Who’s Buying Who: The Acquisitions, Mergers, and Shutdowns Redefining the Experimentation Space in 2026

The A/B testing and experimentation space looks very different today than it did 5–10 years ago.

Private equity firms are consolidating independent tools, tech giants are acquiring experimentation infrastructure outright, and AI is transforming what it means to “run an experiment.”

This article explores the major acquisitions, mergers, and sunsets in the experimentation space, and what each one signals about where experimentation is headed.

Overview of A/B testing tool acquisitions, mergers, and sunsets

1. Optimizely

What the official PR says:

Episerver positioned the deal as the creation of “the industry’s most advanced digital experience platform,” which is able to optimize every customer touchpoint across the entire user journey.

The pitch was simple: combine Episerver’s content and commerce muscle with Optimizely’s experimentation engine, and you get a platform where digital teams can build and test everything in one place.

Post-acquisition, Episerver said it was serving approximately 9,000 brands and over 900 partners globally, spanning retail, financial services, healthcare, and media, among others. Optimizely would continue as a standalone solution but would also be bundled into Episerver’s broader offering immediately.

By January 2021, Episerver had rebranded the entire combined company as “Optimizely”—a telling signal of which brand had more market equity.

What we think is happening behind the scenes:

Episerver’s acquisition of Optimizely set an early template for what was coming. Experimentation had started its journey from a standalone discipline to a bundled platform feature. The question for the market became: do you buy a best-of-breed testing tool, or do you consolidate into a suite?

Here’s our take: Optimizely is the best-case scenario of this model—a bundled suite that didn’t sacrifice experimentation depth. Episerver built a behemoth using one of the first platforms to offer deep, full-stack experimentation to test omni-channel journeys.

That said, not every bundled suite will have the experimentation power of Optimizely, especially as bundled tools often prioritize ease of adoption over depth.

2. Dynamic Yield

What the official PR says:

Worth noting upfront: this was actually the second acquisition for Dynamic Yield. McDonald’s had originally purchased Dynamic Yield in 2019, integrating its decision technology into drive-thrus and ordering kiosks across several markets. Mastercard then bought it from McDonald’s.

The official framing was all about personalization at scale. Mastercard said Dynamic Yield’s technology would slot into its existing Data & Services organization (a team of over 2,000 data scientists and consultants serving customers in 70+ countries) to help brands deliver more relevant, right-place, right-time consumer experiences.

McDonald’s, for its part, wasn’t walking away. It planned to continue scaling Dynamic Yield’s capabilities globally and across its ordering channels.

What we think is happening behind the scenes:

Rather than building a personalization tool from scratch, acquiring Dynamic Yield gave Mastercard immediate access to a proven platform (a new revenue stream) and the behavioral data its users were already generating. That data now informs how Mastercard shapes and sells its own engagement services to the merchants and banks in its network.

On the flip side, Dynamic Yield’s personalization and testing capabilities are now being applied at a large scale across eCommerce, media, restaurants, and financial platforms running billions of transactions.

Experimentation at that level looks very different from what most CRO teams run day-to-day, and this acquisition widened that gap.

3. Google Optimize (sunset)

What the official announcement says:

Google kept it short and clinical. The official line was that Optimize “did not have many of the features and services that customers request and need for experimentation testing,” and that Google would instead invest in third-party A/B testing integrations for Google Analytics 4.

No replacement product was launched. No migration path was offered. Google instead pointed users toward third-party tools (specifically AB Tasty, Optimizely, and VWO) and made its APIs publicly available so those tools could plug into GA4.

What we think is happening behind the scenes:

Sunsetting Optimize without a like-for-like replacement was Google nudging the market toward full-stack, server-side experimentation. It’s arguably the loudest industry signal yet that full-stack experiments are the future, and that businesses need to rethink how measurement strategy is built from the ground up.

4. Intellimize

What the official PR says:

Webflow framed this acquisition as a major step toward becoming the world’s first “Website Experience Platform” (WXP), a unified platform combining visual site building, CMS, hosting, personalization, and optimization.

The pitch was that building a great website and optimizing it for conversions had always been two separate, painful workflows. This acquisition was meant to collapse that gap.

Intellimize’s co-founder Guy Yalif joined Webflow as Head of Personalization, with the majority of Intellimize’s ~50-person team making the move, too. The deal was reportedly in the eight-figure range—so not a blockbuster exit, but a meaningful one for an eight-year-old startup that had raised just over $50 million.

What we think is happening behind the scenes:

Webflow’s move confirmed that the no-code website building space and the experimentation space are on a collision course. If you can build and optimize in the same visual canvas, without having to touch code, the barrier to running experiments drops dramatically.

Webflow had over 200,000 customers at the time of the acquisition, and most of them were marketers and designers who’d never run a structured experiment. Putting personalization and A/B testing directly in their workflow significantly expands who gets to participate in experimentation.

