How to Help Customers Find the Right Product in a Large or Technical Catalogue

Why ordinary category menus fail technical buyers

A category menu assumes the customer already understands how the seller organises the catalogue. That works for simple ranges, but it breaks down when a buyer starts with a machine model, dimensions, operating conditions, compatibility question or replacement part number. The customer may recognise the problem while having no idea which internal category contains the answer.

A guided finder reverses that journey. It begins with what the customer knows, asks only questions that narrow the valid choices and returns a shortlist supported by catalogue evidence. Search and filters still matter, but they become parts of a wider decision path rather than the only ways into the catalogue.

Start with the customer decision, not every available field

Before building a finder, choose one decision it must help someone make. “Find the correct replacement belt for this machine” is testable. “Make the catalogue easier” is not. Observe the questions customers ask sales staff, the details support teams request and the reasons products are returned or abandoned.

Turn those observations into a short decision tree. Ask high-value questions first and hide fields that do not change the recommendation. If machine make and model determine the compatible family, ask for them before colour or delivery preference. If a measured size is essential, explain how to take the measurement and which unit to use.

• What does the customer usually know when they arrive?
• Which answer removes the largest number of unsuitable products?
• Which compatibility rule must never be treated as a preference?
• What evidence should appear beside the recommendation?
• When should the finder stop and ask for human help?

Build the finder on structured product facts

A reliable recommendation needs fields that mean the same thing across the catalogue. Normalise units, attribute names, brand and model references, product types and compatibility relationships. Keep display copy separate from the values used for matching so a wording change does not alter the underlying rule.

Stable identifiers are equally important. GS1 explains that a GTIN uniquely identifies a trade item. Where GTINs are appropriate, preserve them alongside the platform product and variant IDs, manufacturer references and internal SKUs. Do not rely on an editable title as the identity of the product being recommended or updated.

1. List the requirements customers use to distinguish a suitable product.
2. Map each requirement to an approved catalogue field or relationship.
3. Normalise values and units before writing matching rules.
4. Keep stable product and variant identifiers attached to every result.
5. Flag missing or conflicting facts for review instead of guessing.

Separate hard compatibility rules from preferences

A hard rule decides whether a product is valid. A preference helps rank several valid products. A part that does not fit the customer’s machine should never appear because it is popular, in stock or commercially attractive. Once invalid options are removed, preferences such as brand, price range, delivery time or material can order the remaining results.

Make exclusions understandable. “Not compatible with model year 2022” is more useful than a result silently disappearing. When the catalogue does not contain enough evidence for a safe answer, say what is missing and offer a clear route to support. An honest no-result path protects trust better than an invented recommendation.

Use filters and search as part of the guided journey

Shopify’s official guidance allows merchants to configure storefront filters and customise product discovery through Search & Discovery. Those controls are useful when product data is complete and customers understand the available attributes. A finder can use the same approved product facts while presenting them as questions in the order a buyer naturally answers them.

Keep the two experiences connected. A customer who arrives through search should be able to refine by compatible attributes. A finder result should link to the real product or variant page rather than a duplicated landing page. If Shopify is the storefront, use the authorised store connection and preserve its product and variant identities throughout the matching and hand-off.

A concrete example: choosing a replacement part

Imagine a catalogue containing thousands of belts used across lawn equipment, motorcycles, industrial drives and vehicles. A buyer searches for “belt for Model 123” but several manufacturers use similar model names. The finder first asks for industry, then manufacturer and product type, then the exact model. It can request dimensions only when the model record is incomplete.

The result is not a long search page. It is a short list of products whose approved compatibility records match the answers. Each result shows the matched machine, relevant dimensions, product reference and any qualification the customer must check. If two products remain possible, the finder explains the difference instead of choosing silently.

Design the result page to support a confident decision

Show why each product was selected. Repeat the customer’s important answers, highlight the attributes that matched and distinguish required checks from optional preferences. Google’s product-variant documentation describes ways to represent variations of a parent product; the same discipline is useful inside the catalogue because size, material or pattern must remain attached to the correct selectable variant.

A good result page also preserves normal commerce behaviour: current availability, price, delivery information, product imagery and a direct route to the correct product detail page. Do not turn a recommendation into a dead end that forces the customer to repeat the search.

• A concise explanation of why the result fits
• The exact product or variant identity
• Matched compatibility facts and dimensions
• Any condition the customer still needs to verify
• Current commercial information from the storefront
• A human-help route for ambiguous cases

Measure whether the finder is genuinely useful

Track completed finder journeys, no-result reasons, refinements, product clicks and requests for help. Where consent and platform capabilities allow, compare those signals with add-to-cart, purchase and return outcomes. The purpose is to find catalogue gaps and confusing questions, not to claim that every click proves a correct recommendation.

Review failed journeys with catalogue and customer-service teams. A frequent no-result may indicate a missing product relationship, an unrecognised customer term or a genuinely unsupported requirement. Improvements should update the governed product facts or question path so the same issue is resolved consistently for the next customer.