The slide was called Barriers to Purchase. By the second revision it was called Friction Points. By the version that actually got presented on Wednesday it was called Considerations, and the bullet underneath it said customers expressed some uncertainty during the final step.
What eleven of fourteen participants actually said was that the checkout flow shipped in September made them think they had been charged twice, so they closed the tab. The person who approved that flow was sitting at the head of the table, and every softening edit had happened between the first draft and his calendar invite.
Nobody in that chain did anything unethical. That is what makes it so hard to fix.
A debrief that flatters everyone in the room has already failed, because the only findings that survive that filter are the ones nobody needed research to discover.
I have sat on both sides of this. I have written the sanded-down version, and I have been the person at the head of the table whose decision the data was quietly indicting. Neither seat feels like cowardice from the inside. It feels like tact. It feels like protecting the relationship so you get invited to the next study. That instinct is exactly how a company spends sixty thousand dollars to be told what it already believed in October.
Three things help, and they are structural rather than brave, which is the only kind of fix that survives contact with a real client.
One, get the room to pre-commit. Before fieldwork, ask everyone what result would change their mind, and write it down with names attached. A finding is much harder to soften when the person doing the softening agreed in advance that this exact result would matter.
Two, do not paraphrase. Play the clip. A participant saying I thought it charged me twice, in her own voice, with her own face, cannot be edited down into expressed some uncertainty. Attribution to a real human being moves the finding out of the analyst's opinion column and into the record.
Three, separate the finding from the verdict, out loud and on purpose. The checkout flow is costing conversions is a finding. Shipping it was a mistake is a verdict. Research owns the first and has to visibly refuse the second, because the fastest way to make a room defensive is to let those two collapse into one sentence.
It feels like tact. It feels like protecting the relationship so you get invited to the next study.
I sing at a bar in Southeast Portland twice a month, badly and with complete commitment. The host who calls everybody baby told me once, very kindly, that I come in early on the chorus every single time. I do. I had been doing it for three years, in front of the same people, and not one of them had ever mentioned it.