The caveat, however, is that more people running tests doesn’t automatically mean better experimentation. Without proper research, a trustworthy stats engine, and a strong understanding of statistical significance, programs may generate false positives that don’t move the needle.

5. Wingify (VWO)

What the official PR says:

This one’s a bit different from the others in that it’s a private equity buyout, not a strategic acquisition. Everstone Capital acquired a majority stake in Wingify for around $200 million, with the stated goal of accelerating VWO’s product innovation and international growth.

VWO’s founder, Paras Chopra, built Wingify from a two-person startup in 2010 to a profitable business with over $50 million in ARR, entirely bootstrapped.

He stepped back from the CEO role but stayed on as a shareholder and board member, with co-founder Sparsh Gupta taking the reins and leading the company into its next phase alongside the rest of the leadership team, all of whom retained a meaningful equity stake.

What we think is happening behind the scenes:

A PE firm buying a bootstrapped, profitable A/B testing company is a straightforward growth-and-exit play. Everstone gets a cash-generative business with strong margins, and VWO gets capital to push harder on enterprise sales in the US and Europe.

VWO built its reputation as an independent, founder-led tool, fairly priced and not tied to any platform agenda. PE ownership changes that. The pressure shifts toward growth targets and margin optimization, which typically means pricing moves upmarket.

If you chose VWO specifically for its affordable pricing, keep an eye on the pricing as the company enters this new phase.

6. SiteSpect

What the official PR says:

Monetate (a portfolio company of Centre Lane Partners, a private equity firm) positioned the deal as a way to deliver unified client-side and server-side experimentation with full personalization capabilities in a single enterprise solution—something it claims no other platform currently offers.

The technical fit here is pretty clear. Monetate brought AI-driven, real-time personalization, while SiteSpect brought its patented zero-flicker testing technology and a strong foothold in regulated industries, like healthcare and financial services (which Monetate previously couldn’t fully serve).

What we think is happening behind the scenes:

SiteSpect had carved out a specific niche: server-side testing for large enterprises with strict security and compliance requirements (like HIPAA and PCI). That’s a hard niche to build, so Monetate acquired that capability rather than building it from scratch.

Regulated industries have historically had to choose between robust experimentation and compliance. So, a platform that handles both removes a blocker for teams that have been running limited testing programs precisely because their compliance requirements ruled out most tools.

That said, there are signals in the market that Monetate may be deprecating SiteSpect. So if you’re a current SiteSpect customer, keep a close eye on how the product roadmap evolves.

7. OfferFit

What the official PR says:

Braze acquired OfferFit for $325 million, positioning it as a move to deepen its agentic AI capabilities in customer engagement. The two companies had already been working together as technology partners, so the integration wasn’t starting from zero.

OfferFit’s core technology is a reinforcement learning engine. Rather than running traditional A/B tests, it uses AI agents to autonomously make decisions about what to send each customer, learning and optimizing continuously without manual test setup.

Braze’s goal is to fold this into its broader platform alongside Project Catalyst, its native AI agent, to deliver 1:1 personalization across the full customer journey.

What we think is happening behind the scenes:

OfferFit wasn’t just an experimentation tool. It was built specifically to replace A/B testing with autonomous AI decisioning, and Braze paid $325 million for that thesis.

The largest customer engagement platforms today are actively betting that AI-driven optimization will handle decisions that marketers currently make through structured experiments.

That doesn’t mean A/B testing will disappear, but the traditional ‘set up a test, wait for results, pick a winner’ framework is no longer enough on its own. Experimenters now need to figure out where human-designed hypothesis testing still wins over autonomous AI decisioning, and how to build a program that balances both.

8. Eppo

What the official PR says:

Datadog framed the acquisition as the missing piece in its product analytics story, with Eppo’s feature flagging and experimentation capabilities plugging directly into its existing Product Analytics suite to create a full end-to-end solution on one platform.

The AI angle was front and center. Datadog’s VP of Product noted that as teams deploy multiple AI models, experimentation becomes essential for quantifying the business impact of different models, agent behaviors, prompts, and UI changes side-by-side.

What we think is happening behind the scenes:

Eppo was purpose-built for the modern data stack, with warehouse-native metrics, rigorous statistics, and a workflow that brought engineers, product managers, and data scientists into the same experimentation loop. That’s a very different DNA from traditional A/B testing tools, and it’s exactly what Datadog needed.

Datadog’s acquisition moves experimentation closer to engineering infrastructure—think feature flags, large data volumes, deployment pipelines, and canary releases—rather than marketing use cases.

9. Statsig

What the official PR says:

OpenAI acquired Statsig for $1.1 billion in an all-stock deal (one of its largest acquisitions to date), bringing Statsig’s founder and CEO Vijaye Raji in as CTO of Applications, where he’ll head product engineering for ChatGPT and Codex.

OpenAI described Statsig as one of the most trusted experimentation platforms in the industry, already powering A/B testing, feature flagging, and real-time decisioning for OpenAI itself. Statsig will continue operating independently from its Seattle office and serving its existing customer base.

