. but they still require our help and they still turn to us for guidance. And so you see that reflected among our financial services clients, our telecommunications clients and our retail clients; and so every one of those verticals is quite large. I believe we've mentioned before that software and technology makes up roughly 30% of our total, but each of those others is double digit, and the financial services one is expanding quite rapidly. So you see a broad array of customers stepping forward; and it isn't by accident that we show those different types of clients on our websites and in our scripts, because we want people to understand that Amplitude isn't just for technology providers; it's really for any company that's trying to digitally engage with their client in an effective way and improve those experiences.

Rob Oliver

analyst
#27

Great. Okay. That's really useful. And then, Andrew, one more follow-up question for you. Specifically about gross margins — I mean, you've all been very explicit regarding a few of the stat effects and the inference costs that come ahead of revenue, and things of that nature. So when we consider how gross margins will trend, do you feel good about the current level? Perhaps you could speak a bit about what the activity levels around data ingestion are telling you regarding where that might land, and it would be helpful to know how much Statsig contributed to gross margin this quarter. Sure. So to begin, I would say that when we acquired the Statsig business, we inherited a hosting environment that wasn't optimized. We were moving quickly to make certain the main priority was guaranteeing there'd be no interruption in service and that clients felt Amplitude was definitely the right home for them to keep getting their experimentation services from. That mattered most; and what we didn't do at that time was take a long-term look at whether we had the optimal provider partner for those hosting services. We lacked a long-term contract in nature. If you look now in our Q, we actually disclosed that we did extend a long-term contract with Google to supply the support services — the hosting services for the Statsig environment. And so we're working toward better gross margins in areas where we know historically we're able to make improvements; we're also working on engineering optimizations, both in our AWS and our Google environments. So all of those things — you can count on us doing our best to try and make optimizations without sacrificing any level of service quality. Now the other side of this is you're pointing to the data ingestion rates and AI utilization, and those are strong predictors for us regarding future revenue. And I'm very pleased we're at this point now where utilization versus entitlement is so positive. And because it starts giving you an early signal, we can have the right kinds of discussions with our enterprise clients about broader expansions, whether that's an upsell or it gives the sales rep an opportunity to have a conversation about other products that they can get value from as they're adding more surfaces to that data ingestion. So for us, it's a great leading indicator, and we're quite pleased with the progress.

John Streppa

executive
#28

Our next question comes from Koji Ikeda from Bank of America, followed by Jackson Ader.

Koji Ikeda

analyst
#29

So, out of that $21 million sequential ARR increase, $6 million was from Statsig. That leaves $15 million — could you give some color on the different components, things like logos, usage expansion, cross-sell, pricing and contraction — which of those improved the most compared with the second quarter?

Andrew Casey

executive
#30

Koji, I'd say you're still seeing a really strong cross-sell motion in our business. I mean, the responses we're getting are great from — believe it or not — our traditional experimentation and our analytics product; we mentioned a really large telecommunication provider we closed during the quarter. That one was a classic analytics sale; we're also seeing contributions from upsell, which is great. In the past that contribution was negative to flat, but now upsells are happening too, with the utilization rate climbing higher and higher, and in some cases customers going beyond their entitlement. Still, I'd tell you cross-sell is the biggest contributor. Demand remains great for our session replay and our guides and surveys products, and the overall platform sale is going very strong.

Koji Ikeda

analyst
#31

Perhaps my follow-up is about NRR, and more precisely that pro forma NRR figure of 105%. It was 105% in this quarter and 105% in the prior one. So what is keeping it from going higher? Cross-sell sounds healthy, but what other factors should we be considering when it comes to NRR?

Andrew Casey

executive
#32

The first thing I'd point to is that we brought on a customer base from the Statsig business, and we didn't concentrate much on pushing upsell or cross-sell. Our emphasis was more on getting renewals done. What you're seeing is that a base of ARR now sits in there which, if you will, blunts the other expansions from an impact standpoint on net dollar retention. Another point I mentioned: there were quite a few solid expansions this quarter involving customers who had signed up as new customers less than a year ago. That cohort therefore hasn't caught up enough to really capture the positives tied to the expansion; in fact, one customer signed up close to 6 months ago, and it was only a few hundred thousand or perhaps just under $200,000. Today it's well above $1 million in ARR; that expansion did occur, and it won't affect net dollar retention at all until we pass the quarter in which it became a new customer; so those are some of the elements having an impact.

Koji Ikeda

analyst
#33

That seems stronger than what was reported.

John Streppa

executive
#34

Next up, we have a question from Jackson Ader at KeyBanc, and after that, Clark Wright with D.A. Davidson. Please proceed, Nate.

