How to turn customer interviews into a positioning brief

A memorable sentence can dominate a workshop because it sounds true. That makes qualitative research vulnerable to anecdote bias. One articulate customer may describe the problem beautifully while representing a rare buying situation.

Preserve verbatim language, but place it beside context: who said it, what triggered the purchase, which alternatives were considered, what mattered during evaluation, and what happened after the decision. The strategic question is less about which quote sounds best and more about which pattern deserves investment.

Treat the interview corpus as a decision dataset. The output should explain what the business will prioritise because of the evidence, rather than simply summarising what participants said.

Write four columns during synthesis. Observation contains the evidence itself. Interpretation explains what the researcher thinks it means. Implication describes the possible consequence for the brand. Decision records what the business chooses to do.

This separation makes assumptions inspectable. A customer saying they hired the firm after a failed internal attempt is an observation. Interpreting that as a demand for specialist rescue is plausible. Choosing to position entirely around rescue work is a much larger decision and needs support from the wider evidence and business direction.

When these levels collapse into one another, strategy can acquire false certainty. Keeping them separate makes disagreement productive because the team can identify whether it disputes the evidence, the interpretation, or the commercial choice.

Begin with a small coding system that follows the decision rather than every topic discussed. Useful codes include trigger, previous workaround, alternative, selection criterion, risk, proof, objection, desired progress, outcome, language, and memory.

Add participant attributes only where they may explain meaningful differences: customer stage, service purchased, buying role, business size, geography, or relationship length. Avoid building dozens of demographic tags before the research shows they matter.

A spreadsheet is enough for a small study. Each row can represent one evidence fragment, with columns for participant, code, verbatim text, context, researcher note, and confidence. The aim is traceability rather than research theatre.

Create a matrix with candidate patterns on one axis and participants or segments on the other. Mark where the pattern appears, how specific the evidence is, and whether the participant's behaviour supports the stated opinion.

Frequency matters, but it is only one signal. A pattern mentioned by fewer customers may still be strategically important if it describes the highest value buying situation or exposes a distinctive capability competitors rarely provide.

Look for co occurrence. A trigger may repeatedly appear with one alternative and one proof requirement. Those combinations often reveal a stronger positioning opportunity than isolated themes such as trust or quality.

Contradictory interviews may reveal different buying situations rather than poor research. One group may value senior access while another values process speed. One may compare the business with agencies while another compares it with hiring internally.

Trace contradictions back to context. If different situations consistently produce different alternatives or criteria, the positioning problem may require segmentation, offer architecture, or a clearer priority rather than a sentence broad enough to satisfy everyone.

Record minority patterns that carry high commercial importance. Synthesis should reduce noise without deleting inconvenient evidence.

Write the brief as decisions supported by evidence. Begin with the priority audience and buying situation. Define the category or comparison frame customers already use. Name the tension that creates urgency. State the distinctive choice the business can repeatedly deliver. Attach proof. Then compress the system into a message another person can understand.

For each decision, include evidence references and a confidence rating. High confidence may mean several relevant participants described the same behaviour with concrete examples. Medium confidence may mean the pattern is plausible but uneven. Low confidence belongs in the validation backlog rather than the headline.

Generate a small number of candidate positions from the evidence, then score them against relevance, distinctiveness, credibility, commercial value, operational fit, and memory potential.

A candidate can be strongly supported by customer language and still be a poor strategic choice if the business has no desire to build around that demand. Research informs strategy; it never hands leadership decisions to the sample.

Document the tradeoff behind the chosen position. Naming what the business is deprioritising makes the decision more durable when new opportunities appear.

Before a full rollout, test whether the proposed position improves understanding in real contexts. Rewrite the homepage opening, one service description, one sales explanation, and one proposal introduction. Ask people to explain the business back after brief exposure.

Return to customers where appropriate with specific comprehension questions rather than asking whether they like the positioning. Observe whether the category, relevance, difference, and proof are understood without extensive explanation.

Keep the research archive connected with the final brief. Future revisions should be able to trace a strategic claim back to evidence rather than reconstructing the rationale from memory.