The most common fate of a study is not being wrong. It is being filed. Here is the only metric of research value I trust, and the three questions that keep a study alive.
Every researcher I know has a folder like this. Well-run studies, good samples, clean reports — and no trace of any decision that changed because of them. The report was presented, thanked, and archived. Nobody argued with it. That was the problem: nobody needed to.
I used to think this was a failure of communication. Better slides, sharper insights, a tighter executive summary. I no longer think that. A study dies on the drive when it was never attached to a decision in the first place, and no amount of polish attaches it afterwards.
Ask any research team how they measure their value and you'll hear about insights delivered, studies completed, stakeholders satisfied. I use a cruder number: how many decisions were made after the research that would have been made differently without it. Count them. If the answer is zero, the time and money were most likely wasted, however good the report looked.
This metric is unkind to researchers, and it is meant to be. It is also unkind to organisations, because a zero often means the decision was taken before the study was commissioned and the study was ordered as decoration. Both failures produce the same folder.
There's a second test I apply to my own work. If a study brings the CEO a pile of things about customers that were entirely unknown to them, I get suspicious of the study. Insights can be things people half-guessed and can now confirm, or things they suspected and can now drop. But if the whole report is shock and novelty to the person running the company, then either the research was aimed at the wrong questions — or leadership is so detached from its own customers that no report will fix it. Either way, filing is the likely outcome.
What keeps a study alive is not method. It is being anchored to a decision at three points in time.
Before, one question: what decision could change depending on the answer? Not "what do we want to learn" — that produces curiosity, and curiosity produces folders. Name the decision. Name who owns it. Name the options that are actually on the table. If nobody can name a decision that could go differently, don't run the study; run a smaller one, or none. "Why do users churn?" is a research question. "Which of the reasons for cancelling in the first three days can we realistically change through onboarding, which are expectation problems from acquisition, and which are structural?" is a decision.
During, one question: what evidence would actually change this decision? Ask it of the decision owner before the data exist, and write down the answer. This is the most protective sentence in research practice. It commits the owner to a threshold while they are still uncertain. And it exposes the studies that are decoration: if the honest answer is "nothing would change it", you have learned something about the decision that the study itself never could.
After, one question: what changed? Not "was it well received". What changed. A roadmap item moved, a price held, a launch delayed, a segment dropped. Write it next to the study. Over a year that list is the only research-impact report that means anything, and it is the one that stops the folder from growing.
Research teams are organised to produce reports, so they produce reports. But a report is an intermediate artefact. The product is a decision that is better than it would have been. If you measure the artefact, you get beautiful artefacts and a drive full of them.
This is the argument behind the playbook we've just published — research as a route from a business question to a decision, with the dead ends drawn in rather than smoothed away. It exists because I got tired of teaching people to run interviews well and then watching the interviews die in a folder like everyone else's.
If you've got a folder like that, you already know what I mean.
The playbook: So They Lied to You / «Так тобі й збрехали»
Nina Somkina · Founder, Institute for Human Reasoning
Institute for Human Reasoning is a research and education organisation strengthening human reasoning for consequential decisions in a technology-shaped world.
Seventy years of research on evidence and decisions, in six findings that still embarrass every company calling itself data-driven.
ReadEvery company will say yes. The only measurement that means anything is where AI enters a decision — and whether anyone checks it when it disagrees with the boss.
Read