What we think is happening behind the scenes:

The relationship between OpenAI and Statsig wasn’t new. OpenAI was already a customer, and we believe Statsig was providing some infrastructure to support OpenAI’s own testing cadences and experiments.

At ChatGPT’s scale—around 2.5 billion prompts a day—that represents a massive experimentation workload. Bringing that infrastructure in-house through a $1.1 billion acquisition likely reduces the long-term cost of running experiments through a third-party platform.

More importantly, it signals how central experimentation has become to AI product development. The deeper question for the experimentation space is whether other AI companies will follow the same path, treating experimentation infrastructure less like SaaS and more like core product infrastructure.

10. VWO + AB Tasty merger

What the official PR says:

VWO and AB Tasty announced a definitive agreement to merge, creating a combined platform with over $100 million in ARR, more than 4,000 enterprise customers globally, and roughly 90% of revenue from the US and Europe.

Everstone Capital, which already held a majority stake in VWO, is injecting additional capital into the deal and will remain the largest institutional shareholder in the merged entity. Sparsh Gupta of VWO will serve as CEO of the combined company, with co-founders from both sides taking senior roles across product, revenue, and strategy.

What we think is happening behind the scenes:

Less than a year after acquiring VWO, Everstone used it as a platform to absorb a direct competitor.

VWO’s CEO Gupta acknowledged in a recent interview that while the company has grown at a CAGR of 25-30%, the cost of that growth is climbing. Also, the total addressable market (TAM) for web experimentation is largely tapped, making new customer acquisition increasingly expensive.

Merging with AB Tasty sidesteps that problem by buying an existing customer base rather than fighting for one.

For experimenters, this consolidation raises questions about how customers will be treated going forward. AB Tasty is known for great customer experience, while VWO has had a weaker reputation in that area. Teams, especially those loyal to AB Tasty, will be watching closely to see which approach carries through in the merged entity.

What awaits experimentation in 2027?

We gathered qualitative signals from 11 experimentation practitioners—spanning in-house teams, agencies, and independent consultants—on what they want from their A/B testing tools and where they see experimentation heading.

What experimenters want from their tools

• Prompt-based workflows. Testers want to be able to describe what they want to test in plain language and have the platform translate that into a structured experiment. This means less setup friction and a faster time to hypothesis.
• Automation for repetitive tasks. Things like tagging tests, designing variants, writing variation copy, generating reports, and documenting results manually are eating into time that should be spent on analysis and strategy. Practitioners want platforms to handle more of this automatically.
• Better reporting and documentation. Test results shouldn’t live in spreadsheets or scattered Slack threads. Testers want built-in learning repositories—a proper knowledge base that clearly captures what was tested, what was found, and who owns what.
• Access to raw data and better integrations. Experimenters want to be able to inspect their data directly, rather than simply trusting a dashboard summary. And they want their testing tool to integrate seamlessly with their data warehouse, analytics stack, and CDP without custom engineering work.
  
  Warehouse-native experimentation solves this by pulling metrics directly from the source of truth rather than a separate, siloed dataset.
  
• Transparency and a trustworthy stats engine. Beyond features, experimenters want to be able to understand how their testing platform works, explain its outputs to stakeholders, and defend the results. Clear documentation, a trustworthy stats engine, and honest pricing were all cited as trust signals.

Trends from the acquisitions: What’s coming for personalization and experimentation

The acquisitions we covered above point to a few shifts worth watching.

• AI agents are autonomous systems that can make decisions, take actions, and optimize outcomes without a human triggering each step.
  
  For example, Braze (which acquired OfferFit) uses AI agents to autonomously decide what to send each customer, across which channel, and at what time, replacing the manual A/B test workflow for lifecycle marketing.
  
  As more platforms move in this direction, experimenters will need to decide which decisions they’re comfortable delegating to an agent and which ones still need a human in the loop.
  
  Also, AI agents are increasingly browsing the web as users, researching, comparing, and making decisions on behalf of humans. So, websites have to figure out what converts an AI agent (as it may be different from what converts a human). Does this mean websites will become well-parsed markdown files? Probably not.
  
• With 1:1 personalization, instead of showing the same experience to broad audience segments, every visitor gets a version of the page tailored to their individual behavior, context, and history in real time.
  
  AI makes this personalization feasible at scale in a way that rule-based personalization never could. The question now is whether 1:1 personalization can be done with both relevance and respect for user privacy.

The bigger picture from the survey

The clearest takeaway from the survey wasn’t about any specific feature. Instead, it was that, while experimenters use AI for reporting, synthesizing insights, generating test ideas, and coding support, they don’t see it as a replacement for experimentation itself.

What practitioners want is simple: AI that reduces manual effort, flags anomalies, and speeds up analysis, while they remain responsible for deciding winners, interpreting results, and making the final call.

As AI overviews shrink top-of-funnel traffic (with CTR drops of up to 58%), experimentation programs need to be more focused, higher-impact, and strategically chosen to generate meaningful learning.