2learn

From findings to decisions

What it is

Turning research into decisions means compressing sessions of notes into a small number of findings, each tied to evidence, and then attaching a change and an owner to each. It is the stage where research either pays for itself or becomes an interesting document nobody acts on.

Why it matters

Research that ends in a report changes nothing: the report is read once and the next sprint continues as before. The value appears only when a finding becomes a line item someone owns. This stage is also where the team without the researcher present gets convinced — a finding shown with two quotes and a short clip is far harder to argue with than a summary paragraph.

Core ideas

Work from evidence to finding, not the reverse

Group the raw observations first, then name what the group means. If you start from a finding and look for quotes, you will find them whether or not they represent the sessions.

Observations (14 from 6 sessions):
 "scrolled past search three times"      x5
 "used the recent tab before searching"  x6
 "asked a colleague for the order number" x3
Finding: people reach the last order by browsing, not searching.

Keep the count and the words

Report how many participants did something and show at least one verbatim quote. “Several users” cannot be checked; “five of six did X” can, and it survives the meeting where somebody disagrees.

Rank by frequency times severity

Two axes: how often it happened, and how much it cost when it did. A rare, total block outranks a frequent mild annoyance. Write the ranking down so it can be argued with.

1 payment failure, no recovery path   rare x total block    -> fix now
2 address field asks for company no.  common x 40s detour   -> fix soon
3 label wording on the second button  common x mild         -> fix if cheap

Turn each finding into a change with an owner

A finding is not actionable until it names what will change and who will do it. “Improve onboarding” is not a change; “make the phone number optional and re-run the test next month, owner: Maya” is.

Present it in the room the work is decided

Show the findings where priorities are set, not in a separate readout. Bring one clip, one quote and the count.

Practice

Cluster one study’s notes

Take the notes from one study — at least five sessions. Write every observation on a separate line, group them, and name each group in the participants’ words.

Done when: every finding has a count, one verbatim quote, and at least one observation that contradicts it noted beneath.

Write the decision table

For each finding, write the change, the owner and the date, and rank the rows by frequency times severity.

Done when: the table has no row whose change is a verb, and no row without a named owner and date.

Common mistakes

  • A report instead of a change — the document is filed and the product is untouched; fix: attach a change, an owner and a date to every finding.
  • Losing the count — findings cannot be checked and get argued away; fix: report how many of how many, and quote at least one participant.
  • Cherry-picking quotes — the evidence is selected to fit a preferred answer; fix: cluster first, name second, and note contradictory observations.
  • Ranking by loudness — the last stakeholder to speak sets the priority; fix: rank by frequency times severity before the meeting.
  • Presenting to the wrong room — the people who decide priorities are absent; fix: take the findings to the planning session, not a separate review.

Go deeper

Worked examples

Analytics cookies are set only if you allow them. Nothing is loaded, and nothing is sent, until you choose.