Meet your community.

In a fictional community-booking product, a team is testing whether hosts need a shared equipment list. The prototype has been available for a while, but the relevant booking task happens infrequently. The team keeps extending the trial and adding small interface changes without deciding what the original question was meant to establish.

We are considering a reset: write the learning question, define a meaningful observation and agree what would justify continuing, changing the method or stopping. We also want to distinguish an unclear prototype from evidence that the task is unimportant. Participants should know what happens to their contributions when the trial ends, rather than finding a temporary screen left in place indefinitely.

What stopping condition has helped your team avoid a drifting experiment? How do you handle ambiguous evidence without using it as a reason to continue forever? This is a fictional demo scenario, not a report of actual product usage or customer participation.