Case study — Funnel Diagnosis
Diagnosing why a majority of new partners stalled before their first listing went live
Emerging Travel Group
Every abandoned registration was a property that never joined the marketplace
Registering a new property is the first step in a partner's entire relationship with the platform. But a large share of people who registered an account never created a listing at all — and of those who did start, a substantial portion never finished. Nobody had a clear picture of exactly where people were getting stuck, or why.
Data first, then qualitative — in that order, deliberately
I started with quantitative discovery rather than jumping straight to interviews, because the team didn't yet know where in the funnel people were dropping — guessing at qualitative questions before that would have wasted respondents' time. I built custom SQL views to segment the funnel by step, device, property type, and whether it was a partner's first or subsequent listing.
Only once the data pointed clearly at two friction points — contract signing and identity verification — did I move to a mixed qualitative and quantitative validation phase: moderated interviews to understand why, and a structured survey to confirm which issues generalized across the wider partner base. (71 respondents is a strong sample here — this is a B2B partner population, not a mass-market consumer audience, so the addressable pool is inherently smaller than a typical B2C survey.)
What the data showed, step by step
Two steps carried almost all of the friction
Only about a third of respondents understood why the platform needed their documents — and the verification process itself was often physically awkward, not just confusing.
"I had to record myself twelve times because the head-turn movements didn't match what was required — very difficult to turn my head that way."
"I had to go outside to film part of the verification video, then film inside — but there were guests staying at the time, so it felt very awkward to do in front of them."
Around a third of partners simply didn't have the paperwork on hand when they reached this step — a preparation problem the product could soften with a simple checklist earlier in the flow.
A third of respondents lost entered data; a fifth couldn't find a listing they'd started earlier. Small, unglamorous, and entirely fixable.
Most comparable platforms keep identity verification outside the main registration wizard — treating it as a final, optional confirmation rather than a blocking gate. This directly shaped a follow-up proposal to test the same approach.
Two teams acted before the research had even finished
Once the funnel analysis alone was ready, two separate teams picked up work directly from it: retention marketing revisited onboarding messaging, and the data/analytics function added error logging in the registration wizard — before the qualitative or survey phases were even complete.
The full set of findings shaped a prioritized backlog: draft-saving, cross-device continuity, clearer identity-verification copy, and a "what you'll need" checklist ahead of the contract step.
The project took longer than I'd planned — partly a deliberate trade-off. Analytics support was scarce at the time, so I agreed with the product owner that I'd run the SQL analysis myself rather than queue for a dedicated analyst. The right call given the constraints, but it stretched the timeline. Next time, I'd time-box the analytical phase more tightly and bring in a data partner earlier, so qualitative and quantitative work could run in parallel rather than in sequence.
The shipped changes were minor relative to the full backlog, and conversion wasn't tracked afterward — so there's no clean before/after number to point to here.