AskCraft1. Material: what the person actually said
“The article said twenty minutes, but then it needed another hour to marinate.”
This illustrative quote concerns one particular dinner. A working report should include the participant ID, a reference to the source and the context. Without those, the team cannot inspect what came before the sentence.
2. Observation: what happened
The participant selected a recipe using the stated time, discovered additional waiting and ordered food. Preserve that sequence. “People dislike long recipes” is a generalisation that one account does not support.
3. Interpretation: our explanation
One possibility is that the person needed the total time until dinner before committing. Another is that they had no energy to cook that evening. Compare other accounts and investigate whether the same difficulty occurred under different conditions.
4. Decision: what to test
A prototype recipe card could separate active preparation, waiting and time until the meal is ready. Give someone a concrete task: “You want to eat in forty minutes. Which option would work?” This checks understanding rather than visual preference.
A table that preserves the chain
Use five columns: source; observation; interpretation; alternative explanation; next test. Mark unsupported ideas as team assumptions. Place conflicting accounts alongside the finding instead of hiding them in an appendix.
You may find that people avoid elaborate cooking on weekdays but deliberately seek detailed explanations on weekends. That is not a defect in the data. Different circumstances may call for different content.
Discuss the decision with the client
Return to the decision that prompted the research. Show important evidence and limits, then agree which hypothesis to test first, who will make the prototype and what observation would change your mind. A collection of quotations does not do this on its own.
Qualitative material helps build and refine explanations. Population prevalence and the effect of a change require separate checks.