As of November 8, 2026, the most useful number in retail availability data is not the stockout rate on any single day. It is the decay rate. A study covering 98 consumer brands, 2,833 products, 103,131 stores and 66 retail chains found that 0.8 percent of products in stock today are reported unavailable tomorrow, 3.4 percent are gone a week later, and 6.3 percent disappear within thirty days. For a brand planning the holiday window, that curve is the operating constraint, not the headline growth forecast.
What a 3.4 percent weekly decay actually measures
The study is built from what retailers publish about their own stores rather than from physical shelf audits, and that choice shapes how the number should be read. A listing that still exists in a chain's system but cannot be bought on the shelf is functionally absent, and the shopper experience is identical to a delisted item. Daily website level availability is the closest proxy a brand can get without sending people into stores with a clipboard.
The thirty day figure is the one that should set the review cadence. A monthly availability report arrives after most of the damage is done, because a product that disappears in the second week of December does not come back in January with the lost sales attached to it. Brands that already read the data weekly still learn about a third of the decay only after it has happened, which is why the useful question is not the average rate but where the failures concentrate.
The nearest store is the substitute, and it is often close
Substitution is usually modelled as a category behaviour, but this data suggests it is first a geography behaviour. When a store runs out of an item that another store in the same chain holds that day, the nearest such store is within five miles in 41 percent of cases, within ten miles in 62 percent and within twenty five miles in 79 percent. In most urban trade areas the alternative is not far away at all.
| Distance to the nearest same chain store holding the item | Share of out of stock cases |
|---|---|
| Within 5 miles | 41 percent |
| Within 10 miles | 62 percent |
| Within 25 miles | 79 percent |
That proximity cuts both ways. It raises the chance that the shopper keeps the category but loses the brand at that particular store, and it lowers the chance that the brand loses the shopper altogether. A nearby recovery is good for the retailer's basket and neutral to bad for the brand that was absent, which is exactly why availability failures rarely show up as a category level sales decline in the data a brand usually monitors.
What shoppers do before they leave the house
Consumer research points at the same gap from the other side. In a 2026 survey, 57 percent of shoppers said an item marked available online was out of stock when they reached the store, and 78 percent said they had tried to buy something in store only to find it sold out or nearly gone. Forty three percent said they would visit stores more often if availability simply held, and 22 percent now compare stock online specifically to avoid a wasted trip.
Replenishment interval separates a logistics failure from a demand spike
Replenishment interval is the diagnostic variable that most availability reports leave out. When availability drops in a store, comparing the failure dates against the delivery dates separates a logistics problem from a demand spike. If failures cluster in the two days just before a delivery, the store is running too lean for its own traffic and the fix is a parameter change rather than a larger order. If failures cluster right after a delivery, the order itself is too small for the rate of sale.
Concentration makes that check affordable. Within a chain, the ten percent of a brand's products with the most out of stock days account for 36 percent of all the out of stock days, and the ten percent of stores with the most account for 28 percent. A brand does not need to fix a network to move the number. It needs to fix a list, and the list is short enough to be reviewed product by product, store by store, in one meeting.
Dark store networks changed what coverage means
India has mapped 5,625 quick commerce stores across 408 cities
Coverage and availability are different facts and are routinely reported as one number. Coverage is a listing decision and moves slowly; availability is a daily operational fact and moves fast. Combining the two produces a figure that describes neither the assortment nor the shelf, and it hides precisely the stores where the gap is widening. The distinction matters more as delivery networks multiply the number of nodes a brand has to hold.
| Platform | Mapped stores in the directory |
|---|---|
| Blinkit | 1,955 |
| Zepto | 1,088 |
| Swiggy Instamart | 1,038 |
| Flipkart Minutes | 880 |
| BigBasket | 664 |
The mapped counts also show how unevenly the network sits. Bangalore lists 629 stores and Delhi 474 across five platforms, while whole regions run on a few dozen. For a brand, that means an availability target cannot be a single national number, because the same percentage implies a very different absolute exposure in a city with 474 nodes and in a city with 48.
A four week check a brand can run without new tooling
A four week check does not need new tooling if the brand already receives retailer availability feeds. The structure that works is weekly capture, a fixed comparison against delivery dates, one named owner per chain, and a decision rule agreed in advance. The rule is the part that usually goes missing. Without a threshold that automatically triggers a replenishment review, the report turns into a monthly argument about causes instead of a weekly list of fixes.
Who does what, and which evidence closes the loop
FAQ
What counts as on-shelf availability in this data?
It is the share of covered stores where the item could actually be bought on the check date. It is not distribution coverage, which only records that a listing exists somewhere in the retailer's system, and the two are often averaged together by mistake.
Is a 3.4 percent weekly decay rate high?
On its own it is a median across 96 brands in one market, with a 95 percent range of 3.0 to 3.9 percent. What matters for a brand is its own rate inside its top chains and whether the failures cluster before delivery dates.
Does the nearby store solve the problem?
Only partly. It protects the category and the retailer's basket, but the brand still loses that store's sale. Treat a nearby store as a reason the loss is invisible in category data, not as a reason the loss is small.
How often should the check run?
Weekly during peak season and monthly outside it. The decay between the seven day and thirty day marks is where most of the recoverable loss sits, and a monthly cadence cannot see it.
