A large product catalogue needs more than more categories
Organise product codes, variants, filters and imports so customers can find products and your team can maintain the catalogue.
Alphabite · Ecommerce design and development team ·
The first problem in a large catalogue is often inconsistent information. Before adding filters, decide what each product attribute means and whether the data is complete enough to use it.
Start with a difficult sample.
Choose a few dozen representative products: a bestseller, a product with many variants, an unavailable item and something that belongs in more than one collection. This is a practical test sample, not a required catalogue size.
Give every sellable variant a stable code. Decide whether “gold” describes colour, material or plating. If those meanings share one field, the filter can mislead shoppers even when the software works correctly.
Categories, filters and variants do different jobs.
A shopper should not have to understand your supplier’s spreadsheet to find a product. Build navigation around the selection they are making.
| Structure | Purpose | Example |
|---|---|---|
| Category | Choose the kind of product | Necklaces |
| Filter | Narrow the selection | Material or adjustable length |
| Variant | Choose a sellable version | A specific length with its own stock |
| Collection | Browse a shared idea | A gift selection across product categories |
Example for a jewellery catalogue
Category
- Purpose
- Choose the kind of product
- Example
- Necklaces
Filter
- Purpose
- Narrow the selection
- Example
- Material or adjustable length
Variant
- Purpose
- Choose a sellable version
- Example
- A specific length with its own stock
Collection
- Purpose
- Browse a shared idea
- Example
- A gift selection across product categories
Make the assortment easy to explore.
Lumies Jewelry is a useful example from our work: a broad jewellery assortment needs both visual browsing and dependable product structure. Attractive collection imagery gets people started; accurate attributes help them continue.
Test the update, not only the first import.
Import your sample twice. Correct a price, remove an image and mark a variant unavailable. Check that the second import updates the intended records instead of duplicating products or restoring old information. Keep a record of rejected rows that a person can understand.
On the public site, decide which category and product pages should be discoverable in search. Google documents ProductGroup markup for related variants; structured data must describe the variants people can actually see and select. It cannot repair missing product data.