35 percent fewer out-of-stocks and 28 percent shorter gaps on the shelf are the two numbers from a 60-day pilot DeCicco and Sons ran with Simbe. In the same week Instacart began rolling out a tool that prices every empty facing at store level and at UPC level. Read together, they point at one shift: on-shelf availability is moving from a weekly percentage to a daily position on a specific aisle. For a brand team, the measure now has a location, and that location can be bought, audited or disputed, a reading that holds through November 10, 2026.
Two vendors, one measurement shift
Both vendors sell the same idea in different units. The Simbe release frames the gain as time: less time with a gap on the shelf, fewer hours spent hunting for problems. Instacart frames it as money: an estimate of lost sales per UPC per day, plus a substitution path that shows whether a shopper moved to a competitor SKU or simply took another pack size.
The difference matters because the two units drive different decisions. A time-based number sends a store team to an aisle. A money-based number sends a field sales team to a buyer meeting. Retailers rarely buy both at once, so vendors are competing over which number becomes the shared language between store operations and the brand.
Where distribution coverage ends
Distribution coverage tells a brand how many doors carry the item. On-shelf availability tells it whether the item can actually be bought in those doors. The two are frequently reported as one figure, which is how a brand ends up with 95 percent distribution and a shelf that is empty a fifth of the time.
What a shelf scan still cannot see
A scan reports whether a position is occupied. It does not separate a missing carton in the back room from a delisting, and it does not know whether the gap was created by demand or by a planogram change that moved the item two aisles away. Those three causes need three different fixes, and only the first is a replenishment task.
| What the system reports | Unit of measure | Refresh | Known limit |
|---|---|---|---|
| Empty facing | Aisle position | Daily | Cannot see back room stock |
| Lost sales estimate | Store and UPC | Daily | Modelled, not invoiced |
| Substitution path | Shopper basket | Daily where orders exist | Thin in low-order stores |
The refresh rate is where most pilots fail. A daily estimate is useful for a replenishment cycle that runs twice a week. It is not useful for a chain that already replenishes fresh categories every morning, because the gap has been closed before the report lands.
Substitution is the number most reports skip
A stockout rate says how often an item was missing. It does not say where the demand went. If the shopper bought another pack size of the same brand, the loss is smaller than the rate suggests. If the shopper walked to a competitor facing, the loss compounds, because a new habit can outlast the refill.
Linking the two requires order history rather than shelf imagery. That is why the platforms with a delivery network have an advantage in this category: they already hold the basket data that turns an empty facing into a substitution path.
Sampling is not even
Visit frequency per store is the sampling frame, and it is not uniform. Dense urban stores generate far more passes than a suburban store with a thin order book, so a national availability figure can reflect the store mix of the sample as much as the state of the shelf. Before carrying a number into a buyer meeting, a brand should check whether the sample matches its own distribution footprint.
How to test the data before signing
Run the tool against a store list the brand already audits by hand for four weeks. Compare three things: the stores flagged as having a gap, the size of the gap in facings, and the substitution path. If the vendor and the manual audit disagree on fewer than one store in ten, the data is probably good enough to negotiate with.
The commercial question is usually left until last: who pays for continuous visibility, the retailer or the brand. Vendors rarely publish that split, and it decides whether the data becomes a shared operating metric or another report that only one side can afford.
FAQ
Is an empty facing the same as an out of stock?
No. An empty facing can be back room stock, a delisting, or a planogram move. Each has a different fix.
Why does the unit of measure matter?
A time-based number drives store labour. A money-based number drives field sales. Vendors compete over which becomes the shared metric.
Can a national availability figure be trusted?
Only if the store sample matches the brand distribution. Dense urban stores are over-sampled in most networks.
What should a pilot measure?
Agreement with a manual audit on flagged stores, facing counts and substitution paths over at least four weeks.
