Ecommerce catalogue architecture: help customers find the right product

Map buying decisions before drawing menus

Invite a salesperson, a customer support colleague and someone responsible for product data to describe how buyers narrow their choices. Capture the language customers use, the questions they ask and the mistakes they make. An internal supplier category may be useful for purchasing while being meaningless to a first-time visitor.

For each product family, identify the primary decision: application, compatibility, material, size or desired result. Keep those dimensions separate. A customer should not need to guess whether “bathroom”, “waterproof” and “decorative” are competing categories or overlapping properties. Clear relationships let the website offer several sensible routes without turning every possible combination into a separate collection.

Give categories, filters and guides different jobs

A category offers a stable destination for a recognisable shopping intention. A filter narrows a set of products. A guide helps somebody understand which criteria matter. Confusing those jobs creates menus that grow indefinitely and thin pages that say almost the same thing.

Google’s ecommerce guidance explains that navigation and links help it understand a site’s structure. That supports a practical approach: make important categories reachable through ordinary links, and give each useful category a clear purpose. The commercial reason is equally strong. A category should answer “am I in the right place?” before asking the customer to inspect individual products.

A hypothetical building-products catalogue

Imagine a retailer selling wall panels, flooring, trims and installation materials. Its current menu follows manufacturers. A homeowner looking for a kitchen wall finish has to understand the suppliers before understanding the products. The proposed architecture gives wall panels a clear main category, with filters for application, finish and dimensions.

Installation guidance explains when trims or adhesives are needed, while product pages link to compatible accessories. Trade buyers can still browse by brand through a secondary route. This is a design hypothesis, not evidence of a sales uplift. Test it by asking people with realistic buying tasks to find a suitable product and explain why they chose it.

Create a product-data contract

The structure will only work if the data supports it. Define what each attribute means, its permitted values and who maintains it. “Natural oak”, “oak natural” and “light oak effect” may describe different things or accidental duplicates. Resolve that ambiguity before building a filter that exposes it to customers.

Make important attributes required for new products and document how unknown values are handled. A missing compatibility value should not silently imply compatibility. Give the merchandising team a simple exception list, so gaps can be corrected without a developer. Also decide how discontinued products, replacement items and temporary stock shortages should appear in the customer journey.

Use this catalogue review checklist

Review a representative family first. Include bestsellers, unusual variants, a discontinued item and products with incomplete data. That sample will reveal more useful problems than polishing the easiest category alone.

• Can a first-time buyer identify the correct category without knowing a supplier name?
• Does each filter describe a meaningful difference and show understandable values?
• Can customers compare the attributes that actually change the purchase decision?
• Do accessories and alternatives have explicit, maintainable relationships?
• Can the team explain which pages deserve search visibility and which combinations are only browsing controls?
• Is there an owner for category names, product attributes and obsolete content?

Make the structure governable

More categories offer more tailored entry points, but each needs useful content and maintenance. Fewer categories reduce administration, but can force buyers through too much filtering. The right balance depends on the range and the customer’s familiarity with it. Avoid a universal rule such as a fixed maximum menu depth without testing the actual tasks.

Google’s URL guidance is useful when deciding how categories, variants and parameters will be represented technically. Agree those rules with development before migration. Then measure failed searches, repeated filtering, support questions and product discovery alongside sales. A better catalogue is one customers can understand and the business can keep accurate.