AskCraftReturn to the original decision
Interesting quotes can take over after interviewing. Put the client’s decision at the top of the document: what are they choosing and what can they change? Material about choosing a one-off consultation may not answer which social platform to develop. Keep useful side questions separately so they do not displace the original question.
Inventory the material: available interviews, permitted storage locations, participant eligibility and missing information. Assign neutral participant IDs. Shared analysis usually needs identifiers rather than names or contact details.
Use one row for one episode or observation
Do not place an entire interview in one cell. Separate episodes that concern different situations. One person may choose a short recipe on a weekday evening and a detailed technique explanation at the weekend. Collapsing these into “likes recipes” loses the useful distinction.
A row needs an ID, source reference or timestamp, situation and exact passage. Put your observation in a separate column. Another researcher should be able to distinguish the original material from your interpretation of it.
Separate observation from explanation
Fictional passage: “I noticed the hour of marinating after I had taken the ingredients out.” Observation: the participant began preparation and then discovered additional waiting time. Explanation: total time to the meal mattered to their choice. The explanation remains a hypothesis, even when it seems convincing.
A useful check is whether someone can find the observation in the source without extra reasoning. If not, you may already be interpreting it. Interpretation is necessary; the aim is to make its place in the reasoning visible.
Compare situations, not just repeated words
Two people can say “I need it quickly” and mean different things. One has a meal deadline; another wants to understand a technique quickly and is happy to cook slowly. Compare circumstances, desired progress, alternatives and selection criteria. Words help locate passages but cannot replace context.
Keep contrasting accounts alongside one another. If one person avoids long content and another seeks it, do not select only the convenient quote. Investigate the circumstances. The implication may be a choice of formats rather than shortening every article.
Consider an alternative explanation
“The timing was unclear” competes with “the person was already tired” and “the ingredients were unavailable anyway”. Note what supports each explanation and what is missing. This helps select the next question and prevents a recommendation built around one memorable account.
Avoid assigning arbitrary confidence percentages. Be specific instead: two relevant accounts support the idea; one did not explore the cause; a contrasting account involved different circumstances. This is more useful than a number with no defensible basis.
Turn a hypothesis into a testable change
“Make the content more useful” does not define an action. A concrete proposal adds total time, waiting time and substitutions to a recipe card. Its basis is accounts where waiting or a missing ingredient affected choice. Its limitation is that presentation cannot remove tiredness or guarantee cooking.
Specify the prototype, situation, observable action, owner and a result that would make you reconsider. A person might understand the timing yet reject the recipe because it looks difficult. That would indicate the problem extends beyond the timing label.
Discuss priorities and preserve limitations
Prioritise a change tied to an important decision, supported by understandable evidence and feasible now. Do not automatically choose the most repeated word. Mention counts in a small qualitative sample do not establish prevalence across the whole audience.
Bring a concise summary and access to the underlying material. Record what the team will test, what it has deferred and what evidence is needed next. A useful handover makes the decision, evidence, limitations and next step clear.
How to use the workbook
The template contains a brief, evidence, interpretations and tests. Start with the brief, then add a few passages you are allowed to use. Link interpretations to evidence IDs and tests to hypothesis IDs. Blank fields are for your project; instructional prompts are not research data.
Before handover, ask someone else to trace one test back to its source passage. If the connection is unclear, add context or narrow the claim. State when the material is insufficient. A workbook supports discussion of the reasoning; it does not perform the research for you.
Apply this to your own question
A blank workbook with instructions: brief, evidence, interpretations and tests. XLSX, four worksheets.