User research analysis is the process of interpreting participant evidence to answer a research question. Start with the decision the team faces, choose an appropriate method, and keep each finding connected to the material that supports it. Separate what participants said, what you infer, and what the team should test next.
Use the editable research evidence worksheet to document that chain. It is a Markdown text file with fields for source location, code, interpretation, counterevidence, coverage, and action. The worked example below uses the published shopper study so readers can inspect the audio, transcript, and sample presentation rather than relying on an invented participant quote.
What Analysis Methods Produce the Richest Insights?
Four analysis methods cover the majority of user research analytical needs. Each has specific strengths, and the choice should match the research question and the type of insight needed.
Thematic analysis develops patterns of meaning across a dataset. Coding may begin from the data, an existing research framework, or a combination; it is not necessarily purely inductive. Different thematic approaches make different assumptions about researcher interpretation and coding consistency. State which approach you use and why it fits the question.
Braun and Clarke describe six phases for reflexive thematic analysis: familiarization, coding, developing initial themes, reviewing and developing themes, refining and naming themes, and writing. Their process is iterative rather than a checklist completed once. Their overview of thematic approaches also explains why coding reliability is not a universal quality requirement across all forms of thematic analysis.
Affinity mapping organizes data spatially to reveal relationships between findings. Each distinct insight from the transcripts is captured as a discrete unit (typically a sticky note or digital equivalent), then clustered by similarity. The clustering process itself generates analytical insight because it forces the researcher to articulate why certain findings belong together. Affinity mapping works particularly well for discovery research where the landscape of findings is complex and the relationships between insights are not immediately obvious.
Framework analysis organizes cases in a matrix against explicit dimensions, such as journey stage, task, constraint, and outcome. It is useful when a team needs to compare the same questions across participants or segments. Start with dimensions relevant to the brief, but allow unexpected evidence to change the framework rather than forcing every account into the original columns.
Narrative analysis examines how a participant describes an experience over time. Preserve the sequence, turning points, and context rather than fragmenting every account into isolated codes. This is useful for understanding how someone explains a purchase, cancellation, or difficult task. A retrospective account can suggest a mechanism, but it does not establish causation on its own.
How Does AI Change Analysis at Scale?
For larger studies, tools such as User Intuition for user researchers can assist with organizing transcripts and reviewing evidence. The right balance of software and manual analysis depends on the method, participant context, and consequences of the decision. A larger study needs an explicit review process, not an assumption that automation resolves interpretation.
Where AI assistance can help. Retrieval and initial grouping can make a large dataset easier to navigate. Use suggested themes as candidates to investigate. Search deliberately for less common accounts and negative cases, and check the underlying passages: neither a plausible summary nor a frequency count proves the grouping is meaningful.
Evidence links let a researcher check the passage behind a finding. On User Intuition, inspect the linked source and the surrounding conversation before presenting the interpretation. Verify that an excerpt is participant speech, that the moderator did not supply its premise, and that a transcript error has not changed the meaning. Listen to the recording where wording or context is uncertain.
Where human interpretation remains essential. AI analysis identifies patterns but cannot evaluate their significance. A theme that appears in 30% of interviews might be critically important or entirely trivial depending on organizational context that AI does not possess. The researcher determines which findings matter — which patterns represent actionable opportunities, which confirm existing knowledge, and which challenge assumptions in ways that should change product direction.
Interpretive nuance requires attention to context. A transcript can omit tone, interruptions, and uncertainty, and a human reader cannot recover those signals merely by reading harder. Return to the audio where relevant, avoid guessing an emotional state, and mark ambiguity when the evidence does not resolve it.
The optimal human-AI workflow. The most effective analytical workflow uses AI for initial theme identification and evidence linking, followed by human review for theme evaluation, significance assessment, and interpretive enrichment. Researchers spend analytical time on judgment rather than coding — evaluating whether AI-identified themes are meaningful, connecting themes to organizational strategy, identifying implications that require domain knowledge, and crafting narratives that communicate findings persuasively.
Estimate analytical effort from a pilot rather than applying a universal percentage saving. Log time spent checking source material, revising codes, investigating contradictions, and preparing the decision narrative. Automation is useful when it reduces mechanical work while leaving enough researcher time to evaluate the evidence.
How Should Analysis Outputs Be Structured for Maximum Impact?
The structure of analytical outputs determines whether findings are actionable or academic. Research teams that struggle with stakeholder engagement often have an output structure problem rather than a finding quality problem.
Theme hierarchy. Organize themes in a hierarchy: 3-5 top-level themes that capture the major patterns, each supported by 2-4 sub-themes that provide specificity. This structure helps stakeholders grasp the landscape quickly (top-level themes) before diving into detail (sub-themes). Flat lists of 15-20 themes are overwhelming and obscure the relative importance of different findings.
