How the Recommendation Engine Works

Overview

When a shopper completes your quiz, the recommendation engine:

1. Collects all selected answers
2. Looks up the tag/attribute weights you configured
3. Scores every product in your catalog against those weights
4. Returns the top N products sorted by score

Scoring model

Each answer option can be linked to one or more product tags with a weight (1-10):

Answer: "Oily skin"  →  tag:oily-skin (weight 8), tag:mattifying (weight 6)
Answer: "Sensitive"  →  tag:sensitive (weight 9), tag:fragrance-free (weight 7)

A product's score = sum of weights for all matching tags.

Recommendation modes

| Mode | Behavior | |---|---| | Weighted Score | Default. Products ranked by total weighted score | | Exact Match | Only products matching ALL required tags appear | | Collection Map | Each answer combination maps to a specific Shopify collection | | Manual Override | You pin specific products for specific answer combinations |

Minimum score threshold

Set a minimum score in Results Settings so that low-relevance products never appear. Recommended: 50% of maximum possible score.

Number of results

Configure how many products to show on the results page (1-12). Most brands show 3-4.