Halloween is the one night of the year when ghosts are welcome. In a cooler fleet, they turn up all year round. Every operator tracks shrinkage. Almost none track phantom inventory. Yet the research suggests the loss that goes uncounted may be the larger of the two. One shows up the moment the cash drawer fails to match the shelf. The other never appears on a loss report at all, because a sale that never happened leaves no trace. This guide looks at what each loss actually costs, why the two share one root cause, and how a cooler that counts its own contents addresses both.

The loss everyone counts

Shrinkage is the gap between the stock a system expects and what is physically on the shelf. It runs about 1.6% of sales for U.S. retailers (National Retail Federation).

The NRF’s 2023 National Retail Security Survey put it at $112.1 billion in fiscal 2022, the last edition before the NRF stopped publishing the annual shrink report. Theft takes most of the blame, yet Retail Dive’s reporting on shrink data found that much of the loss traces back to inventory control: miscounts, receiving errors, and stock that was never tracked correctly.

Either way, shrinkage is visible. The cash says one thing, the shelf says another, and the difference has a name.

The loss almost no one counts

Phantom inventory is stock a system records as available that is not actually on the shelf. It is the leading cause of out-of-stocks, which cost U.S. retailers an estimated $82 billion in lost sales in 2021 alone (NielsenIQ).

Retail also calls it ghost inventory, and the name fits: it haunts the records without ever appearing on the shelf. The system shows product on hand; the shelf is empty. Because the record says there is stock, the automatic reorder never fires. The shopper arrives, finds nothing, and leaves, often for a competitor. No alarm sounds, no variance is logged, and the lost sale never enters a loss report, because from the system’s point of view nothing was lost at all.

The scale is striking once measured against shrinkage. Research by Corsten, Gruen and Bharadwaj established that out-of-stocks cost retailers roughly 4% of sales, more than double the headline shrinkage rate, and NielsenIQ estimated that out-of-stock and out-of-shelf items cost U.S. retailers $82 billion in 2021 alone. Underneath it sits a sobering statistic: the average U.S. retailer operates at only about 65% inventory accuracy, according to Auburn University’s RFID Lab. For every hundred records, roughly a third do not match the shelf.

Why they are the same problem

Shrinkage and phantom inventory are connected at the root: shrinkage corrupts the inventory records that drive reordering, and corrupted records produce stockouts. Theft and the empty shelf feed each other through the same bad data.

Treating the two as separate line items is the central mistake. When product walks out unrecorded, the system still believes it is on the shelf, that is phantom inventory, created directly by shrinkage. The reorder does not trigger, the gap goes unfilled, and a second loss stacks on top of the first. The retailer pays once for the missing product and again for the sale it can no longer make. The link is not theory; it is the mechanism inventory-accuracy research keeps surfacing.

Read across the table and the pattern is clear: the two columns describe one chain of events, not two problems. The common thread is the distance between what the system believes and what is physically on the shelf.

Where smart coolers change the math

Smart coolers, freezers and cabinets put a door and a payment step between the shopper and the product, and they can measure what is inside them. Both losses respond to that.

Operators are already moving this way. In Automatic Merchandiser’s 2026 State of the Vending and Micro Market Industry report, smart coolers made up 33.5% of the equipment respondents deployed in 2025. A closed, payment-gated format ties every item that leaves to a transaction, which takes most unrecorded removal off the table.

The inventory side matters just as much. Many vending-management systems still work from telemetry: transaction logs and mechanical counts that estimate what is left. A route driver restocks on that estimate. When it drifts, the truck arrives with the wrong product in the wrong quantity, and the empty slot stays empty until the next visit. That drift is where phantom inventory lives, and a cooler that weighs its own contents removes it.

How SHEKEL closes the gap

Both losses shrink when the system’s record matches what is physically inside the cooler. WeightAI™ does that job inside smart coolers, freezers and cabinets.

