On-shelf availability (OSA) in retail, explained

What on-shelf availability is

On-shelf availability (OSA) is the percentage of time a product is physically on the shelf and buyable by a shopper, measured at the store and SKU level. It is the number that decides whether the demand you worked so hard to forecast actually converts into a ring at the register. I have watched a SKU at Kroger show 96% warehouse in-stock and still lose sales every weekend, because the case was sitting in the back room, not on the shelf where a shopper could grab it.

That gap is the whole point of OSA. A product can be in the building and still be out of stock from the only perspective that matters, the shopper's. The classic case study is a 6 oz yogurt SKU at a large banner: warehouse said in-stock, the planogram said two facings, and the shelf was empty by Saturday afternoon every week. Nobody flagged it because the system the buyer looked at was green.

OSA is not warehouse in-stock

These two numbers get used interchangeably and they should not be. They measure different things at different points in the chain, and the spread between them is where lost sales hide.

Each layer leaks a little. The DC can be 98% in-stock while the store is 96% (a truck got skipped), and the store can be 96% while OSA is 92% (product is in the back room, not faced). Stack the leaks and a SKU that reads 98% at the warehouse is buyable only 92% of the time at shelf. That 6-point gap is not rounding. On a fast mover it is real volume walking out the door.

Phantom inventory and hidden voids

The worst version of an OSA problem is phantom inventory: the system believes there are units on the shelf, so it never triggers a reorder, but the shelf is empty. The perpetual inventory record says 8 units, reality says 0, and the replenishment engine sits quiet because it is trusting the record. These are also called hidden voids because they never show up as a formal out-of-stock alert.

Phantom inventory is poison for a demand forecast. Your POS shows zero sales for that SKU that week, the model reads it as collapsing demand and lowers the baseline, and now you are forecasting down on a product that was simply not on the shelf. The demand was there. The shelf was not. The model cannot tell the difference unless someone tells it the void existed.

The lost-sales dollar math

OSA is easy to wave away until you put a dollar figure on it. Take a single SKU, a 12 oz cold brew at $4.49 retail, selling 14 units per store per week when fully stocked, across 600 stores.

The arithmetic: 14 x 9% = 1.26 lost units per store per week, times 600 stores is 756 units, times $4.49 is about $3,394 a week, or roughly $176,500 a year on one SKU. Lift OSA from 91% to 96% and you recover most of that without spending a dollar on trade or media. This is why I always pushed OSA ahead of incremental promotions on the priority list. You are not buying new demand, you are stopping the leak on demand you already have.

Where Scout fits

Scout is a demand-side analytics layer. It connects your SPINS, Circana, or retailer POS data so you can spot the stores and weeks where velocity drops to zero against a baseline that says it should not, which is the signature of a void or phantom inventory, and size the lost-sales dollars the way the table above does. It is not a store-execution or retail-audit tool. It does not walk the aisle, scan the shelf, or fire a replenishment order. It surfaces the availability gap in the data so you know which stores to send the question to.

The short version

• On-shelf availability is the percentage of time a SKU is actually on the shelf and buyable, measured on the sales floor, not in the warehouse.
• Warehouse in-stock can read 98% while OSA is 92%. The gap between layers is where lost sales hide.
• Phantom inventory (the system thinks the shelf has stock when it is empty) suppresses reorders and corrupts the demand forecast.
• The lost-sales math is large: a 9-point OSA gap on one mid-velocity SKU ran roughly $176,500 a year in the worked example.