Every sprint we run ends with one of three words: Build, Reframe, or Kill. The first two are easy to say out loud in a meeting. The third one, telling a team that spent a grant cycle wanting this thing that the need underneath it is weaker than it looked, is not.
We say it anyway, and we’ve come to believe it’s the most valuable of the three.
The incentive problem with “yes”. Most vendors are structured, deliberately or not, to say yes. A build shop is paid to build; an agency’s next contract depends on this one going well; a consultancy’s report is easier to write if it confirms the premise it was hired to explore. None of these incentives are evil. They’re just misaligned with the thing a research team actually needs, which is an honest read on whether an idea deserves the next six months of budget.
We’re structured differently, on purpose. Our commercial relationship with a client is the sprint, not the build that might follow it. That means a Kill verdict costs us nothing we were counting on, and a Build verdict we can’t back up with evidence costs us the thing we’re actually selling: a reputation for telling teams the truth.
What a Kill verdict actually looks like. It is not “this was a bad idea” delivered as a eulogy. It’s specific: here is the need we tested, here is what we observed when real people tried to use a real prototype, and here is why the need turns out to be smaller, differently shaped, or already served by something else. A good Kill verdict usually surfaces exactly what would have to be true for the idea to work, which is often useful information on its own.
The counterfactual is the point. Picture the alternative: a year-long build, a grant spent, a team’s credibility with their funder on the line, all in service of something that, had it been tested in week two, would have shown the same weak signal it showed in our sprint. A Kill verdict two weeks in is not a loss. It’s the six-figure mistake that didn’t happen, and the honest sentence a PI gets to say to their board instead of a difficult one eighteen months from now.
Why this is a trust argument, not a marketing one. We don’t publish Kill verdicts to look modest. We publish them because a studio that never says Kill hasn’t been tested yet, and a research sector allergic to hype can tell the difference between a vendor that’s confident and one that’s just agreeable.
If a sprint with us ends in Kill, we’ll tell you why, in detail, and we’ll mean it as a good outcome. That’s the deal.