user-research-synthesis

Use thematic analysis: read all notes first, code each observation descriptively, group codes into candidate themes, then review and refine themes against the data. Affinity mapping offers a collaborative alternative where observations are clustered without pre-defined categories.

Use qualitative methods first to generate hypotheses about what and why, then validate at scale with surveys and analytics. Present combined evidence, such as pairing survey percentages with interview explanations, and investigate disagreements between sources rather than ignoring them.

Good personas emerge from behavioral clusters in research data, not demographics or assumptions. Each persona needs goals, pain points, usage context, representative quotes, and quantitative sizing. Keep the set to 3-5 personas and update them as the product evolves.

Score each opportunity on impact (users affected times frequency times severity), evidence strength, strategic alignment, and feasibility. Use ranges instead of false precision and rank opportunities relative to each other rather than relying on absolute scores.

Common mistakes include reporting averages without distributions, ignoring non-response bias, over-interpreting small differences, treating Likert scales as interval data, and confusing correlation with causation. Always examine response distributions and segment results before drawing conclusions.