Scent-finder quiz, how do you build one that actually moves conversion?

Scent is invisible online. A quiz does the work the candle counter does in-store. Built right, recommended-product PDP conversion runs 4-6% vs. 1.2-1.8% for unguided category browsing, a 2.5-4x lift on first-touch traffic.

The stack I default to:

• Octane AI, quiz UX, branching logic, Klaviyo native integration. ~$50, $500/mo depending on volume. Best for fragrance specifically because the branching trees (mood × season × occasion × intensity) get deep.
• Klaviyo, segments the customer by quiz result (e.g. “cozy / woody”, “fresh / citrus”, “floral / soft”) and runs an abandoned-quiz email flow if they don’t convert. Recovery rate runs 8-14%.
• Magento integration, quiz result writes to a customer attribute (scent_profile) and pre-fills the cart with the matching recommended candle / diffuser / perfume. Re-visit experience: “Welcome back, based on your scent profile, here’s the new arrival you’d like.”

Quiz design rules I’ve learned the hard way: 5-7 questions max (drop-off jumps past 8), use sensory imagery not text-only options, never make the user type, end on a single hero recommendation (not a list of 5, analysis paralysis), let logged-in users skip the quiz.

Diptyque, Otherland, Boy Smells, and P.F. Candle Co. all run variants of this. The pattern works.