A research report is the last mile of your study. You've written the survey, recruited participants, and collected 200+ responses. The insight is in the data. And then it dies — in a Google Doc that nobody opens, a spreadsheet that only you can read, or a presentation that took 4 hours to build and was outdated by the time you sent it.
This article gives you two things: a free template that structures your report the right way, and a walkthrough of how Lumor's AI Doc Report fills it automatically so you can spend that 4 hours doing something that actually requires your judgment.
Why most research reports fail stakeholders
The average research report is written for researchers, not for the people who need to act on it. It leads with methodology. It buries the headline finding in section 4. It includes every data point because the researcher is reluctant to discard work. And it uses language — "statistically significant", "p-value", "n=47 subset" — that creates distance between the reader and the decision.
Stakeholders — PMs, designers, founders, marketing leads — don't need comprehensiveness. They need clarity. They need to know, in the first 60 seconds of reading, what you found, what it means, and what you think they should do about it.
The template below is structured around that reading behaviour. The most important content comes first. Supporting evidence comes after. Methodology and raw data are at the end, for those who need to verify — not for those who need to decide.
📐 Design principle
Structure your report the way a newspaper structures its front page — the headline, then the lede, then the supporting detail. Not the other way around. Your stakeholders read like journalists scan headlines, not like academics read abstracts.
Download the free template
Free Download
Lumor UX Research Report Template
A ready-to-use Google Doc template structured for stakeholder clarity. Eight sections, pre-written headings, placeholder guidance on every section, and an AI-ready layout that Lumor's Doc Report can auto-populate.
No email required. Word document opens immediately.
What each section does and why it's there
The template has eight sections. Here's what each one does — and the specific mistake it's designed to prevent.
3 bullet points. Top finding, top implication, recommended action. This is the only section most stakeholders will read in full. Everything else supports it.
What were you trying to find out and why? One paragraph. Prevents the report from being read in isolation — gives stakeholders the "why this study" context before the "what we found".
How did you run the study? Who responded? Brief and factual. This section is for credibility and reproducibility — not the main event. Keep it under 150 words.
One finding per sub-section. Each finding is a declarative statement — not a chart description. Chart or data point follows the finding statement, not the other way around. This is the inversion that most reports get wrong.
Open-text responses grouped into named themes. 3–5 themes maximum. Each theme gets a name, a 1-sentence description, and 2 representative verbatim quotes. Verbatim quotes are the evidence that brings numbers to life.
Where do different audience groups diverge? The most interesting findings are often in the deltas — not the averages. This section surfaces the "but for this group..." insight that changes product decisions.
3 specific, actionable recommendations written in plain language. Each one references the finding it comes from. Avoids the generic "do more research" recommendation that every report ends with. One recommendation per finding.
Everything that belongs in the report for completeness but not for the main reading experience. Full question list, response distributions, methodology details, survey link. Optional — most stakeholders never open this section.
Filling it manually vs with AI
There are two ways to use this template. The first is manual — you fill in each section yourself, pulling data from your survey tool and writing the findings and recommendations from scratch. The second is automated — you run your survey in Lumor and let the AI Doc Report populate the template for you.
Here's the honest comparison:
✕ Manual workflow
- Export CSV from survey tool
- Open spreadsheet, calculate percentages manually
- Copy-paste data into the template one section at a time
- Write findings and recommendations from a blank page
- Format charts, resize images, align layout
- Search through open-text responses to find good quotes
- Identify themes in qualitative data by hand
- Total time: 3–5 hours for 200 responses
✦ AI Doc Report workflow
- Close survey in Lumor
- Click Studio → Doc Report
- AI reads all responses and identifies key patterns
- AI writes finding statements and recommendations
- AI surfaces top verbatim quotes from open-text
- AI groups themes from qualitative responses
- Review, edit, add your editorial layer, export
- Total time: under 15 minutes
✓ The right approach
Use the manual template if you're analysing data from a tool other than Lumor, or if you prefer to write the report yourself first and use AI as a review pass. Use the AI Doc Report if you're working in Lumor — it's the same template structure, populated automatically. Either way, your editorial judgment on findings and recommendations is irreplaceable.
Step-by-step: the AI Doc Report workflow
If you're running your survey in Lumor, here's exactly how the AI Doc Report workflow works — from survey close to finished document.
Close your survey
In the Data Collection tab, toggle your collector offline. Once closed, responses are locked and the AI can process the full dataset without incomplete responses skewing the analysis.
Open the Result tab → Studio
Navigate to the Result tab on your survey. On the right side you'll see the Studio panel. Select "Doc Report" (as opposed to "Slides" — Doc Report produces a formatted written document, Slides produces a presentation deck).
AI generates the full report
Lumor's AI reads every response — quantitative and qualitative. It calculates response distributions, identifies the top statistical patterns, groups open-text themes, and selects the most representative verbatim quotes. This takes 60–120 seconds depending on response volume.
Review the eight sections
The output mirrors the template structure above — Executive Summary, Research Objective, Methodology, Key Findings, Qualitative Themes, Segment Comparisons, Recommendations, Appendix. Read through once with your stakeholder's eye. Is every finding a declarative statement? Is the executive summary tight?
Add your editorial layer
The AI handles production. You handle judgment. Edit the Recommendations section to reflect your team's specific context. Add the "so what for our sprint" note to the executive summary. Flag any finding where the sample size is too small to be conclusive. This is the 5–10 minutes that makes the difference between a document and a decision.
Export as PDF or share via link
Download as a formatted PDF for email distribution, or share a live link directly from Lumor. Team members with view access can read the report without downloading anything — useful for async review before a stakeholder meeting.
5 tips for a report stakeholders actually read
Whether you're using the manual template or the AI Doc Report, these five habits separate research communication that changes decisions from research communication that gets archived.
- 1.Lead with the finding, not the method. — Move your methodology section to the back. Put your top finding in the first 100 words. If a stakeholder reads nothing else, they should walk away knowing your most important result.
- 2.Use a declarative headline for every finding. — "Section 3: Payment step data" is a label. "Users abandon at payment 3× more on mobile than desktop" is a finding. Your section headings should be findings, not topic names.
- 3.Include exactly one verbatim quote per finding. — One well-chosen quote does more work than 10 data points. It makes the number human. Pick quotes that are specific, vivid, and representative — not quotes that are extreme or unusual.
- 4.Write your recommendations as actions, not observations. — "Users find onboarding confusing" is an observation. "Remove step 3 from the onboarding flow and test a 2-step alternative" is a recommendation. If your stakeholder can't do something with it, it's not a recommendation.
- 5.State what this study does not answer. — Every study has a scope. One sentence — "This study does not address mobile vs desktop split or regional variation" — sets the right expectations and pre-empts the meeting question that would have derailed the conversation.
Key takeaways
- 1.Structure your report for the reader who will act on it, not the researcher who wrote it. Decision-makers read like journalists, not academics.
- 2.The eight-section template above is ordered by stakeholder reading priority — executive summary first, raw data last.
- 3.AI Doc Report in Lumor produces the same structure automatically. Your job is the editorial layer: context, judgment, and the "so what for our team" annotation.
- 4.The gap between "data in" and "report shared" should be measured in minutes, not days. Research that arrives late is research that doesn't influence the decision.
Somnath Chakraborty
Co-founder, Lumor
Building Lumor for research teams. Previously led UX research at B2B SaaS companies. Writes about research methods, product design, and building in public.