Key takeaways
- Choose a Shopify product recommendations app by the merchandising jobs it must perform, not by the length of its feature list.
- Define each placement separately because product-page substitutes, cart add-ons, collection discovery, and search recommendations solve different customer problems.
- Record catalog constraints before evaluating apps, including variant availability, product relationships, seasonal inventory, taxonomy quality, markets, and collection structure.
- Reject any app that fails a must-have acceptance test, even when its overall feature score looks strong.
This worksheet turns a broad app search into a purchasing decision. Complete one row for every recommendation placement, then mark each requirement as must-have, useful, or unnecessary. The selected Shopify product recommendations app should advance the identified merchandising job without exposing unavailable products, suggesting incompatible items, or creating more manual work than the team can maintain. Use the completed worksheet during app reviews, demos, trials, and implementation planning. Do not award points for capabilities that the store will not use during the next merchandising cycle.
Define the recommendation job before the feature
Start with the customer decision that needs help. A label such as related products is too vague to evaluate because it could mean substitutes, accessories, next-step products, or items sharing a collection. Write the job as a specific outcome: help a shopper compare similar trail shoes, add compatible filters to a coffee machine, or move from an unavailable dress to an in-stock alternative.
Create a separate worksheet row for each placement. Common jobs include product-page alternatives, product-page complements, cart add-ons, collection discovery, search-assisted recommendations, and post-purchase suggestions. For every row, name the trigger, placement, product relationship, and exclusion rule. A useful example is: when a shopper views a $120 espresso machine, show three compatible accessories under $40, exclude other machines, and suppress unavailable items.
Do not combine bundles and complementary recommendations without checking the purchase logic. A fixed kit has different pricing and inventory implications from an optional cross-sell. Use the bundles versus complementary products decision guide when that distinction is unclear. If the actual problem is poor search or collection discovery, review the guidance on improving Shopify product discovery before adding another placement tool.
Complete one worksheet row for every placement
The worksheet should make every vendor prove fit against the same inputs. Copy this table into a spreadsheet, create one row per placement, and add columns for priority, owner, vendor result, and notes. A store with six placements should have at least six rows rather than one store-wide rating.
| Criterion | What to check | Why it matters |
|---|---|---|
| Merchandising job | Substitute, complement, comparison, discovery, add-on, or repeat purchase | Each job needs different product logic |
| Trigger and placement | Product view, search, collection, cart, or completed order | Context changes what counts as relevant |
| Eligible products | Collections, product types, tags, price ranges, or explicit product sets | Broad eligibility can produce weak suggestions |
| Exclusions | Unavailable items, incompatible variants, gift cards, samples, or clearance stock | Exclusions prevent commercially harmful results |
| Control level | Automatic, rule-based, manually pinned, or combined | Control determines workload and predictability |
| Acceptance test | Input product, expected results, prohibited results, and failure state | A repeatable test prevents subjective scoring |
Classify each row as must-have, useful, or out of scope. A must-have should represent a current commercial or customer-experience requirement, not a possible future campaign. Set the scoring rule before reviewing vendors: reject an app that fails any must-have compatibility or exclusion test. Score useful requirements from zero to two, where zero means unsupported, one means partially supported, and two means supported under the store's actual operating conditions.
For a broader planning framework, compare these rows with the product discovery app requirements worksheet. Keep recommendation results separate so a strong search capability does not conceal a weak cross-sell result, or vice versa.
Which catalog constraints can change the result?
Catalog structure determines whether recommendation logic can return dependable products. Record the fields available for matching, including product type, vendor, collection, tags, metafields, price, options, and inventory state. Document where those fields are incomplete or inconsistent. If half of a furniture catalog lacks dimensions, size-compatible accessory recommendations are not ready for evaluation.
Test difficult catalog segments rather than only best sellers. Select at least five fixtures: a high-traffic product, a new product with little behavioral history, an unavailable item, a product with many variants, and a niche item with few valid complements. For every fixture, list acceptable and prohibited outputs. A red phone case should never be recommended for a phone model it does not fit, even when both products share a broad accessories collection.