Analyst

analyst
#35

This is Nate speaking for Jackson, so perhaps this question is meant for Andrew. It's been a while now since the pricing model change, so I'm just curious about what usually happens to customer spend once they move over to the new pricing model, and presumably there's less friction. And are you seeing an inflection in spend, whether that's higher volume or cross-sell opportunities? Any color on that would be great.

Andrew Casey

executive
#36

Yes. Thanks for the question. I'm really pleased with how effectively we've been executing on the rollout of our new pricing and packaging. More than 90% of transactions completed during the quarter, both new business and renewals, were closed on the new pricing and packaging. That has now brought us to 38% of our total ARR being on the new pricing and packaging. And by year-end, we expect it will exceed 50% — when we implemented this, we intended to accomplish a few things. First, we wanted to ensure it was simplified for our sellers to quote, that it was transparent for our customers to understand the cost as they expanded further into multiproduct, and that we had a path through which to raise overall ASP and contract duration for clients. I'd say all of those things are positives. We're seeing higher ASPs among customers who are renewing or starting on new pricing and packaging. We're seeing more customers become multiproduct customers as a result of pricing and packaging, and we're seeing contract duration lengthen; so all of those are outcomes as we move further and further along in the rollout. We've done it very methodically, and in how we're approaching customers, not forcing them to migrate, but incentivizing them to migrate as much as possible; and much of that comes from gaining access to our AI products if they move to the new pricing and packaging. So personally, I'm very, very pleased with how well the teams have executed on that. And there's still some road ahead. Obviously, we're at 38%. We're not at 100%. Reaching that level will take some time, but we are making great progress.

Analyst

analyst
#37

That’s useful context. Then I suppose one more thing for Spenser. I understand you mentioned the initial focus for Wave is on landing bigger customers. But which customer profile do you generally expect will take up Wave? And after that, what has to occur before it turns into a bigger commercial opportunity?

Spenser Skates

executive
#38

Well, I believe this is going to be an enormous commercial opportunity. I wish I could pull out a pile of projections and spreadsheets and say, oh yes, it will be a 33.333 repeating, naturally, percentage increase or something like that. But we don’t — it’s early. My view is that the opportunity will be enormous because the uplift may be 50% to 100% of the existing contract with the rest of the platform included, since we’re moving beyond merely running a lot of data pipelines and analytics and session replay and tools to running the entire — we’re genuinely making someone’s business better. And so a company will pay 10x or 100x for that impact compared with what they pay purely for the data tooling that sits upstream of it. So in any case, I think it’s an enormous opportunity. I think the right approach is to go deep with a few customers first. I think if it takes time, we’ll see plenty of acceleration in the business.

John Streppa

executive
#39

Our next question comes from Clark Wright from D.A. Davidson, followed by Arjun Bhatia.

Clark Wright

analyst
#40

During the prior quarter, you mentioned the 40 AI native customers that are currently working with Amplitude, and I felt that was a fantastic stat to call out. How — could you perhaps discuss that cohort and how it behaves compared with traditional software customers when it comes to data consumption, product adoption and expansion? And then if you can, could you also address whether these customers are turning into larger accounts materially faster than what you've observed before?

Spenser Skates

executive
#41

That's 40 accounts. It isn't a massive sample set. It's clearly wonderful to be working with leading-edge companies. Their preference, I'd say, is closer to: keep the tooling solid and don't bother us — not the hands-on assistance we extend to nearly all our other customers. Feedback from them reaches us all the time and is genuinely valuable. Statsig, I think, has served as a strong initial entry point with them as well. Statsig is ours — clearly the very tooling OpenAI operates with in-house. Beyond that, additional labs and AI-native companies also use Statsig — their mindset is simply, what’s the strongest product, or the most advanced software, for my particular use case, and they’ll work out how to fit it together with the other building blocks they already have. By contrast, I’d say many bigger enterprises — and most of the remaining customer base — come to us wanting considerably more hands-on assistance, and that’s exactly where we see a very large opportunity.

Clark Wright

analyst
#42

Understood. And then, Andrew, regarding the data upselling you described — could you explain how that fits into the growth formula ahead, and where you expect event growth to trend relative to the revenue growth line going forward?

Andrew Casey

executive
#43

I think event volume growth will always outpace the revenue growth profile. That's because we want to keep incenting customers to adopt more. We want them adding more services to the platform base. And we should be passing on the marginal incremental cost reductions that we're seeing to our customers. In fact, that shows up in our pricing strategy: as customers adopt more volume, they get greater and greater discounts off our list prices. So I think some of what we've been doing recently around expanding the services, adding more capabilities into our products where customers are blending them with their own data warehouse capabilities, is leading to greater and greater outcomes, and that just incents more adoption. So for us, it's — as I think Spenser mentioned earlier, it's really about thinking of Amplitude as the application that provides the best outcomes given all the different surface areas and data that you're — our businesses and their customers are interacting with. And so we want that to be increasing. We want that to show greater and greater, and we want that to be something our customers look at and feel they're paying a fair value for the value they're getting from the Amplitude investment.

