Cobl solves a problem every sales team knows: writing proposals. You hand it your call notes, your website, and your branding, and it generates a full presentation in one shot, which you then edit to the last mile. Miguel Schuppan is the product manager behind it, and when he joined he was given an unusual mandate. Not just ship faster. Ship things that are right.

When You Can't Just Call Your Five Biggest Customers

Cobl used to sell to corporate and enterprise clients. In that model, listening to users is a calendar entry: a handful of accounts, weekly calls, requirements written down and implemented. Then the company pivoted to product led growth, with many more users and no standing meetings with any of them.

Miguel had seen the failure mode of building without that feedback loop, in his own work and in the wider product world he had studied closely.

Proper discovery work, user interviews, and user testing were new to Cobl as a company. Miguel had management buy-in to build that half of the process, so he went looking for a tool.

8 Platforms, and Half Ruled Out Before the Free Trial

Miguel ran a full benchmark of around eight user testing platforms. Half of them eliminated themselves immediately, not on features but on how they sold.

The enterprise options were also priced for a different kind of company. One quoted an entry level of $15,000 for 70 credits, enough for roughly seven hour-long sessions, which by Miguel's calculation made it about 10 times more expensive than what he ended up with.

His requirements were specific: self-service, affordable monthly pricing with no year paid up front, and above all the ability to run tests on a live website. Half the tools could not do that properly, or only worked with apps. One platform he tried before Userbrain was built around an embed rather than a browser plugin, which meant fighting cookies and authentication to get a session running at all. It never worked.

Userbrain's plugin worked the first time. Miguel then did the unit economics, worked out the cost per session, bought around 100 tests in bulk, and started.

Hypothesis Monday Morning, Answers After Lunch

What that unlocked is a loop rather than a project. A hypothesis on Monday morning becomes a quick prototype, and the prototype is in front of real testers before the team breaks for lunch, with the target audience narrowed by screening questions to the salespeople Cobl builds for.

That speed changed the shape of the week. Testing stopped being a phase that delays a release and became a step inside it.

The Failure Nothing Else Would Have Caught

A month after launch, Cobl had healthy traffic, signups, and usage. Every dashboard said things were working. Miguel ran a short test with 10 users on the core flow, the one where the product does the thing it exists to do and generates a document.

Two of them got there.

That is the case for watching real people rather than reading aggregate numbers, in one afternoon. Analytics told Cobl how many people arrived. Only the recordings told them how many walked away, and why.

Evidence the Whole Team Wants

Testing at Cobl is no longer Miguel's private practice. The tech team, the product designer, and the CEO all use what comes out of it, with the CEO going straight to the AI Insights summaries for the bird's eye view of what worked and what did not. Sessions cost at least 50% less than the next best option Miguel benchmarked, which makes the internal argument short.

Results at a Glance