Equals One (d.b.a. Mida.so)

AI-powered A/B testing and behaviour analytics — lightweight, fast, and built for the post-Google Optimize era

Google shut down Google Optimize in September 2023, leaving millions of websites without the free A/B testing tool they had built their conversion workflows around. The alternatives were either expensive enterprise platforms built for teams that don't need simplicity, or simplified tools that added so much script weight they hurt the very page performance metrics their users were trying to improve. Mida.so was already positioned to fill exactly that gap — and it was built by the same team that had spent the pandemic years building Howuku, a full-stack CRO platform that taught them everything about what the market actually needed.

The CRO stack problem: useful tools that don't talk to each other

Conversion rate optimisation — the practice of understanding how users behave on a website and then systematically improving those experiences to increase the percentage who complete a desired action — has a fragmentation problem. The mature tooling in the category is distributed across multiple vendors: Google Analytics for traffic, Hotjar for heatmaps and session recordings, a separate A/B testing platform, another tool for on-site feedback, and potentially more. Each produces useful data in isolation. The challenge is that insights from a heatmap, a session recording, and an A/B test result require a human to hold all three mental models simultaneously and synthesise them into a coherent hypothesis about what to change and why. Most digital teams lack the time and the methodology to do that synthesis well, which means significant quantities of CRO data are collected but never actioned.

This was the original Howuku thesis: aggregate those individual data streams into a single platform, under a single tracking code, so that the picture of user behaviour is unified rather than fragmented. The product grew to serve over 3,000 organisations globally — a meaningful validation that the consolidation hypothesis was correct and that digital marketers would adopt a well-designed all-in-one alternative to the patchwork stack.

The pivot to Mida: sharper focus, bigger market window

Google's decision to sunset Google Optimize in 2023 created a specific, time-sensitive market opportunity. Optimize had been the default free A/B testing tool for a significant proportion of the global digital marketing community. Its shutdown didn't eliminate the need for A/B testing — it orphaned millions of users who needed to find an alternative quickly, without necessarily wanting to pay for an enterprise platform they would never fully use.

The team's response was Mida.so: a standalone A/B testing platform that targets this exact customer — the marketer or CRO specialist who wants to run experiments effectively without the overhead of an Optimizely or VWO, and without the performance tax that most A/B testing scripts impose. Mida's core technical differentiator is script size: at 8kb, the Mida script is approximately 10 times smaller than most competing platforms. The practical consequence is that Mida loads in roughly 20 milliseconds and has a negligible effect on page speed and Core Web Vitals scores — which matters enormously in a category where every competitor's tool creates a measurable page loading delay that can itself depress the conversion rates the tool is supposed to improve.

The AI layer adds a capability that the established players have been slower to deliver: MidaGX generates A/B test variations automatically from a natural language description of the change, or even a drawn annotation on the page, and produces a fully configured experiment ready to run — eliminating the hypothesis-to-test workflow overhead that often prevents teams from running as many experiments as they should.

Why "lightweight" is the actual competitive moat

It's worth dwelling on the script size claim because it's more strategically significant than it might initially appear. A/B testing tools work by injecting a script into the page that intercepts the user before the page finishes loading, serves them either the control or variant, and then records their behaviour. The heavier that script, the longer the delay before the user sees anything — a phenomenon called "flicker" that experienced CRO teams know compromises test validity by introducing friction that isn't present in the production experience.

Enterprise A/B testing platforms have historically accepted this trade-off in exchange for more sophisticated features. Mida rejects it: if the tool itself is degrading the experience you're trying to optimise, the data it produces is systematically biased. The 8kb script and 20ms load time are not just a performance feature — they're an argument about measurement validity that resonates strongly with the technically sophisticated CRO practitioners who are Mida's primary target audience.

This is also a durable technical differentiator rather than a temporary one. The enterprise platforms are constrained by their existing architectures — adding features that customers have already paid for and depend on — in a way that makes substantial script size reduction difficult without breaking existing deployments. A platform built lightweight-first can stay lightweight; a heavyweight platform can't easily become lightweight without significant re-engineering.

The founder: Donald Ng and the product-led approach to global growth

Donald Ng built Howuku during the pandemic with a clear product-led growth philosophy — acquire users through the quality of the product rather than through sales investment, build in public, and compete on genuine utility rather than marketing. The result was a user base across the US, Canada, India, and beyond that arrived through organic discovery, community engagement, and AppSumo deals that got the product into the hands of early adopters who could validate the core concepts.

The same philosophy carries into Mida. The product is free at up to 50,000 monthly tested users, which is sufficient for most SMEs and individual CRO practitioners to run meaningful experiments without paying anything. The pricing scales with usage rather than charging by feature, which removes a common friction point in the category — the sense that you need to upgrade to unlock functionality you actually need. The CRO agency directory, the free tool suite (hypothesis generators, sample size calculators, significance calculators), and the free A/B test idea bank are all demand generation assets that attract the platform's core users through genuine value delivery rather than advertising spend.

The customer case studies tell a consistent story: a 50% sales increase for Tile Giant through search results optimisation, a 47% conversion lift for Alpensattel, a 16% conversion rate increase for Dstrezzed, a 13% increase for NetworkLessons.com. These are not cherry-picked outliers — they're representative of what happens when a team that understands web experimentation has a tool precise enough to test meaningful hypotheses without introducing confounds through script weight.

What would have to be true for this not to work

The competitive intensity in the A/B testing category is real. The Google Optimize vacancy has attracted significant VC-backed alternatives — PostHog, VWO, and several newer entrants are all competing for the same migrating user base. Mida's differentiation on script performance is genuine, but performance alone doesn't guarantee retention; the platform also needs to keep pace on features as users' sophistication and requirements grow.

The product-led growth model also has a ceiling in this market. The practitioners who discover Mida organically through Google searches and AppSumo deals are often individual marketers or small agency teams. Converting them into larger, stickier accounts requires either expanding the enterprise features (which risks adding script weight and complexity) or building a strong agency channel — which Mida has begun with its CRO agency directory and agency-specific pricing. Executing that channel without losing the simplicity that defines the product is the key tension to navigate.

We backed this because the Google Optimize gap created a genuine, time-limited market entry opportunity; Mida's technical architecture is the right foundation to serve that market; and a founder who built a 3,000-company user base for Howuku with minimal marketing spend has demonstrated the product instincts to keep iterating toward the right product form. The A/B testing market is large, structurally changing, and Mida has the right DNA to compete in it.