Product quiz field guide

Product finder catalog-readiness checklist

A reproducible catalog audit for product recommendation quizzes, covering attributes, variants, inventory, exclusions, markets and fallback results.

Answer: a catalog is ready for a product finder when every attribute used in a question exists as structured, consistent data for every eligible product or variant, and when inventory, market and no-match behavior are explicit. Quiz logic cannot repair missing or contradictory product facts.

What is catalog readiness for a product quiz?

Catalog readiness is the state in which product data can support repeatable recommendation rules. Titles and descriptions help shoppers understand an item, but matching usually needs more precise fields. Shopify's current product model includes category, price, inventory, variants and metafields. Category metafields can provide standardized attributes, while custom metafields hold specialized facts such as material, compatibility or product-specific dates.

The 12-field readiness matrix

FieldWhy the finder needs itReadiness test
Stable product IDPrevents title changes from breaking mappingsUnique and never reused
Variant IDSupports shade, size and material matchesEvery selectable variant has one
CategoryDefines comparable products and available attributesSpecific standardized category assigned
Recommendation attributesConnects answers to productsAllowed values, consistent spelling, no mixed units
Hard exclusionsRemoves unsafe or incompatible matchesBoolean or controlled values, never free text only
PriceSupports budget filtering and result displayCurrency and market behavior documented
AvailabilityPrevents recommending an unavailable choiceUpdate source and fallback rule defined
Market eligibilityHandles region-specific productsProduct and variant availability tested per market
ImageMakes the result recognizableCurrent, descriptive alt text available where rendered
Result reasonExplains why the item fitsReason can be derived from selected attributes
Destination URLProvides a stable next stepCanonical product or plan URL returns 200
FallbackHandles empty valid setsAlternative product, waitlist or human route is named

Run a field coverage audit

For each field used by quiz logic, calculate populated eligible items divided by all eligible items. Do not average away a critical gap. A catalog can be 95% complete overall and still be unusable if the only missing values belong to a mandatory compatibility field. Report hard-exclusion and availability coverage separately.

In the 24-SKU Northstar skincare brief, the recommendation fields are skin type, sensitivity, concern, fragrance, price, compatibility, inventory and market. If two variants lack sensitivity data, the coverage for that field is 22 divided by 24, or 91.7%. The correct response is to complete or exclude those variants, not to assume that missing means suitable.

Define a controlled field contract

Each field needs a name, type, allowed values, unit, scope, data owner, update source and missing-value behavior. For example, sensitivity_fit might allow suitable, conditional and exclude. Free-text phrases such as gentle, extra gentle and good for most sensitive skin are harder to map consistently and can hide meaningful differences.

Variant and inventory rules

Decide whether a quiz recommends a product family or a purchasable variant. A shade finder that recommends only the parent product has not finished the matching job. Shopify records price and inventory behavior at product and variant levels, so the sync and result action should preserve the selected variant when it matters.

The minimum logic test pack

  1. One common shopper who should receive the expected bestseller.
  2. One path for every hard exclusion.
  3. Three conflicting combinations where preferences point in different directions.
  4. One boundary value for every numeric rule.
  5. One out-of-stock best match with an available fallback.
  6. One market where a recommended item is not published.
  7. One answer set that should produce no automatic recommendation.

Write the expected result before running the quiz. Record the catalog snapshot, logic version, actual result and discrepancy. This makes a retest meaningful after inventory or product data changes.

Limits of the checklist

A complete catalog does not prove that the recommendation model is correct. Domain judgment is still needed to choose attributes, exclusions and weights. Live inventory can also change between result and checkout, so the cart must validate the selected variant again.

Sources

Method: a reusable field-coverage and edge-case audit applied to the 24-SKU Northstar Skincare Finder brief. Limitation: platform data models and sync intervals differ, so verify field availability in the selected builder. Next retest: after any catalog schema, market or inventory-sync change.

Corrections: send the page URL, the exact statement and a current source through the contact form.

Compare the eight product recommendation quiz builders or read the full testing protocol.