The verdict in three sentences
Variants multiply SKUs: a model in 5 sizes x 4 colors = 20 combinations to track separately. The right 2026 architecture is a parent product plus a variant matrix, each variant carrying its own SKU, stock, price and photo. Done right, stock stays accurate, checkout handles per-variant stockouts, and a size chart cuts returns by 10 to 25 %.
The data model that doesn't break
The classic mistake: tracking stock at the parent-product level. You think you have 30 sneakers but miss that only 2 are left in size 42. You must track stock per variant.
| Field | Parent level | Variant level |
|---|---|---|
| Product name | ✅ | inherited |
| Description | ✅ | inherited |
| Category | ✅ | inherited |
| SKU | no | ✅ unique |
| Stock | no | ✅ per combination |
| Price | base price | ✅ adjustable |
| Photo | generic | ✅ per color |
| Size / Color | no | ✅ attributes |
A clear SKU naming convention avoids chaos: TSH-RED-M, TSH-RED-L, TSH-BLUE-M. Readable in the warehouse and online alike.
Naming, stockouts and returns
| Topic | Bad practice | Good 2026 practice |
|---|---|---|
| SKU | random codes | MODEL-COLOR-SIZE |
| Variant stock | tracked at parent | tracked per combination |
| Out-of-stock variant | buyable then canceled | greyed out at checkout |
| Price by size | single forced price | +2,000 FCFA for XL possible |
| Size chart | missing | cm table per size |
| Size-fit returns | 20-30 % | 10-15 % |
| Photo | 1 photo for all | 1 per color |
At checkout, a zero-stock variant must be greyed out, never buyable then canceled: that is the top cause of Wave refunds and bad reviews.
Need a professional website?
Kolonell builds websites that attract clients, optimized for the Sénégalese market. Free quote in 2 minutes.
Mini case study
Chidi, who runs a fashion store in Lagos, sells shirts in 4 sizes and 3 colors (12 variants). At first, stock was tracked at the parent: he showed "available" while blue M was sold out, driving ~18 % returns and cancellations. After switching to per-variant tracking with a size chart, returns fell to 11 %. On 200 sales/month at 12,000 FCFA, cutting returns by 7 points saves ~14 canceled sales, about ~168,000 FCFA/month of revenue preserved.
FAQ
How many variants can I manage? Technically unlimited, but keep it readable: beyond 30-40 combinations per product, group by ranges. Each variant keeps its own SKU, stock and price.
Can I set a different price per size? Yes. The base price sits on the parent, but each variant can adjust it, for example +2,000 FCFA for an XL or a premium finish.
How do I avoid selling a sold-out size? Stock is tracked per variant and checkout automatically greys out any zero combination, with an optional restock alert.
Does a size chart really cut returns? Yes: a cm-per-size table lowers returns by 10 to 25 % depending on category, because customers pick the right fit first time.
Do I need a photo per variant? At minimum one per color. The buyer wants to see the exact red before buying; a single generic photo raises returns and WhatsApp questions.
Let's talk about your project. We structure your catalog into clean variants, with per-combination stock and a size chart. WhatsApp +221 77 596 93 33.
Mohamed Bah
Fondateur, Kolonell
Passionate about digital and entrepreneurship in Africa, Mohamed has been helping Sénégalese businesses with their digital transformation since 2020. Founder of Kolonell, he believes every SME deserves a professional and accessible online présence.

