SKU rationalization: pruning the tail, explained

What SKU rationalization is

SKU rationalization is the deliberate cutting of low-velocity products from a line so the survivors get more shelf, more inventory budget, and more attention. The first time I ran this exercise at a natural-products brand, we had grown to nineteen SKUs almost by accident, a flavor here for one retailer, a pack size there for a club test, and four of them together did less than 6% of revenue. Pruning them did not dent sales. It freed working capital we had tied up in slow parquet on a warehouse floor.

The instinct that makes this hard is that every SKU has a defender. Somebody launched it, somebody promised a buyer it would stay, somebody has a customer who loves it. Rationalization is the process of overriding those stories with velocity numbers.

The 80/20 of any line

Almost every CPG line obeys a brutal version of the Pareto split. A small set of SKUs does the overwhelming majority of the volume, and a long tail of items does almost nothing while consuming the same setup, inventory, and reporting overhead as the winners. Here is a representative twelve-SKU line ranked by units per store per week, with a cut line drawn in.

The top five SKUs do 84% of volume. The bottom five do 6% combined. Each unit in that bottom group still needs its own UPC, its own slot in the planogram, its own minimum production run, and its own line on every reporting deck. The math for cutting them is not subtle.

The velocity floor

The cleanest rationalization tool is a velocity floor: a units-per-store-per-week threshold below which a SKU has to justify its existence or go. In the table above I would set the floor around 3 units per store per week. Everything under it is a candidate, which puts SKUs 8 through 12 on the block and SKUs 6 and 7 on watch.

The floor is not a religion. A SKU can sit below it and still earn a stay for real reasons: it is six months old and still ramping, it is a strategic item that anchors a club listing, it is a halo flavor that drives trial. The point of the floor is to flip the default. Instead of every SKU staying until someone argues to kill it, every below-floor SKU goes until someone argues to keep it. That one inversion is most of the value.

Check cannibalization before you cut

Here is the trap that turns a clean rationalization into a self-inflicted sales loss. Before you cut a low-velocity SKU, you have to know whether its buyers will trade up to another item in your line or walk to a competitor. That is a cannibalization question, and getting it wrong is expensive.

Say you cut Habanero 6oz at 2 units per store per week. If those shoppers were heat-seekers who will happily switch to Sriracha 6oz, you lose almost nothing and your Sriracha velocity ticks up. But if Habanero was the only super-hot item in your line and those shoppers were buying it specifically, cutting it sends them to a competitor's hot sauce, and you may quietly lose the basket entirely. Two SKUs with identical 2-unit velocity can have opposite cut decisions depending purely on where their buyers go next.

The practical check: before pruning, look at whether the candidate SKU overlaps heavily with a survivor (high transfer, safe to cut) or occupies a distinct position with no nearby substitute in your line (low transfer, cut at your peril). Velocity tells you which SKUs are weak. Cannibalization analysis tells you which weak SKUs are safe to remove.

SKU rationalization from the retailer's side

The same words describe a different exercise when a retailer runs it. A brand rationalizes its own line; a retailer rationalizes a whole category across competing suppliers, and the decision rule changes as a result.

The retailer's cut line is not a velocity floor, it is a carrying cost. An item earns its position when its marginal contribution beats the cost of the space it occupies. In a center-store set returning $288 in annual sales per selling square foot, a single-facing item generating under roughly $260 a year is not covering its own shelf, whatever its velocity looks like in isolation.

That threshold usually cuts deeper than a velocity floor and shallower than the raw ranking suggests. In one 62-store operator's center-store set, the bottom three deciles were 30% of SKU count and 3.6% of dollars, which reads as an easy cut until the four protective filters are applied: destination items carrying a basket, items that are under-distributed rather than slow, items still inside their launch window, and price-tier anchors. About half the raw list survives those.

The cannibalization check above still applies, with one addition. A retailer cutting a SKU can recover the volume elsewhere in the category, so the number to watch after a reset is lost-item recovery: how much of each cut item's volume reappeared in the surviving set. Recover 80% and the cut was right and could go deeper. Recover 30% and those items were serving a need the rest of the set does not, and the cut line is too aggressive for that category.

For the full framework, see assortment optimization.

Where Scout fits

Ranking a line by velocity per store and spotting the tail is exactly the kind of slice that takes an afternoon in spreadsheets and seconds once the data is connected. Scout reads your SPINS or retailer data and surfaces per-SKU velocity, cumulative volume share, and the candidates below any floor you set, so the rationalization conversation starts from numbers instead of opinions. Scout measures velocity and flags the tail. The decision to cut, and the buyer conversation that follows, stays yours.

The short version

• SKU rationalization is cutting low-velocity products to free shelf space, inventory budget, and focus for the SKUs that actually sell.
• Most lines follow an 80/20 split. In the example, the top five SKUs do 84% of volume and the bottom five do 6%.
• Use a velocity floor (units per store per week) to flip the default: below-floor SKUs go unless someone argues to keep them.
• Always check cannibalization first. A SKU is only safe to cut if its buyers trade up within your line rather than walking to a competitor.