Products
Solutions
Resources
Statsig and Amplitude’s announcement in May was a record-scratch moment for experimentation. The next track starts today.
Statsig joined the Amplitude family on May 5. Since that announcement, we haven’t said much more publicly about what that means.
That’s because we had work to do before we started making claims. We’ve been learning. We’ve been maintaining. We’ve been getting into gear to match the build velocity and customer support that Statsig is known for.
(And we have said a lot privately; our team has been in customer conversations every day since the announcement. It’s been inspiring getting to know you all, and if we haven’t met yet, I look forward to talking soon.)
The past six weeks have been Phase 0. Here’s what we did, and where we’re headed next.
Phase 0: What’s been happening since May
First off, I feel like I should introduce myself: Hi! I’m Chris Yu, VP of product management at Amplitude. And now Statsig.
Since Statsig joined the Amplitude family in May, I’ve been leading a crack team of about 50 Ampliteers maintaining the Statsig platform, supporting its customers, and planning its future. Along with me, other key leaders of the Statsig + Amplitude team are:
Curtis Liu, co-founder and CTO of Amplitude
Larry Xu, senior director of engineering at Amplitude who has led the engineering teams on Amplitude Experimentation, Session Replay, SDK, and more
Eric Metelka, director of experimentation at Amplitude and former head of product at Eppo
Vivian Magida, global solutions architect at Amplitude and former director of product experimentation at Peacock
Our top priority in Phase 0 was making the transition as smooth as possible. That meant two things: learn fast and keep shipping.
On the learning front, the original Statsig team, now at OpenAI, has been an incredible partner. They walked us through the platform, codebase, operations, and roadmap. They answered everything, even when it meant digging up years-old documentation. It’s clear how much they care about Statsig and its customers, and we intend to live up to the standard they set.
We also learned from talking with you. The highlights: you love what you’ve had with Statsig, and you don’t want to see that go away. (Some of you used more explicit language to describe this.) We intend to meet and exceed your expectations.
Learning was a parallel effort to actually doing the work. In Phase 0, we addressed customer needs such as MCP version history tools, fixed bugs as they arose, and identified quick wins to improve the service. Check out the Product Updates page for the full list.
We spent Phase 0 getting up to speed. Now we’re ready to show what we can do.
Phase 1: Expanding the platform
Statsig’s platform is impressive. The warehouse-native tech, the dynamic config tools that let customers change their product without a deploy, the stats engine … we get it now. We understand why some of you chose it over Amplitude.
But as Statsig’s original team would be the first to admit, there’s no such thing as a perfect product. Perfection is a limit you approach through learning, experimenting, and working your ass off. That’s what we intend to do.
We’re going to keep investing in and expanding Statsig, shipping at the same speed Statsig is known for. In Phase 1, our top priorities include:
Control and governance. We’ve already released updates for MCP governance, and we’re working on more granular permissioning. We’re enhancing API controls. We’ll be launching release pipelines for experiments in the next month and a half.
AI-native workflows. Evals and LLM traces will be hitting GA in the next couple of months. And, AI experiment reports will be next. This is an area where we’ve been working closely with Amplitude’s AI teams, like the ones building Agent Analytics. There’s a lot we’re excited to bring over and connect.
Platform scale and performance. Amplitude has spent over a decade building one of the fastest, most performant data pipelines on the market. We’re bringing that performance into Statsig.
As we expand, we’ll continue to strengthen and innovate on Statsig’s world-class stats engine and warehouse native capabilities. They’re two pillars that make Statsig truly unique, and they’re essential to the continuous learning loop we’re building with Statsig and Amplitude together. They’re only going to get better.
Statsig is also unique in its commitment to building what customers ask for. We’re keeping that commitment strong, and we’ll keep fielding your requests as quickly as we can. Things like the CMAB beta are moving forward. We’re hosting events to connect (keep an eye on our socials and your inbox). Your energy and interest are amazing; keep the feedback coming.
And to support all that, we’re hiring.
And … what about integration? Is that like Phase 2?
Great question!
We are actively building the integrated Statsig and Amplitude experience. Over the next four months, we'll connect the two products at the data layer: your Amplitude events and cohorts will work inside Statsig, and Statsig experiment outcomes will show up inside Amplitude. By the end of Q3, customers of both products will be able to use them together without having to choose.
After that, we go deeper with the goal of full interoperability. Our vision is a system where product engineers ship safely, test new solutions, and prove their fixes work, all from a single workflow in a continuous learning loop. That’s what uniting Statsig’s best-in-class validation layer and Amplitude's best-in-class behavioral layer will enable. That’s Phase 2.
As we integrate the platforms, we’ll be working closely with customers to ensure you have the best possible experience, whichever platform you currently call home. I’ll share more as we build.
All that really matters
Today, when you ship a gated feature, how do you know if it’s working or not? You find out from a customer complaint, three tools open, 3.5 hours later. That doesn’t scale with the speed you’re now able to build at.
We're building for a world where you can move faster because you’re shipping safer. Gates ship with guardrails attached. Rollback is automatic if checkout drops 4%. “3,200 users affected” instead of “fired 10,000 times.” Fix it, ship the fix, see the issue go to zero. Then get notified about what to test next. That’s what it looks like to build at the speed of AI.
We believe that’s the future of experimentation. We believe that the future isn’t very far off. And we believe we’re the right ones to build it. Amplitude has the curiosity, drive, and technical chops to carry on the original Statsig team’s legacy, as we’ve demonstrated for over a decade building analytics and experimentation tools. (We also love dogs. The dogs aren’t going anywhere.)
I can’t wait to show you what we’re building. Let’s go.