Marcus Andrey Study
Ler em português Library

UX Research · Labs · 3 min read

Write a research plan

A guided lab: pick a real decision your team faces, and write a one-page research plan for it, from the research questions to the participants, the method, consent and how you'll share the results. Then review it with one stakeholder. About an hour and a half, with any document tool.

A plan is where you find out the study is the wrong study, while it still costs nothing to change. You'll write a one-page plan for a real decision, then hand it to the person who'll use the results. If they can't say what they'll do differently after the study, you've just saved a month.

What you'll do:

  • Name the decision the research will inform
  • Write and rank three to five research questions
  • Choose a method, the participants and a schedule
  • Cover consent, data and how you'll share the results
  • Review the plan with one stakeholder

You'll need: a document tool or paper, a real product question from your work or a side project, one colleague for 15 minutes, about an hour and a half.


1. Start from the decision

Write one sentence:

After this study, ______ will decide ______, by ______.

For example: "After this study, the product team will decide whether to build group-chat sharing for splits, by the planning meeting on 12 March." If you can't name who decides what, go back to Research questions that lead to decisions. NN/g's first rule applies: "All UX-research studies should start with a plan" (Maria Rosala, NN/g, 2024), and a plan starts with why.


2. Write the research questions, then rank them

Write three to five questions the study must answer. They are what you want to learn, not what you'll ask participants (that's the guide, in the next lab). Then rank them: the GOV.UK Service Manual suggests agreeing and prioritising research questions with the team, so each round knows what it's for (GOV.UK Service Manual, Plan user research for your service).

Good research questions are open, about behaviour, and answerable by this study. "Do people want group-chat sharing?" fails all three. "How do households tell each other about a new bill today?" passes.


3. Fill in the plan

Research plan: how households split bills

Owner: product designer, Splits teamVersion: 2, 18 FebReviewed by: product manager, 19 Feb

Decision

The product team decides whether to build group-chat sharing for splits, at planning on 12 March.

Background

Half of new splits are never marked as paid. Analytics show where people stop, not why.

Research questionsranked

  1. How do households tell each other about a new bill today?
  2. What makes someone question a split?
  3. Who keeps track of who owes what, and what does it cost them?
  4. Where does our app fit in that, and where does it get in the way?

Method

8 remote interviews, 45 minutes, with a show-me of the chat or notes people use. Interviews, because RQ1–3 are about past behaviour and reasons, not about our screens.

Participants

  • 8 people from 8 different homes: 5 flatshares, 3 couples
  • Split at least one bill in the last month; at least 2 not current users
  • Recruited from our panel and a recruitment agency; screener attached
  • Incentive: €40 voucher

Teamwho does what

Designer moderates. Engineer and PM each take notes in 2 sessions. Everyone joins the analysis on 6 March.

Consent and data

Consent form v3 (recording, AI transcription named). Recordings in the research drive, team access only, deleted 6 months after the study. Notes use codes P1–P8, no names.

Schedule

Recruit 20–27 Feb · Pilot 28 Feb · Sessions 2–5 March · Analysis 6 March · Top-line 7 March · Readout 10 March

Sharing

Top-line in the team channel; 10-minute readout at the sprint review; deck and nuggets in the repository.

Risks

Couples are hard to recruit: if fewer than 3 by 25 Feb, ask the panel again with a higher incentive.

A research plan on one page, filled in. It starts from a decision with a date, and every later line serves it: the readout lands two days before planning.Example study, invented for this course. Plan: Marcus Andrey, CC BY 4.0. Sections after Maria Rosala, Research Plans (NN/g, 2024) and the GOV.UK Service Manual.

Use the worked example as a model, not a form to copy. NN/g's research plan has four parts: the purpose and goals, the participants, the method and procedure, and the related documents such as the consent form and screener (NN/g, 2024). The example adds three lines that save trouble later: who on the team does what, where the data lives, and when the results will be shared.

For the method, use Choosing a research method. For the participants and how to find them, Recruiting participants and Screeners and sampling. For consent and storage, Research ethics and consent and Research data and privacy.


4. Check the schedule against the decision

Work backwards from the decision date. Recruiting usually takes longest; analysis takes longer than people expect (GOV.UK suggests about one hour of analysis for every two hours of research, GOV.UK Service Manual, Analyse a research session). If the results can't land before the decision, change the scope now, not later: fewer sessions, a narrower question, or a top-line on a fixed date (see Research reports people read).


5. Review it with one stakeholder

Send the plan to the person who'll use the results, and ask three questions in 15 minutes:

  1. If we learn what we hope to, what will you do differently?
  2. Is there a question missing that you'll ask me after the study?
  3. Who else should watch the sessions?

Change the plan. Then add one line at the top: "Reviewed by ______ on ______."


6. Check your work

Reflection questions: Which research question did the stakeholder's answers change? What would you cut first if you had half the time?

Writing a research plan · Research questions that lead to decisions · Choosing a research method · Run three interviews

Saved only on this device.

↑ ↓ to move · Enter to open · Esc to close