A stock keeping unit (SKU) is a code a business creates to identify one specific item it sells, down to its size, colour or other variation. A navy jumper in medium and the same jumper in large are two SKUs, because they are counted, stocked and sold separately.
How SKUs work
A SKU is internal. You choose the format, and it only has to be unique within your own business. Many retailers build it from readable parts, such as JMP-WOOL-NVY-M for a navy wool jumper in medium. Others use plain numbers generated by the ecommerce platform or stock system. Either approach works as long as every sellable variation has exactly one code and that code never changes.
A SKU is different from a GTIN, the barcode number (usually an EAN in the UK) that identifies a product worldwide and is issued through GS1, normally by the brand. Two shops selling the same kettle share its GTIN but each has its own SKU for it. It also differs from the manufacturer part number (MPN), which the brand assigns.
Inside the shop, each product variant normally carries its own SKU, and the variants sit under one parent product. That structure carries through to everything that depends on it: stock control, the warehouse or fulfilment partner, accounting, and the product feed sent to Google and Meta, where the SKU is often used as the item ID and the parent code as the item group ID.
Why it matters
Clean SKUs are what let separate systems agree on what was sold. When codes are missing, duplicated or reused, stock figures drift, the wrong item gets picked, and reports cannot tell you which size or colour actually sells.
For marketing, the SKU is the thread that links an ad click to a product. Shopping ads, Performance Max and Meta catalogue ads report results by item ID, and the purchase event on your website should send the same ID in its item data. If the feed uses one code and the site tracking uses another, the platforms cannot match purchases to products, and dynamic retargeting may show people items that are out of stock or no longer sold.
Changing item IDs also wipes history. Google and Meta treat a new ID as a new product, so recoding your range or moving ecommerce platform carelessly can throw away months of performance data on your best sellers.
Common mistakes
- Reusing a discontinued product’s SKU for a new item, which merges two products’ sales history.
- Giving every size and colour the same SKU, so stock and sales cannot be tracked by variant.
- Letting the website, the warehouse and the feed each use different codes for the same item.
- Changing every SKU during a platform move without keeping a map from old to new.
- Building meaning into codes that later changes, such as a supplier’s name, then having to break the pattern.
- Typing the SKU into the GTIN field because the feed demanded a value there.
How to act on it
Export your product list and check three things: every sellable variant has a SKU, no SKU appears twice, and none are blank. Then compare that list with the item IDs in your product feed and the item IDs your purchase tracking sends. They should match exactly.
Write down a simple format for new products and keep to it. If you are changing ecommerce platform, keep existing SKUs where you can; where you cannot, keep a mapping table so stock, reports and feeds survive the move.
Once SKUs, feeds and tracking line up, catalogue-based advertising becomes much easier to judge. Checking that data is the first step when I set up catalogue and dynamic product ads, and the same clean feed serves Google Merchant Center for Shopping ads.