• Every item tied to a payment. The door opens for a shopper who has started a transaction, and weight sensors register each product the moment it is lifted, so what leaves is what gets charged. SHEKEL customers run shrinkage below 1% (SHEKEL figures).
• Live inventory, item by item. Each shelf inside the cooler reports its contents in real time, so the stock level on the dashboard is a measurement, with no telemetry estimate in between. Phantom stock has nowhere to hide.
• Routes planned from real counts. Drivers load the truck with what each cooler actually needs, best sellers get reordered before they run out, and fewer trips go to locations that are still full.
• Built into the coolers operators already trust. Manufacturers integrate WeightAI™ into their own smart coolers, freezers and cabinets, so operators keep their equipment partner and gain 99.99% uptime.

The takeaway

Shrinkage earns the attention because it is the loss with a number attached. But the research suggests the larger drag may be the loss that never shows up the sale foregone because the system thought a shelf was stocked when it was empty. The two are not rivals for the loss-prevention budget. They are the same failure measured at two different moments. Operators who close the gap between what their system reports and what is physically in the cooler reduce both at once. In a smart cooler that measures its own contents, the record and the shelf become the same number. Leave the ghosts to Halloween.

THE INTELLIGENCE LAYER

WeightAI™ turns smart coolers, freezers and cabinets into self-counting stores: every item weighed, every sale recorded, every restock planned from real counts. Shrinkage below 1%, live inventory, and routes built on what is actually inside the machine. See how it works at shekel.ai.

Frequently asked questions

Is shrinkage the same as theft?

No. Theft is one cause of shrinkage, alongside internal loss and administrative error. Industry analyses note that theft is often blamed for losses that are really inventory-control failures, so treating shrinkage as a theft problem alone can miss the larger share.

What is a good retail shrinkage rate?

The NRF’s 2023 National Retail Security Survey put the U.S. average at 1.6% of sales for fiscal 2022, with a median of 1.4% (NRF via Statista). Acceptable rates vary widely by category, so the most useful benchmark is a location’s own rate tracked over time.

What is retail shortage control?

Retail shortage control is the practice of keeping the inventory a system reports aligned with what is physically on the shelf, so that losses are caught and reorders fire correctly. It combines loss prevention, accurate counting, and real-time shelf visibility.

How does phantom inventory cause stockouts?

When records show stock that is not physically present, the automatic reorder never triggers. The shelf stays empty, the shopper finds nothing, and the sale is lost — without the loss ever appearing on a report.

SOURCES

1. National Retail Federation — National Retail Security Survey 2023 (U.S. average shrink 1.6% of sales; $112.1B in fiscal 2022). NRF has since stopped publishing the annual shrink report (Retail Dive).
2. Statista — Median inventory shrinkage rate of U.S. retailers (1.4% of sales in fiscal 2022; NRF data).
3. Corsten, D., Gruen, T. & Bharadwaj, S. — Retail Out-of-Stocks: A Worldwide Examination of Extent, Causes and Consumer Responses (GMA / FMI / CIES, 2002); see also Corsten & Gruen, International Journal of Retail & Distribution Management, 2003 (~4% of sales lost to out-of-stocks).
4. NielsenIQ — On-Shelf Availability (OSA) Barometer for U.S. retailers, February 2022 (out-of-stocks cost U.S. retailers $82B in 2021).
5. Auburn University RFID Lab — ~65% average retail inventory accuracy, as cited in Zebra Technologies, Inventory Visibility white paper.
6. Retail Dive — “Retailers have a crime problem. It’s in the numbers.” and “NRF updates crime report to address faulty numbers” (2023), on inventory-control errors being counted as theft.
7. Automatic Merchandiser — “State of the industry 2026: Smart coolers move mainstream as revenue tops $40 billion” (smart coolers 33.5% of equipment deployed by respondents, 2025 data).

Note: U.S. cost figures are drawn from the most recent published surveys and industry estimates; the NRF stopped publishing its annual shrink report after the 2023 edition, so headline shrink totals reflect fiscal 2022 data and are presented as baselines, not current readings.