UX and research · Beginner · 13 min
Report research findings so the team acts on them
By Padding · Updated 18 Nov 2026
Any document editor and a calendar · November 2026
You will have a short findings report that opens with the decisions it supports and backs each finding with evidence, plus a 15-minute playback plan that ends with a named owner for each recommendation.
On this page
A research report fails quietly. Nobody argues with it. People read the first page, agree it was interesting, and the roadmap stays the same. The usual cause is the order: the report starts with the method, spends three pages on who was recruited and how, and reaches the part a product manager needs (what should we change, and how sure are you) on page eight.
This tutorial turns that order round. You start from the decisions your team has to make, write the answer to each one first, and put the method at the back where anyone who wants to check it can. Then you plan a 15-minute playback, because most decisions happen in a room, not in a document.
The worked example continues the meal-kit study from the insights tutorial: eight interviews with people who paused or cancelled a weekly meal-kit subscription in the last three months, and the five insights that came out of it. Swap in your own study as you go.
Step 1 · List the decisions the research has to inform
Before writing a word of the report, find out what the people reading it are about to decide. Look at the next one or two planning cycles and ask the product manager, the design lead and whoever owns the metric your study touches one question each: "What are you deciding in the next six weeks that this research could change?"
Write the answers as a numbered list at the top of a blank document, in the form "Decide whether to…" or "Choose between…". Beside each, note who makes the call and when.
For the meal-kit study the list looked like this:
- Decide whether to build flexible delivery days before the spring campaign. Owner: product manager, at the planning meeting on the 3rd.
- Choose between shorter recipes and a quick-meals filter for the weeknight problem. Owner: design lead, this month.
- Decide whether the pause flow needs work, or whether people pausing is fine. Owner: head of retention, next quarter.
If nobody can name a decision, your report will be read as background. That is worth knowing now: say so to your manager and agree what the research is for before you put time into the write-up.
Check it worked
The top of your document has two to five decisions, each starting with a verb, each with a named owner and a date or meeting. You could read the list aloud to any of those owners and they would recognise their own decision in it.
Step 2 · Match each finding to a decision
Lay your insights beside the decision list. For each insight, write the number of the decision it bears on. Some insights will touch two decisions; some will touch none.
Insights that touch no decision are not wasted. They go in a short "Also worth knowing" section near the end, or into the repository for later. They do not go at the front, however surprising they were to you.
Then look the other way: is there a decision with no finding beside it? That is a gap, and it becomes an open question in a later step rather than a guess dressed up as a finding.
The meal-kit study's five insights mapped like this:
- "Food arrives when people cannot use it" (7 of 8, High) bears on decision 1.
- "Total time, not cooking time" (5 of 8, High) bears on decision 2.
- "The skip deadline arrives before people know their week" (2 of 8, Low) is the only insight that touches decision 3, the pause flow. Two participants is too little to settle that decision, so decision 3 is marked as a gap and becomes an open question.
- "Price is compared per meal, not per box" (7 of 8, Medium) touches none of the three decisions, even though more participants mentioned it than anything else. It goes in "Also worth knowing", with a note to whoever owns pricing.
- "Mixed diets break the plan" (2 of 8, Low) is about why people cancel, which nobody is deciding on this cycle. It also goes in "Also worth knowing".
A well-supported insight with no decision beside it is still not a headline. Count tells you how sure you are; the decision list tells you whether anyone needs it now.
Check it worked
Every insight in your list has either a decision number or the label "Also worth knowing" beside it. Every decision has at least one insight, or is marked as a gap.
Step 3 · Write the summary page first
The first page of the report is the only page some readers will see, so write it before anything else and make it stand on its own. Use this structure:
- What we set out to learn: the research question in one sentence.
- What we did: one sentence. "Eight 45-minute video interviews with people who paused or cancelled between August and October."
- What we recommend: one line per decision from the first step, in the same order, each giving the answer and how confident you are.
- What we do not know yet: one or two lines.
The recommendation lines carry the weight. Write them as answers, not topics:
Weak: Delivery flexibility came up a lot.
Better: Build flexible delivery days before the spring campaign.
Seven of eight people had boxes arrive on days they could
not cook them. High confidence.Give confidence in plain words (high, medium, low) and one clause on why. "High: seven of eight, and it matches what support tickets already say" is more useful to a reader than a percentage on a sample of eight.
Check it worked
Your summary fits on one page. Cover everything below it and read it as the product manager would: you can tell what to do about each decision, and how sure the researcher is, without turning the page. No method detail appears beyond the single "What we did" sentence.
Step 4 · Write each finding with its evidence
After the summary, give each finding its own short section, in the same order as the decisions. Each section has four parts:
- The finding as a heading, written as a full sentence: "Food arrives when people cannot use it."
- What we saw: two to four sentences describing the behaviour, with the participant count ("seven of eight").
