How to Analyze Concept Interviews When People Say “Sounds Good”

“Sounds good” is a response worth examining

Concept interviews help you understand how people interpret a proposed solution, where it fits their experience, and what might prevent adoption. Praise or stated intent alone cannot establish purchasing, continued use, or market size.

“That looks useful. I'd probably use it.” It is an encouraging thing to hear. When reviewing the transcript, though, I would pause: did the participant understand the same feature we intend to build? Have they recently faced the problem? What prompted that answer? Tone alone cannot settle those questions.

NN/g's research methods guide distinguishes attitudes from behavior. What people say and what they do answer different questions. Attitude data can help; the finding needs to respect its limits.

We will use an expense-submission assistant as an example. All participants, quotations and events below are fictional teaching material, not ScribeFlash customer research or product capabilities.

Keep a record of what people actually saw

Suppose the concept card says: “Upload receipt photos to organize the documents for your expense submission.” Alex reads it and says, “Great, so I won't have to ask my manager to approve expenses anymore.”

The team intends to organize documents. Alex expects it to remove approval. Writing “Users welcome automated expenses” would hide that mismatch before the discussion has even started.

For each session, record the concept version, presentation order, introductory wording, any explanation you added, and the participant's initial description. If you revise the card midway through research, mark the version. Do not count responses to different presentations as if everyone saw the same thing.

Ask, “In your understanding, which part of the process would this handle?” Listen before explaining the intended scope. Keep answers from before and after clarification separate. The misunderstanding is useful evidence in its own right.

Follow “Would you use it?” with a specific experience

If Alex says expense reporting is frustrating, ask about the most recent submission. What happened between collecting the receipts and sending them? Perhaps a taxi receipt was missing on Friday afternoon, and Alex had to search messages for the attachment at home.

That gives you a situation to investigate. It is still an account of past events. Unless you saw the process or relevant records, do not describe it as directly observed behavior.

Useful follow-ups include:

• “Which part made you say you liked it?” Find out what the praise refers to.
• “How did you get the missing document that time?” Explore the existing workaround.
• “Where would this fit into that particular submission? What would stay the same?” Locate a possible use.
• “What would you need to know before trying it with your own material?” Explore adoption conditions.

Avoid “Automatic sorting would save you lots of time, wouldn't it?” GOV.UK's interview guidance recommends neutral questions about concrete experiences. An answer that rejects your proposed benefit is useful too.

Keep different kinds of material separate

This table is a manual analysis aid. These categories are not scores automatically assigned by ScribeFlash.

| Fictional material | A reasonable record | What it does not establish | | --- | --- | --- | | “Once I upload photos, approval is taken care of.” | A description of the concept that conflicts with its intended scope | Correct understanding or acceptance | | “I had to find a missing taxi receipt last week.” | A reported recent experience | Direct observation of that expense process | | “I'd probably use it when it launches.” | Stated intention | Registration, payment or repeat use | | Cannot find the missing-receipt option in a prototype task | Observed behavior under the recorded test conditions | Completion rates in the released product | | Books Tuesday's trial and actually attends | One completed follow-up action | Willingness to purchase or long-term retention |

A sentence can contain several kinds of information. Preserve its context rather than cutting it into neat but misleading fragments. Keep the participant code, interview source and concept version beside the quote. Add a recording timestamp only when one exists.

Read the less enthusiastic responses as carefully. Another fictional participant, Sam, says their company already recognizes receipts: the difficulty is knowing which expenses are eligible. That points toward policy interpretation. It should not disappear inside a broad theme such as “People want easier expenses.”

Give the next round something concrete to observe

A finding from our fictional example could read:

Alex interpreted document preparation as removing approval, while Sam already had receipt recognition. We cannot yet judge whether automatic organization is worth adopting. Revise the concept wording, then investigate the actual work of gathering missing documents and checking expense rules separately.

That gives the team a task. To check comprehension, show the revised card to new participants and ask them to explain it. To check task completion, prepare an interactive prototype. To investigate payment, present clear pricing, delivery scope and trial terms, then observe actual choices.

Leaving an email address, booking a trial, attending and paying each mean something different. Record them separately. Ask about missing actions too: unavailable documents, a scheduling conflict or lack of purchasing authority are different explanations from a lack of need. By the end of analysis, identify the evidence you most need next.

Use the concept feedback template in ScribeFlash

Organize interviews and prototype sessions separately. ScribeFlash helps you work with imported material and findings; your team still decides whether to build the proposed solution.

1. Open the concept feedback template and sign in when prompted. The project form pre-fills a name, objective and four research questions about interpretation, recent situations, concerns and contradictory responses.
2. Name the project precisely, for example “Expense assistant V1 concept interviews.” Edit the objective and questions around the decision you need to make, then save. These are analysis questions; keep a separate discussion guide for live follow-ups.
3. Import each interview into the project. Use text import for an existing transcript. Remove unnecessary identifying information and label participants consistently. Keep your concept versions and presentation conditions in your research records as well.
4. Review passages where transcription mistakes would change the meaning, especially negatives and conditions attached to an intention. When creating evidence from the transcript, retain the context and link it to the relevant research question.
5. After generating or organizing findings, open their evidence and check the claims. Look for assumptions presented as facts and missing contradictory material. Edit the scope before a researcher confirms the finding.
6. Include confirmed findings in the report, with unresolved questions and the next research task. The research report guide gives more detail. Generated prose does not establish market demand; the report still needs a defensible basis for its judgments.

If you have not started interviewing, review the research templates and adapt the questions to your project. A specific decision gives the conversation, and its analysis, a useful focus.

Sources

• NN/g: When to Use Which User-Experience Research Methods, for the distinction between attitude and behavior and the scope of research methods.
• GOV.UK: Using in-depth interviews, for neutral questions and discussion of concrete experiences.

The expense example, analysis table and proposed follow-up tasks are original teaching examples, not reported research results.