You have the research. How do you turn it into a decision?

There is a particular kind of stuck that follows a finished piece of research. You know more than you did. You can describe the competitors, the pricing, the gaps. And yet the decision the research was meant to serve is exactly where it was a fortnight ago, because knowing more and deciding are different acts, and only one of them was on the plan.

Every other guide in this library ends at the finding. This one starts there.

Write down the decision before you read the findings

The single most useful habit is also the least natural: state what you are deciding, and what would move you each way, before you look at the evidence. One paragraph is enough. The decision, the options including the option of doing nothing, and for each option, roughly what you would need to see to choose it.

Do this afterwards instead and you will almost certainly find the evidence agreeing with you. Not because you are dishonest, but because a set of findings is ambiguous enough to support several readings, and the one that arrives first is the one you already held. Written-down criteria are the only cheap defence against that.

If you have already done the research and skipped this step, you can still recover most of the value. Write the criteria now, but write them as a stranger would — or better, have someone who has not read the findings write down what would convince them. Then check your findings against that.

Sort every finding into observation, inference, or assumption

Go through the research line by line and mark each claim as one of three things:

• Observation — something you can point at. A price on a page, a review count, a job posting, a filing. Someone else looking today would see the same thing.
• Inference — a reasonable reading of observations. "They are moving upmarket" is an inference from a price change plus a change in the case studies they publish.
• Assumption — something you believe but have not checked, often so basic it never got written down. "Customers in this market care about turnaround time" is usually an assumption.

The sort matters because these three carry completely different weight and are almost always presented in the same voice. Once separated, two things usually become visible: the decision is resting on fewer observations than it felt like, and one or two unexamined assumptions are doing most of the structural work.

Those load-bearing assumptions are your real research agenda. They are worth more attention than another round of competitor detail.

Match the evidence bar to how reversible the decision is

Not every decision deserves the same confidence. The useful question is not "am I sure?" but "what does it cost me to be wrong, and can I undo it?"

A reversible, cheap decision — trying a different headline, testing a price with the next five enquiries, writing one article aimed at a new segment — should be made on thin evidence, quickly, because running it is cheaper than researching it. Waiting for certainty here is the expensive mistake, and it is the more common one.

A decision that is hard to unwind — a rebrand, a hire, a long lease, dropping a product line, a rewrite — earns a higher bar and a slower process. For those, the question worth asking is whether some part of the commitment can be made reversible first: a contractor before a hire, one campaign before a repositioning, a landing page before a build.

Most decisions people agonise over are more reversible than they feel in the moment, and a few that get made casually are not reversible at all. Sorting them honestly is most of the work.

The third verdict people forget

Act, or do not act, are the two verdicts everyone reaches for. The third is buy more information — and it is a legitimate answer only if it comes with three specifics: what exact question you are answering, what it will cost in time or money, and what you will do differently depending on the answer.

That last one is the test. If both answers lead to the same action, the information is not worth buying no matter how interesting it is. This is the line between genuine diligence and comfortable procrastination, and they are indistinguishable from the inside without it.

Give the question a deadline too. "I will spend three days on this and then decide with what I have" converts an open loop into a bounded one, and it forces the decision to happen even if the answer is unsatisfying.

Ask what would have to be true

When a decision is genuinely close, invert it. Instead of asking whether the evidence supports the option, ask: what would have to be true for this to be the right call? Then look at that list and ask which items you have actually checked.

This surfaces the difference between a weak option and an option resting on one unverified belief. The second is often worth pursuing — you go and check the belief. The first is not, and inversion is what tells them apart, because both feel the same when you are only weighing evidence for and against.

Reduce the finding to the smallest action that tests it

A finding is not a plan. "Nobody in this market publishes pricing" is an observation; the decision it implies could be anything from publishing your own prices to concluding that the market has a reason not to.

Ask what the smallest version of the action is that would produce real evidence. Not a pilot in the corporate sense — the genuinely small thing. One page. One week. One offer to the next ten enquiries. If the finding is right, the small version should show some signal; if it shows nothing, you have spent very little to learn that.

The value of the small version is not caution. It is that it converts an argument about what the market will do into an observation of what it did.

Set a review date and a tripwire

Write down when you will look at the decision again, and what would make you look at it sooner. The tripwire is the more useful half: a specific observable event — a competitor launching the thing, enquiries dropping below some number, a supplier changing terms — that means the reasoning no longer holds.

Without a tripwire, decisions quietly outlive their evidence. The research that justified the choice was true when you did it, and nobody ever schedules the meeting where it stops being true.

Where this goes wrong

• Deciding first, researching second. If you cannot describe what evidence would have changed your mind, the research was decoration.
• Following the loudest finding. The most surprising or vivid item in a report is rarely the one the decision rests on. Sort by load-bearing, not by interesting.
• Treating the report as the deliverable. A research document is finished when a decision changes or is deliberately confirmed. Filing it is not finishing it.
• Sunk research cost. Effort spent gathering evidence is not a reason to act on it. "We have done all this work" is not a finding.
• Waiting for certainty on a reversible call. The cost of a slow decision is real, gets paid daily, and never appears on anyone's list of risks.
• Confusing a finding with a cause. Noticing that the leading competitor does something does not tell you it is why they lead. It might be a habit they can afford.
• Never writing the decision down. If it exists only as a conversation, in three months nobody will agree on what was decided or why, and the reasoning cannot be reviewed when the tripwire fires.

What good looks like

A short written note: the decision, the two or three observations it rests on, the assumptions you are knowingly accepting, what you expect to happen, when you will review it, and what would make you revisit sooner. Half a page.

That note is worth more than the research folder, because it is the only artefact that can be checked against reality later. When the review date arrives you can see exactly which part of the reasoning held and which did not — and that is the thing that makes the next decision better, rather than merely better-informed.