Bubble

At Bubble, we use DoesQA to automate testing across some of the most critical areas of our app, which is the entry point and backbone of our business that serves 6M+ users. It gives us reliable regression coverage, reduces repetitive manual testing for each release, and gives us greater confidence that changes will not create unexpected consequences elsewhere in the product.

Before DoesQA

Before DoesQA, we were lightly using synthetic testing tooling, which let us automate some coverage, but the reliability was never good enough to trust in our CI pipeline. False positives meant we still spent time sorting real failures from noise.

Before that, my team had no automated testing on the Bubble app, even though other engineering teams in the wider codebase already had robust suites. Without reliable automation, we manually tested every iteration of a flow, especially where branching logic mattered, such as new-user onboarding.

What we needed was automated regression testing we could rely on.

Building our suite in DoesQA

With DoesQA, we can automatically sweep critical areas of our app for regressions. We have built 23 Flows with around 800 Test Steps, covering onboarding, payments, SSO, and Bubble AI. Those Flows mix browser interactions with Integration API steps, plus specialized journeys such as multi-factor authenticated logins and scanning QR codes.

As a Bubble user, DoesQA’s interface feels incredibly familiar. We design Flows in DoesQA's Flow Builder, an infinite canvas where we drag Test Steps and use Flow Branches for branching paths. Step Groups work much like Bubble's reusable elements, so shared logic stays modular, and a change in one place updates every Flow that uses the group.

Fitting DoesQA into how we work

Due to the nature of our work, we don’t ship through GitHub with a conventional Actions-based CI/CD pipeline. Instead, we have built our own pipeline from tooling and internal processes, so flexibility mattered. We built tooling that triggers DoesQA suites from Slack and returns the results when a Run finishes, which made our test suite extremely convenient.

We also use DoesQA for periodic health checks alongside other monitoring. Around three quarters of our Runs are now on a Schedule, with the rest triggered manually when needed. In a typical month, the suite runs about 650-850 times.

Less manual testing, more confidence

A reliable suite has changed where we spend time before a release. Instead of manually testing every iteration of every critical flow, we lean on DoesQA for broader regression coverage and focus manual effort on the areas we are explicitly changing. I would estimate that saves the team about 5-20 hours of manual testing each month. The bigger win is confidence: we can change critical paths knowing they are tested consistently, with scheduled health checks adding reassurance between releases.

Finding the unexpected

While preparing to launch a new AI-chat onboarding experience, we ran our DoesQA suite and saw a handful of failures. Someone on my team found that the tests had exposed a race condition elsewhere in the codebase, outside our normal surface area. It had been quietly creating duplicate metadata records.

The user-facing impact at the time was limited, and the issue was relatively rare in production. Left unnoticed, though, it could have resulted in growing data inconsistency as usage of the new experience increased. It is the kind of error that would have been nearly impossible to catch by hand.

Working with the DoesQA team

Support from the DoesQA team has been an important part of the experience. During evaluation, they gave us product walkthroughs. As we built the suite, they gave hands-on advice on how to structure and use the product well, which helped us get over the initial barrier of rebuilding our suite.

We now have a suite that runs hundreds of times each month, covers some of Bubble's most critical areas, and sits inside our wider release and monitoring process. Ultimately, it gave us something we did not have before: reliable automated regression testing we can genuinely build our processes around.