# Completed fictional interview analysis example

Source: fictional Northstar interview NS-07. No real participant or company is represented.

## Research question

What makes meeting and customer-interview follow-up difficult, and what must useful notes preserve?

## First-read memo

Maya wants a shorter current-state view without losing the discussion that changed it. She treats automatic output as a draft, checks sensitive details against the source, and distinguishes a requested feature from the problem behind it. Her initial statement about ten-minute follow-up is later narrowed to routine weekly meetings.

## Evidence table

| Source locator | Short passage | Code | Interpretation |
| --- | --- | --- | --- |
| 01:23-01:48 | Ten minutes for a weekly check-in; longer for customer or planning calls | effort depends on meeting type | The first time estimate needs its condition |
| 01:51-02:29 | Copy was approved, then legal reopened review | decision changed inside conversation | Current state and change history both matter |
| 02:52-03:20 | Check exact wording and ownership; write “owner to confirm” when unclear | verify sensitive details | A useful note can preserve uncertainty |
| 03:23-04:31 | Dashboard request pointed to cross-team status problem | request differs from underlying need | Keep the request as evidence without treating it as the finding |
| 04:53-05:28 | Pull out useful material and keep its source nearby; first pass lacks internal context | evidence-linked editable draft | Generated notes still need review |
| 07:24-08:28 | One interview supports an observation, not a broad pattern | limit claims to evidence | Follow-up interviews must test the condition |
| 08:52-09:18 | “Launch Thursday” depended on legal approval by Tuesday | preserve condition | Short bullets can erase the part that governs action |
| 12:07-end | Correct the ten-minute answer so the condition stays attached | participant correction | The final note must replace the broad reading |

## Codes

- effort depends on meeting type
- decision changed inside conversation
- verify sensitive details
- owner remains uncertain
- feature request differs from need
- evidence-linked editable draft
- one account cannot establish a broad pattern
- preserve conditions and exceptions
- participant corrected earlier answer

## Candidate themes

### Useful compression keeps the condition attached

The participant wants a concise current state, but several examples show that removing sequence or conditions changes the meaning. Supporting codes: decision changed inside conversation; preserve conditions and exceptions; participant corrected earlier answer.

### Trust comes from review and traceability

The participant treats generated output as a draft and checks sensitive decisions, names, and ownership against the source. Supporting codes: verify sensitive details; owner remains uncertain; evidence-linked editable draft.

## Narrow findings from this interview

1. This participant needs notes to show a current decision and the earlier discussion that changed it. Evidence: 01:51-02:29 and 10:22-10:50.
2. This participant checks sensitive wording and ownership against the source rather than treating an automatic summary as the record. Evidence: 02:52-03:20 and 05:25-05:51.
3. Requested features are useful evidence, but this participant tries to record the underlying work problem separately. Evidence: 03:23-04:31.
4. The participant’s ten-minute follow-up estimate applies to routine weekly meetings, not customer or planning calls. Evidence: 01:23-01:48 and 12:03-end.

These are observations from one fictional interview. They do not establish frequency, saturation, or a population-wide pattern.

## Contradictions and corrections

- Earlier: follow-up takes ten minutes.
- Later correction: ten minutes applies to an easy weekly check-in; complicated calls take longer.
- Final note: keep the corrected, conditional statement and cite both passages.

## Follow-up questions

- When do other project leads consider a spoken proposal a confirmed decision?
- Which meeting types require checking audio as well as the transcript?
- Do other researchers separate a requested feature from the need beneath it?
- Who reviews generated notes before they become a team record?
- How should access differ between raw transcripts and shared findings?

This fictional example is provided by Vyxora Editorial under CC BY 4.0. Attribution: “Completed fictional interview analysis example by Vyxora, https://vyxora.app/blog/analyze-interview-transcripts”.