# When should you re-ask quiz questions? | PrefEngine

Source: https://prefengine.com/blog/2026-09-26-when-to-reask-quiz-questions/

PrefEngine answers


When the answer is likely to have changed, or when the behavior says it has. A quiz answer is a snapshot of the shopper at one moment, and moments pass: skin changes with the seasons, budgets move, tastes evolve. The best programs treat preference data as perishable, with different expiry dates for different questions, and re-ask at the moment the shopper gets value from answering again.

## Why answers go stale

Zero-party data is prized because it is explicit and accurate, but explicit does not mean permanent. A shopper who said their budget was under $50 gets promoted and stops caring about price caps. A skincare quiz answered in winter describes different skin than the same shopper has in July. A coffee drinker who loved dark roast discovers light roast and never tells anyone.

Stale answers are worse than no answers. A personalization engine acting on a three-year-old quiz confidently recommends the wrong products, and the shopper cannot tell why the recommendations feel off. They just stop trusting them.

## Not all answers expire equally

* Stable facts rarely need re-asking: skin tone, shoe size, dietary restrictions. Ask once, keep forever, and let the shopper update them in a preference center.
* Seasonal answers expire on a schedule: skin concerns, fragrance preferences, wardrobe needs. Re-ask when the season turns, not on an arbitrary anniversary.
* Intent answers expire fast: budget, occasion, gift recipient. These describe a mission, not the person, and the next mission is usually different. Treat them as single-session data unless the shopper says otherwise.
* Taste answers decay slowly and silently: style preferences, flavor profiles, brand affinities. Nobody announces a taste change; the behavior tells you, which is the signal to re-ask.

## The behavioral triggers that say re-ask

You do not need a calendar to know an answer died; the shopper's behavior usually confesses. A "budget under $50" shopper who keeps buying $120 items is voting against their old answer. A shopper whose recommendations get ignored for weeks is telling you the profile is wrong. Build triggers on the mismatch: when behavior contradicts the stated preference consistently, it is time to ask again.

The trigger should fire before the frustration does. Three consecutive ignored recommendation sets, a category shift that persists for a month, or a return pattern that suggests the matches were wrong: any of these is a better prompt than a yearly "update your preferences" email nobody opens.

## How to re-ask without annoying people

* Ask one question, not the whole quiz. "Still shopping for sensitive skin?" with a yes/no is a refresh; a twelve-question quiz is a chore.
* Prefill the old answer and let them confirm or change it. Editing is easier than answering from scratch, and the confirmation itself is a signal.
* Time it to a natural moment: after a purchase, at the season change, or when they return after a long absence. Random popups asking about preferences feel like interrogation.
* Show the payoff. "Update your skin profile for better matches" works; "Please update your preferences" does not. The shopper should know what they get for the thirty seconds.

## Keep the old answers, timestamp everything

Re-asking does not mean deleting. Keep every answer with its timestamp, because preference history is itself data: the shopper who moves from budget to premium is on an upgrade path worth recognizing. Timestamped answers also let you audit: if re-asked answers flip constantly, the question is probably measuring a mood, not a trait, and should be treated as single-session data.

The operational rule is simple: store answers as a history, not a snapshot. The current answer drives personalization; the history drives the re-ask schedule and the insight. A preference engine that remembers what the shopper used to want is more useful than one that only knows what they want now.

## Reviewed

Published Sep 26, 2026.