John Streppa

executive
#44

The next questioner is Arjun Bhatia from William Blair, and after that we'll go to Elizabeth Porter. It looks like Willow Miller is joining us in place of Arjun.

Willow Miller

analyst
#45

This is Willow, sitting in for Arjun. It was good to hear the latest on Statsig in the prepared remarks. Still, I'd like to know how aware your customer base is of Statsig. Do you need to invest additional time in educating customers and promoting Statsig in order to support cross-sell activity?

Spenser Skates

executive
#46

For sure. Statsig represents an enormous opportunity. When we inform customers that they have it, I believe their reaction is, "Oh, I didn't even realize you offered experimentation even prior to Statsig; now it's like, oh, well, you have the bleeding edge one; that business has — that product has an incredible reputation and deservedly so. So yes, I think there's a lot of opportunity to cross-sell; we've already done that here and there over the last 6 months, and we're going to be doing a lot more of that in Q4 and 2027.

John Streppa

executive
#47

Elizabeth Porter from Morgan Stanley is going to be our next question, coming from her line, and after that we'll hear from Nick Altmann at BTIG. Lucas will be speaking in Elizabeth's place.

Lucas Cerisola

analyst
#48

There are plenty of new products, and demand is still very strong. I'd be interested to hear which specific products you think will drive the most paid expansion from here, particularly as customer pain points shift over the next couple of months?

Andrew Casey

executive
#49

What I'd say is that the ones we're focused on for the biggest expansions, in percentage terms, would be agent analytics and Wave. Even though it's very early with customers, I'd say those two are going to show strong rates of improvement; we're already seeing, as we build and demo for clients, tremendous interest in learning more; and honestly, that also opens up many more parts of the platform for us. But among our existing products, I'd say core analytics, experimentation, the Statsig business — those are showing great growth as we keep driving consolidation in the market and replacing legacy players.

John Streppa

executive
#50

And our last question will come from Nick Dannewitz on for Nick Altmann.

Nicholas Dannewitz

analyst
#51

Great. We're picking up from other businesses we track that there's some softness further down market — things like cost optimization and customers being more price sensitive — while higher up, we're noticing things such as sales cycles stretching out and additional decision-makers getting involved. So with all that in mind, what are you observing on your end when it comes to demand? And I think you mentioned in the press release that you have more than 5,100 customers; and looking at the prior quarter, that figure was around 52. So any insight there would be useful too.

Andrew Casey

executive
#52

What I'd say is, as for me hearing from other peers in the market — I'm not certain who precisely you mean — but hearing them describe weakness in SMB or mid-market doesn't surprise me, since we're moving into a stretch of rising interest rates. That's, I think, what you see in those core markets. Honestly, it's one of the reasons that 2.5 years ago, we began concentrating increasingly on drawing enterprises in and truly pushing consolidation within the market. With enterprise clients, everything improves — how inclined they are to purchase more, their capacity to take on multiple products, their gross retention, their net dollar retention; every one of those measures is stronger in the enterprise. So it doesn't surprise me. And so what I'd tell you is that what we've examined quite carefully is how to return the best information to our investors, grounded in where we're focused. That's why we've shifted further toward concentrating on customers above $100,000, and we concentrate less and less on smaller plus customers and the packages aimed at start-ups, including within our customer counts, because those really aren't all that indicative of what's driving the actual ARR growth; and that's where you see us shifting across the last 2.5 years — simply an ever-greater enterprise emphasis.

John Streppa

executive
#53

Thanks, Nick. That brings our third quarter earnings call to a close. We appreciate your time and interest, and we hope to see you on the road during this quarter at conferences hosted by Needham, UBS and Raymond James. Thank you, all.

Spenser Skates

executive
#54

Thank you, guys.

Andrew Casey

executive
#55

Thank you.

Read the full transcript via the API

You're viewing the first half of this call. Get the complete Amplitude, Inc. transcript — plus 132,000+ transcripts from 7,000+ companies, speaker segments, AI summaries and full-text search — through the EarningsCalls.dev API.

Get the API View API docs →

This call discussed

For developers and AI pipelines

Programmatic access to Amplitude, Inc. earnings transcripts and 132,000+ others is available through the EarningsCalls.dev REST API. Plans from $39.99/month — full transcripts, speaker segments, full-text search, and the recently-added /api/v1/transcripts/recent polling endpoint for ETL pipelines.

More tools for stock research

OptionsBell: an unusual options activity scanner that tracks options flow across all US stocks and emails you when large trades appear in Amplitude, Inc. or any ticker on your watchlist.

StockMarketScan: a stock screener with alerts: 24 pre-built screens refreshed every 5 minutes, watchlists and email notifications. Both offer API and MCP access and fair monthly pricing.