Build a customer research plan by starting with the decision the research must inform. Record what is known, what remains uncertain and what evidence would change the decision. Then define suitable participants, method, recruitment, consent, discussion topics, analysis and a stopping rule. This prevents interviews becoming interesting conversations with no clear use.
Start with the decision, not the method
“We should speak to customers” is an intention, not a research objective. Name the decision owner, the choice they face and the date by which evidence is useful.
A positioning project might need to choose a priority group or competitive frame. Write the decision as a sentence: “By 20 November, the product lead will choose whether the onboarding offer should focus on agencies or in-house teams.”
Research reduces uncertainty; strategy and delivery constraints still matter.
Turn assumptions into research questions
Separate existing evidence from team belief, then convert the riskiest assumptions into answerable questions.
For example:
- Assumption: agencies experience the problem more frequently.
- Evidence held: support notes from three agency accounts.
- Gap: no comparison with in-house teams.
- Research question: in what situations does the problem occur, and how do the two groups respond?
Avoid objectives such as “validate our idea”. Ask what would support, weaken or complicate the current view.
Define participants by relevant experience
Recruit people who can speak from recent, direct experience. Role titles alone are weak criteria because people with the same title may have different authority and workflows.
Specify:
- the event or decision participants must have experienced;
- how recently it occurred;
- organisation or use-case characteristics that matter;
- customer status, such as new, active, former or lost prospect;
- exclusions that could distort the sample; and
- important comparison groups.
The participant recruitment guide explains how to turn these criteria into a screener and outreach process.
Choose a method that fits the question
Interviews reconstruct decisions and context. Observation is stronger when you need to see behaviour. Surveys can estimate defined responses in an appropriate sample, but poorly framed surveys scale ambiguity.
Use the interviews-versus-surveys comparison to choose deliberately. Qualitative work can identify variables for a later survey.
Do not add methods for the appearance of rigour. Every method should answer a named question.
Plan recruitment, consent and data care
Write the invitation, screener, incentive and scheduling process before outreach. Explain what participation involves and how the information will be used.
Decide whether sessions will be recorded, who can access recordings and how long files will be retained. Collect only necessary information; sensitive data may create additional legal and ethical duties.
Consent is a continuing agreement, not a sentence hidden in a calendar invitation. Ask separately before recording and before attributing any quotation.
Build a guide around events and evidence
Create discussion topics rather than a rigid script. Ask participants to reconstruct a recent experience: what happened, what they tried, who became involved and what changed.
The customer interview question guide offers prompts, but choose only those relevant to your decision. Avoid leading questions, speculative feature voting and requests for polished marketing language. Ask “What happened next?” and “Can you show or describe a specific example?”
Pilot the guide and revise questions that produce confusion or generalities.
Decide how evidence will be analysed
Plan analysis before the first session so the team captures comparable information. Create fields for participant context, observed event, alternatives, selection criteria, desired outcome, friction, supporting extract and researcher interpretation.
Keep evidence distinguishable from conclusions. “Participant delayed approval for six weeks” is evidence; “enterprise customers fear change” is an interpretation. Record contrary examples.
Store evidence in a usable customer research repository with access and retention controls.
Set a sample strategy and stopping rule
Plan enough relevant participants to compare the groups that affect the decision, then review evidence quality as work progresses.
A stopping rule might require represented comparison groups, concrete examples and two successive interviews that add no material category. This is not population measurement.
Use the guide to planning interview numbers when setting the first recruitment batch.
Use a one-page planning checklist
Before booking an interview, confirm that the plan names:
- the decision and owner;
- the deadline and constraints;
- current evidence and risky assumptions;
- neutral research questions;
- participant criteria and comparison groups;
- method and recruitment route;
- consent, recording, access and retention arrangements;
- discussion topics and pilot;
- analysis fields and synthesis owner; and
- a stopping rule and decision workshop.
Illustrative example
Imagine a small software company deciding whether to reposition around faster client approvals. Its plan compares recent customers who adopted the product with prospects who retained email-based approvals. Interviews reconstruct the last approval problem, options considered and people involved; a short workflow observation checks what participants actually do.
The team will not claim market prevalence from this small purposive sample. It will use the evidence to refine the proposition and identify questions for a later survey. This is an illustrative scenario, not a reported study.
Account for limitations and trade-offs
A narrow sample produces relevant depth but can miss adjacent situations. A broad sample may surface variety while making comparison difficult. Incentives support participation but do not remove self-selection bias. Interviews reveal remembered accounts, not an objective recording of events.
Product-team research improves exposure to evidence, but prior beliefs can influence interpretation. Neutral facilitation, structured notes and counter-evidence reduce those risks.
Frequently asked questions
How long should a customer research plan be?
One or two pages is often enough if it makes the decision, participants, method, ethics, analysis and stopping rule explicit. Higher-risk work may require more.
Can sales calls count as customer research?
They can contribute evidence, but a sales call has a commercial purpose. Preserve context and do not treat it as a replacement for sampled research.
Who should observe interviews?
Decision-makers benefit from direct exposure, provided the participant agrees and the number of observers does not create pressure. Give observers a note-taking role and keep them from turning the session into a pitch.
Your useful next step
Write the pending decision at the top of a blank page. Under it, list the three assumptions most likely to reverse that decision. Convert the riskiest assumption into one neutral research question and define the recent experience a participant must have to answer it.
