How does a quiz decide which product to recommend?
A product quiz works like a smart filter. Each answer a customer gives maps to one or more product attributes, and the quiz narrows your catalog down to the items that fit. Pick a size, and it keeps that size. Pick a skin type, roast level, or fitness goal, and it filters again. By the last question, only the best-fit products remain.
The signals behind that filtering come from your product data: variants, options, metafields, and tags. Quizzo reads all of it and scores every product against the customer's answers. You don't build a rigid, category-specific tag list by hand. The app generates the eligibility mapping from your catalog, then lets you refine it if you want tighter control.
What attributes does a quiz match against?
A quiz can match almost any structured attribute your products already carry. The exact fields depend on your vertical, but the pattern is the same everywhere: a customer's answer points to an attribute, and products carrying that attribute become eligible. Here is how the same mechanic plays out across four different catalogs.
| Store type | Quiz question | Product attribute matched |
|---|---|---|
| Apparel | "What size are you?" | Variant size (S, M, L) + fit tag |
| Apparel | "Which colors do you like?" | Variant color option |
| Skincare | "What's your skin type?" | Metafield: skin_type (oily, dry, combination) |
| Coffee | "How do you like your roast?" | Metafield: roast_level (light, medium, dark) |
| Coffee | "How do you brew?" | Grind option (whole bean, espresso, filter) |
| Supplements | "What's your main goal?" | Metafield: goal (energy, recovery, focus) |
| Supplements | "Any dietary needs?" | Tag: vegan, gluten-free |
| Any catalog | "What's your budget?" | Product price range |
Variant options (size, color, format)
These are the attributes you already set on every product. A quiz can read variant size, color, or format directly, no extra work needed. Ask "What size?" and the quiz keeps matching variants. Ask "Whole bean or ground?" and it filters the same way. This data is the cleanest signal a quiz has.
Metafields (skin type, roast level, goal)
Metafields hold the attributes variants can't express, like a serum's skin type or a supplement's goal. If your catalog already uses them, the quiz maps answers to them automatically. If not, adding one or two metafields per product sharpens matching a lot. [PERSONAL EXPERIENCE] In our experience, stores that add a single "goal" or "type" metafield see the biggest jump in quiz relevance.
Tags (flexible extras)
Tags cover everything else: dietary flags, use cases, seasonal collections, or bundles. They're flexible strings, so they're handy for signals that don't fit a variant or metafield. Quizzo can auto-generate eligibility tags from your catalog, so you rarely need to invent a tag scheme yourself. When you do want one, tags are the easiest place to add it.
Why do the best products match more than one answer?
The strongest quiz results come from products that fit several answers at once. A medium blue running tee sized for warm weather can match a size answer, a color answer, and a use-case answer together. Each matching attribute raises that product's score, so it outranks items that fit only one answer. Rich attributes are how your best-fit products rise to the top.
Think of it as a scoring system, not an on/off switch. A quiz collects several data points across the questions, then ranks every eligible product by how many of them it satisfies. A product matching three answers beats one matching a single answer. [UNIQUE INSIGHT] That's why thin product data hurts more than a missing question: a quiz can't rank what your catalog never describes.
How does Quizzo auto-build the quiz from your catalog?
Setting all this up by hand would be slow. Quizzo skips it with one-click auto-generation. Instead of asking you to design questions and hand-tag eligibility, the app reads your catalog and builds both for you. It's the fastest way to launch a quiz that already knows your products. Here's how the flow works.
Auto-generate from your catalog
From the Quizzo dashboard, click Auto-build quiz. The app fetches your full catalog and reads every product's title, type, options, variants, and metafields to understand what you sell.
Review the generated quiz and mapping
Quizzo proposes a set of questions plus the answer-to-product mapping and auto-generated eligibility tags. Every question, answer, and match is editable. Rename questions, remove ones you don't need, or adjust which products a given answer surfaces.
Publish and let eligibility stay in sync
Click Publish to go live as a theme block or a popup. Quizzo keeps eligibility mapped to your catalog, so you're not hand-maintaining a tag taxonomy as products change. Your existing tags stay intact, nothing is removed.
[IMAGE: Shopify product quiz builder dashboard showing auto-generated questions - Pixabay search "dashboard analytics screen"]
How do I refine matching manually?
Auto-generation gets you most of the way, but you'll always know your catalog better than any tool. When you want tighter control over a specific product, or the quiz needs a signal your data doesn't yet carry, add or edit attributes in Shopify Admin. It takes a minute per product.
- Go to Shopify Admin ā Products
- Click the product you want to refine
- Set the right variant options and any relevant metafields (skin type, roast level, goal)
- Scroll to the Tags field and add a signal you want the quiz to catch
- Click Save
For bulk edits, use Shopify's built-in bulk editor: select multiple products from the product list and apply the Add tags or metafield actions to all of them at once. Then reopen the Quizzo dashboard to fold the new signals into your matching.
What happens when you add new products later?
New products don't break the quiz. When you re-run the auto-build, Quizzo reads the additions and folds them into the existing mapping, so items already covered stay untouched and only the new ones get eligibility. That's the payoff of automatic eligibility: your quiz grows with your catalog instead of drifting out of date.
What if a product doesn't get matched?
Auto-generation leans on the data a product carries. If a product has a vague title, no product type, and no metafields, the quiz has little to match it on. This is intentional, the app would rather leave a product unmatched than surface it for the wrong answer and erode trust in the results.
For products with creative names, "The Cloud Nine" instead of "Cotton Crew Tee", the fix is simple. Add a clear product type, a variant option, or a metafield describing the attribute you want the quiz to catch, then re-run the auto-build or map the product by hand in the dashboard.