- Evidence: two or three quotes from different participants, each tagged with the participant code (P2, P5) and a link or reference to the source notes. Pick quotes that show the behaviour, not ones that only express a feeling.
- What it means: one or two sentences connecting it to the decision.
Keep counts honest. With eight participants, write "five of eight", never "62.5%". Percentages on small samples read as survey results and invite the wrong kind of argument.
Add counter-evidence where it exists. Under "Total time, not cooking time", "P7 worked from home and cooked every recipe on the day it was planned" is one line, and it makes the rest of the section more believable. A finding with no exceptions at all is a warning sign to a careful reader.
Check it worked
Each finding section has a full-sentence heading, a participant count written as "n of N", at least two quotes from different participants, and a sentence linking it to a decision. You can follow every quote back to its source in under a minute.
Step 5 · Turn findings into recommendations people can own
A recommendation the team can act on names the change, not the problem, and is small enough that someone can put it into their next plan. Write a short list at the end of the findings, one recommendation per item, each with four parts: the change, the finding it comes from, a rough effort guess, and a suggested owner.
1. Let people pick a delivery day each week, up to the Thursday before.
From: Food arrives when people cannot use it. Effort: medium.
Owner: product manager.
2. Add a "ready in 20 minutes, including preparation" filter to the menu.
From: Total time, not cooking time. Effort: small. Owner: design lead.
3. Interview five more people who paused or missed the skip deadline.
From: Gap on the pause flow (skip deadline insight, 2 of 8).
Effort: small. Owner: head of retention.Mark effort as a guess and leave it to the people who will build it to correct. Your job is to make the recommendation concrete enough that they can estimate it, not to estimate it for them.
Then add two short sections after the list:
- Open questions: what this study could not answer, and how you would find out (another study, analytics, a test).
- How we did it: research question, method, number and type of participants, recruiting criteria, dates, and any limits ("everyone recruited lived in a city"). Link to the full discussion guide and notes.
Check it worked
Each recommendation starts with a verb, points back to a named finding or gap, and has a suggested owner who also appears in your decision list. The method section is the last section in the document, not the first.
Step 6 · Plan a 15-minute playback
Book 30 minutes with the decision owners from the first step and plan to use half of it. The other half is for discussion, which is where commitments are made. Send the report the day before with one line in the invitation, such as "Two decisions to make in this session: delivery days and weeknight recipes."
Write the plan as a short run sheet:
00:00 Why we did this and the decisions on the table 1 min
01:00 What we did, in one sentence 1 min
02:00 Finding 1, one clip or quote, recommendation 4 min
06:00 Finding 2, one clip or quote, recommendation 4 min
10:00 Open questions and what we will not decide today 2 min
12:00 Recommendations: ask for an owner for each 3 min
15:00 DiscussionFor each finding, choose one piece of evidence that does the persuading. A 30-second video clip of a participant describing the moment they cancelled lands harder than a slide of quotes, provided your consent form allows you to show recordings to the team. If it does not, read one quote aloud in the participant's own words.
Leave findings without a decision out of the playback. Point to them in the report instead.
Check it worked
You have a run sheet that adds up to 15 minutes, one piece of evidence chosen per finding, and a calendar invitation sent to every owner named in your decision list. You have checked your consent forms cover any recording you plan to play.
Step 7 · Run the playback and write down who owns what
During the session, keep to the run sheet and let questions wait until the 15-minute mark unless someone is lost. When you reach the recommendations, go through them one by one and ask: "Who will take this, and when will we hear back?" Type the name and date beside each one on screen as people answer.
Some recommendations will be turned down. Record that too, with the reason ("not before the spring campaign; revisit in Q2"). A recorded no is a decision; a polite nod is not.
Within a day, send a short follow-up message to everyone who attended:
- the decisions made, with owners and dates
- the recommendations declined or parked, with the reason
- a link to the report
Then file the report where the team will find it later, alongside the notes it cites, so the evidence outlives the meeting.
Check it worked
Every recommendation in your report now has a name and a date, or a recorded "declined" with a reason. Your follow-up message has gone out, and someone who missed the session could read it and know what was decided.
Next steps
- If your insights are still loose themes, sharpen them first: Turn research themes into insight statements.
- Store this report and its evidence so the next study can build on it: Keep research findings in a searchable repository.
- Check you are allowed to show clips before the playback: Get informed consent and store research data safely.
- Plan the follow-up study for your open questions: Write a one-page research plan.
Part of these paths
Related tutorials
- Turn research themes into insight statementsYou will have five insight statements, each saying what people do, why they do it and what it means for the design, with quotes from several participants and a confidence level beside it.
- Keep research findings in a searchable repositoryYou will have a small spreadsheet repository holding every finding from one study, each row tagged by topic, study and date and linked to its evidence, plus a tag list and house rules a colleague can follow without asking you.
Need this done for you?
Padding does this work for growing companies, from the first diagnosis to a system that keeps working.
Get new tutorials