Re-evaluating personas when engagement is low

WHAT THIS TESTS This tests influence, evidence-based reasoning, and treating research artifacts as falsifiable hypotheses. The interviewer wants to see an engineer drive cross-functional re-examination diplomatically and rigorously, not assign blame.

A GOOD ANSWER COVERS Lead with data, not opinion: present the behavioral analytics showing low engagement, segmented by the user groups the personas describe. Frame personas as hypotheses about users that can be wrong, which depersonalizes the conversation. Then isolate the cause before condemning the persona: low engagement could mean the persona was inaccurate, or the persona was right but the feature did not serve them, or discoverability and execution failed. Propose a structured, joint process with UX and Product: pull analytics to see who actually uses the product versus the assumed persona, run targeted interviews or surveys with real current users, and compare findings against the persona's assumptions. Decide together whether to refine, replace, or retire personas, and document the evidence trail. Position engineering as supplying behavioral signal that complements qualitative research.

COMMON WRONG ANSWERS Unilaterally declaring the personas wrong and useless. Assuming low engagement automatically invalidates the persona without checking execution or discoverability. Going around UX and Product instead of partnering. Bringing opinions instead of segmented data. Treating personas as immutable and refusing to question them at all.

LIKELY FOLLOW-UPS How do you tell whether the persona or the feature was the problem. What data would convince a skeptical product owner. How often should personas be revisited.

ONE CONCRETE EXAMPLE A feature built for a 'busy professional' persona sees almost no use. You pull analytics and find the actual active users skew toward students, not the assumed persona, and the feature assumes desktop use while most sessions are mobile. You bring this to UX and Product as a hypothesis-check, and together you run interviews with real users. The evidence invalidates the original persona's device and context assumptions, so the team updates the persona and reprioritizes a mobile-first redesign, with the decision grounded in data rather than blame.