Reducing an audience to an age range and a few interests makes it difficult to explain why content should work. I want research to produce decisions, not decorative character profiles. What is someone trying to do? Where do they struggle? What would they change about their current solution?
Limit the research question
“What do people think of our brand?” is broad. “Which information is missing when a first-time customer evaluates a product page?” is more actionable. A narrower question clarifies recruitment, the interview flow and where findings will be used.
Do not treat first-time customers, repeat buyers and people who left without buying as an interchangeable group. If you cannot reach the last group, record that limitation. Research based only on existing customers cannot represent the entire potential market. A small, clearly described sample is more useful than a broad claim with unclear origins.
Ask about past behaviour
“Would you like this content?” can invite a polite yes. I prefer the last real experience: “Where did you start looking? Which alternatives did you compare? What did you check before deciding?” Preserve the participant’s wording rather than translating it immediately into marketing vocabulary.
Follow an example instead of mechanically reading a questionnaire. If someone says a page did not feel trustworthy, ask which detail or omission created that impression. Record information relevant to the decision without collecting unnecessary personal details. The objective is to understand a choice, not build a dossier about the participant.
Separate observation, interpretation and action
| Observation | Possible interpretation | Change to test |
|---|---|---|
| A participant could not find dimensions | Dimensions may lack visibility | Move them into the first information group |
| Three product tabs stayed open | Comparison may be difficult | Use a consistent specification order |
| Care instructions were requested from support | The page may omit an answer | Add the care question beside the product |
This is an illustrative worksheet. Keeping these layers separate prevents a colleague from inheriting an assumption as a fact. An observation can support more than one interpretation, so a proposed design change should remain a test until it is evaluated.
Turn an insight into a content decision
“People want information” is too general. “First-time buyers cannot relate the dimensions to their own space” points to a particular problem. Content should then show scale and use, rather than simply add more brand storytelling.
Choose one change for each insight. Test a scale comparison before changing the entire description structure. If everything changes at once, it becomes harder to understand which intervention helped. Agree in advance what evidence would make the team keep, revise or abandon the change.
Keep the research current
A new category, price level or distribution channel can invalidate old assumptions. Repeated new support questions are another reason to revisit research. A single comment, however, may not justify changing direction.
I close the research record with a date, participant context, sample limitations, the decision and a review point. That preserves why the team acted, not merely a presentation of what it found. Real decision situations provide a stronger foundation than invented persona details that sound vivid but cannot guide the work.