SKU Rationalisation

A retailer sells a shirt in several colours and sizes, and each stocked variant may have a different SKU; some sell quickly while others sit on shelves for months. Rationalisation compares those variants using sales, margin, inventory and customer needs, and Shopify describes auditing each SKU's contribution and deciding whether low-demand products should stay.

The unit must be defined first, since a product family can look profitable while one colour or pack size performs badly, and the catalogue should be cleaned because duplicate codes, renamed items and returns can distort sales history. Demand should be measured through sales units, frequency and seasonality, because a slow item may be valuable at a particular time of year.

Margin should be measured using contribution rather than revenue alone, as a high-selling product with heavy discounts may contribute less than a smaller specialty item. Holding costs such as storage, insurance, shrinkage and financing need estimating too, since slow stock ties up cash.

The service role matters, because a rare replacement part may be essential for customer support despite low sales, and baskets deserve a look, since a low-margin item may bring customers who also buy profitable products. Substitutions need watching, as removing one SKU may shift demand to a similar item or lose customers altogether, and Oracle's assortment work considers demand transference.

Channels and regions matter as well, because an item may sell online but not in a physical shop, and climate, local tastes and regulations can make a global deletion wrong for one market. Supplier terms affect the economics of each variant, since minimum order quantities, lead times and return rights differ, and shelf space matters, because a slow SKU can occupy space that could generate more value with another product.

Complexity has a cost even when storage is cheap, since every SKU needs forecasting, replenishment, counting and support, and recently launched items have not had a full demand cycle and deserve a suitable trial period. The exit cost of end-of-life stock should be included too, because discontinuing a SKU leaves inventory that may need markdown, transfer or disposal.

Customers should be asked, because buyers may depend on a niche item that sales reports understate, and criteria with thresholds and exceptions should be set before reviewing favourites to reduce subjective decisions. A decision matrix helps, since keep, improve, substitute and discontinue are all possible outcomes and a binary cut-or-keep rule is crude.

Stockouts, returns and bundles should be examined before removing an item: a low sales figure may reflect poor availability rather than weak demand, a popular item with high returns can be less attractive than gross sales suggest, and a standalone SKU may support a profitable kit or subscription. Supplier exits need planning, because a contract may require notice or minimum purchases, and removals should be tested by taking out a small set first and then watching total category sales and complaints, since a good theory can fail in practice.

Changes should be communicated so sales staff and customers know the alternatives, and cash release should be tracked because inventory value removed is not automatically cash received. The review should be refreshed regularly because a one-time cleanup does not prevent future sprawl, and for owners the best assortment serves customers while using cash and operational attention well, so the right SKU count is an outcome, not the goal itself.