Customer research workshop / 4 MIN READ
Keep what happened separate from what it means
Build a short evidence trail between a customer observation and the marketing change you recommend.
Original editorial learning framework
Editorial and source review: 6 November 2026
A confident summary can turn a team assumption into a supposed customer fact. Separate the recorded behavior, your interpretation and the proposed action so others can challenge each step.
What you will leave with: A three-column evidence ledger that makes the reasoning inspectable.
Use this when
- Research notes contain conclusions without examples.
- A memorable interview quote is driving a broad redesign.
This is not a method for estimating how common a behavior is in the wider market.
Start with a traceable observation
GOV.UK research guidance distinguishes extracting observations, determining findings and deciding actions. That distinction supports the starting principle here. The three-column ledger below is an editorial adaptation for campaign and customer-experience teams.
Write something another observer could check: “The participant reopened the delivery page twice before choosing a date.” Do not write “The participant distrusted our brand” unless that is explicitly their own statement, and even then label it as reported belief. Attach a session identifier and location in the notes, not an unnecessary name.
List competing explanations
Beside the observation, write the explanation you currently favor and at least one credible alternative. Reopening a page could reflect unclear information, comparison with another product, distraction or a desire to confirm details. The point is not to make every explanation equally likely; it is to avoid hiding the inferential jump.
Look for other evidence that separates the explanations. Did the participant verbalize a question? Did others stop at the same place? Did the interface display a contradiction? Keep these observations in their own rows so one vivid story does not become an untraceable pile of supporting claims.
Make the proposed action proportionate
Choose a next action that fits the evidence. A low-cost wording clarification may be sensible after a clear misunderstanding. A complete positioning change requires more than one confusing session. Distinguish fixing an observable defect from betting on an untested psychological explanation.
Name an owner and a check. “Improve trust” is not an action. “Show the delivery date beside the total, then check whether people can identify it without help” gives the team something to build and inspect. Do not promise a conversion increase that the research did not measure.
Preserve disagreement during synthesis
Ask reviewers to challenge the weakest link in the chain: observation accuracy, interpretation or proposed remedy. Record unresolved disagreement instead of averaging it into a vague consensus. A finding can be useful and still uncertain. Include counterexamples that would limit where the conclusion applies.
When sharing the recommendation, show one compact ledger excerpt and explain the sample context. Remove personal details and respect participant consent. Keep the full evidence in an appropriately restricted workspace. The presentation should make the reasoning easier to inspect without distributing raw recordings unnecessarily.
Make the reasoning concrete.
A fictional course provider observes a participant repeatedly opening the syllabus while comparing plans. A stakeholder labels this “price anxiety.”
Diagnosis
The notes show the participant checking whether live feedback is included. Price anxiety is only one possible interpretation; missing plan information is another and has direct supporting evidence.
Add a clear feedback-inclusion comparison and test whether participants can distinguish the plans. Do not reprice the service based solely on the earlier label.
Take the questions into your next brief.
Use these prompts in your own notes. Nothing is submitted from this page.
Observation
What exactly was seen or heard, and where is its record?
Interpretation
What explanation fits, and what alternative also fits?
Counterexample
What evidence would narrow or overturn the conclusion?
Action
What proportionate change or further check follows?
Access
Who needs the evidence, and what participant details can be removed?
Watch for the wrong conclusion.
- Writing inferred emotions as observed facts.
- Selecting only quotes that support the preferred strategy.
- Making an expensive change without checking alternative explanations.
Your next action
Take one recommendation from a recent research deck and rebuild its observation-to-action chain. Flag the first unsupported step.
Sources & scope
The teaching, hypothetical exercises and recommended workflows are original editorial analysis. Linked sources support only the scoped facts identified beside them; they do not validate predicted results or endorse this framework.
- GOV.UK Service Manual: Analyse a research session →Primary public-service research guidance · Supports: Separating observations, findings and actions in research analysis; the evidence ledger and marketing example are original editorial tools.