Evidence quality. Select excerpts that support the specific interpretation, retain their source IDs, and include variation or disagreement. A vivid quote can illustrate a finding but cannot establish how widespread it is. Avoid filling a quote quota with several excerpts from the same person or quoting a moderator as though they were a participant.
Counts and coverage. When counts are appropriate to the method, report participants rather than mentions and state the eligible denominator. For example, an illustrative “6 of 12 participants asked about setup described difficulty” is different from “half of all users struggle.” Record not-asked, unclear, and no-experience cases separately from explicit negative answers. Qualitative recruitment and a large sample do not by themselves support population estimates.
Actionability framing. For each finding, state what it suggests, what remains uncertain, and which decision it can support now. Separate the observation from the recommendation. A report of setup difficulty might justify a usability test; it does not by itself identify the best redesign or its expected commercial return.
Worked example: from W30’s account to a bounded finding
In the sample interviews, select W30 — Phone replacement, unplanned and urgent. The participant describes an urgent replacement purchase and help finding the phone section. This is published study material; the coding below is a teaching example of an analyst’s interpretation, not a claim about all shoppers.
“I could access phones in other store but Walmart was nearest to me.”
That excerpt is reproduced from W30’s transcript. Read the surrounding exchange and listen to the audio before interpreting it. The participant’s later comments about helpful staff concern the experience after entering the store, which should remain separate from the initial choice of retailer.
| Analytical step | Example entry | Review question |
|---|---|---|
| Source | W30, answer to why this store rather than another store or online | Can another researcher locate the full exchange? |
| Initial code | Nearest store for an urgent replacement | Does the code preserve the event and stated reason? |
| Provisional case finding | In this account, proximity helped the shopper act quickly after the old phone broke | Have we distinguished a single case from a cross-case pattern? |
| Alternative explanation | Habit, prior expectations of service, or product availability may also have mattered | Which alternatives were explored, and which remain unanswered? |
| Evidence gap | No complete reconstruction of other sellers considered before departure | Are we treating missing evidence as a negative answer? |
| Next action | Probe the consideration sequence in later interviews and compare urgent with planned trips | What additional evidence would change the interpretation? |
Do not promote this to “Walmart wins on convenience.” That claim generalizes beyond the account. Nor does a positive service experience prove that service caused the store choice. The difference between those statements is the difference between tracing evidence and attaching a quote to a preferred story.
To build a theme, examine other relevant interviews, compare contradictory cases, and refine what the pattern means. W10, the grocery top-up interview in the same preview, provides a different trip context to examine; do not assume an urgent phone replacement explains routine grocery behavior. The two public example calls are useful for learning the method, but they are not the whole study dataset.
Turn the evidence matrix into a client or product readout
Use one row per proposed finding in the downloadable worksheet. Record the research question, participant IDs, source locations, analyst interpretation, and exceptions before writing the recommendation. If you report counts, state who was eligible to answer and whether the question was consistently asked. This makes a claim reviewable even when the final presentation contains only a short summary.
The sample presentation contains a method slide and a final slide on what the 43-shopper base cannot establish. Use that structure when handing findings to a client or product team: explain the question, show the evidence, state its limits, and name the next decision. The presentation is an example of a research deliverable, not evidence that your own audience will behave the same way.
Before signing off, have a colleague trace a consequential claim back to its sources without your explanation. If they cannot tell whether a sentence is participant speech, interpretation, or a proposed action, revise the wording. Keep unresolved questions visible rather than making the report appear more certain than the study allows.
How Do You Choose the Right Analysis Method for Each Study?
Choose the method around the question and intended use. For an exploratory brief, thematic analysis can help develop patterns you did not anticipate. For a consistent comparison across cases, use a framework matrix and keep room for exceptions. Use affinity mapping when the team needs to develop a shared understanding of observations, and narrative analysis when the sequence of events matters. These approaches can complement one another, but explain which work each method is doing rather than treating them as interchangeable labels.
Before selecting software or sample size, test the analytical plan on a few interviews. Define the unit of analysis, source references, treatment of missing answers, and review process. Decide which comparisons would be meaningful with the recruited audience. Expand the study when the additional interviews can answer a remaining question, not simply to produce a larger number beside a theme.
For the operational workflow, see User Intuition for user researchers and the Customer Intelligence Hub. User Intuition voice interviews cost $30 per quality interview with your sample. Standard panel recruitment adds $30 per participant; specialty audiences are quoted separately, and your own incentives are additional. Keep researcher review and client delivery in the project budget. The platform supports the evidence workflow; your team remains responsible for what the findings justify.