Record market, language, currency, and seasonal restrictions where they apply. A valid recommendation in one market may point to an unavailable product in another. If filters, search, and recommendation eligibility rely on the same catalog fields, complete the Shopify storefront filtering readiness checklist. Fixing taxonomy first can reduce configuration work and prevent conflicting discovery rules.
Controls and acceptance tests settle the decision
Require the lowest level of control that protects the merchandising job. Manual curation provides predictability but creates upkeep across large or frequently changing catalogs. Automatic selection reduces routine work but requires strict eligibility and exclusion rules. A combined model can use automatic candidates inside approved boundaries, with manual pinning reserved for launches, campaigns, or contractual priorities.
Write pass-or-fail tests before starting a trial. For a complementary placement, a test might require that viewing SKU MACHINE-01 returns exactly three accessories, every item fits that machine, all three are available, no item exceeds $50, and no competing machine appears. For an alternative-product placement, require the same category and intended use, permit a price band such as 20% below to 30% above, and prohibit the currently viewed product.
Include failure states. Decide what the storefront should do when no eligible recommendation exists, inventory changes, or a rule leaves only one result. Hiding an empty block is usually preferable to filling it with unrelated products, but the correct decision depends on the placement. Test representative mobile layouts as a separate presentation check so a visual defect is not confused with a recommendation-quality failure.
As of August 2026, every scorecard should also name the owner responsible for rules, campaign changes, catalog data, and scheduled test reruns. If no one owns a control, count that control as an operating cost rather than a benefit.
Assess Hyper Search & Filter against the worksheet
Complete the requirements worksheet before assessing Hyper Search & Filter. Bring the placement rows, catalog fixtures, prohibited outputs, and ownership limits to the review. The decision is not whether Hyper Search & Filter has the longest capability list. The decision is whether it supports the discovery jobs documented under the conditions present in the Shopify catalog.
Run the same fixtures against every shortlisted option. Reject a candidate when it fails a must-have compatibility, availability, or exclusion rule. For candidates that pass, compare setup effort, ongoing merchandising time, storefront behavior, and the number of useful requirements supported. Keep subscription cost separate from implementation and maintenance costs. A lower app fee can be offset by weekly manual curation, while greater automation may be unsuitable when exact product control is commercially necessary.
If search relevance is part of the same project, use the Shopify search relevance audit tool as a separate test. Merchants choosing between native capabilities and another app can also consult the native search versus third-party app guide. Make one decision per discovery layer rather than expecting a recommendation score to resolve every storefront problem.
FAQ
Which Shopify product recommendations app should I use?
Use the Shopify product recommendations app that passes every must-have job, exclusion rule, catalog fixture, and ownership requirement in your worksheet. A store needing manually governed compatibility recommendations should not apply the same criteria as a store prioritizing automatic discovery across thousands of loosely related products. Compare pricing only after acceptance tests establish functional fit.
How do related products work on Shopify?
Related products show shoppers items selected through automated logic, merchant configuration, theme behavior, or an installed app. The relationship may represent similarity, a shared category, purchase context, or another configured signal. The Shopify theme must also render the placement. Inspect actual outputs because products classified as related are not necessarily interchangeable or compatible.
How do I add complementary products in Shopify?
Add complementary products by defining sensible additions to a primary product, configuring those relationships in a supported Shopify tool or app, and displaying them through a compatible theme section. Test actual products and variants before release. Complementary products add to the original purchase, while substitutes give the shopper an alternative to the original item.
Is Shopify Search & Discovery free?
Shopify Search & Discovery is generally offered without a separate app subscription charge, but merchants should confirm the current Shopify listing and account eligibility before deciding. A free app can still require theme setup, taxonomy work, testing, and ongoing maintenance. Compare total operating effort rather than treating the subscription price as the full cost.
