Customer interview notes should preserve what the participant said, what the researcher observed, what remains uncertain and what decision the team may change. Notes that immediately convert every comment into a product requirement erase the boundary between evidence and interpretation.

This guide introduces TRACE-8, a reusable interview note and decision-follow-up template for market, product and customer research.

Direct answer

Record eight fields for every interview: context, question, verbatim evidence, observation, interpretation, alternative explanation, decision relevance and follow-up. Keep source evidence separate from analysis, label uncertainty and never invent a quote from memory.

TRACE-8 interview note template

Field What to record Quality check
Target context Participant segment, use situation and relevant background Is the context necessary and authorized?
Research question The question this evidence may inform Is it neutral and decision-linked?
Actual evidence Short verbatim quote or faithful paraphrase Can it be traced to the recording or contemporaneous note?
Contextual observation Behaviour, sequence or artifact observed Is observation separated from judgment?
Explanation Provisional meaning assigned by the researcher Is it labelled as interpretation?
Counter-explanation Another plausible reason for the same evidence What would disconfirm the first interpretation?
Decision relevance Which decision, assumption or criterion may change Is the implication proportionate?
Evidence follow-up Probe, additional participant or other source needed Who owns it and by when?

Before the interview

Define the business decision and research question before recruiting. State the target population, selection logic, consent process, data-retention rule and which claims the sample can support. One useful interview can expose a hypothesis. It cannot estimate population prevalence.

Prepare an interview guide with neutral core questions and optional probes. Avoid questions that contain the preferred answer, combine two topics or ask participants to predict behaviour they have never performed.

During the interview

Use short timestamped notes. Mark direct quotations only when the wording is accurate. Otherwise use a paraphrase label. Record concrete episodes before opinions: what happened, in what order, with which tools, constraints and workarounds.

Do not place sensitive personal, customer or employer data into an unapproved transcription or AI service. Follow the consent, access and retention rules established for the study.

After the interview

Complete the interpretation, counter-explanation and decision fields while the session is fresh. Then preserve the original notes. Corrections should be additive and dated rather than silently replacing the source record.

Use a cross-interview evidence table:

Evidence ID Segment Pattern code Supports Contradicts Decision implication Confidence
I-01 New customer Onboarding delay Setup friction No price objection Test a guided setup Low: one interview
I-02 Existing customer Onboarding delay Setup friction Workaround succeeds Compare guided and self-service setup Moderate: repeated pattern

The example is fictional. Confidence should describe the evidence base, not the researcher's enthusiasm.

A practical coding rule

Use one code when the same underlying issue appears in different language. Keep separate codes when the decision consequence differs. For example, “I could not find the export button” and “the export control was hidden below the fold” may share a discoverability code. “The export failed after clicking” is a reliability issue and should remain distinct.

For every proposed theme, retain:

  1. supporting evidence from more than one source where possible;
  2. counter-evidence or a negative case;
  3. the segments represented and missing;
  4. the decision that could change; and
  5. the next evidence needed.

Worked decision follow-up

Suppose four of eight interviewees describe setup friction, but all four came from one channel and two used an outdated version. The correct conclusion is not “50% of customers struggle.” A defensible statement is: “Setup friction appeared in four interviews, concentrated in one recruitment channel; version and channel effects remain unresolved.”

The next step might be a usability task with the current version across two segments, followed by behavioural data. The interview evidence directs the next test without pretending to measure the market.

Researcher review checklist

  • Are quotes and paraphrases distinguishable?
  • Can each note be traced to an authorized source?
  • Are observation and interpretation in separate fields?
  • Is a plausible counter-explanation recorded?
  • Does the implication name a real decision?
  • Are missing segments and contradictory evidence visible?
  • Are personal and confidential data handled under the approved plan?
  • Does every follow-up have an owner and deadline?

Learning pathway

MTF Institute's Professional Certificate in Market Research covers research briefs, sampling, recruitment, interviews, qualitative coding, surveys, mixed-method analysis and decision-ready reporting. Learners produce an interview session and fieldwork record as part of a broader evidence workflow.

The programme is online professional education and awards a course-completion certificate. It does not authorize participant contact, replace local law or employer procedures, or guarantee research or career outcomes.

Sources