Retail Shrinkage Prevention Technology in 2025: Beyond the EAS Tag

Electronic article surveillance (EAS) was commercialized in the late 1960s. The core principle has not changed since: attach a resonant tag to merchandise, and a detection antenna at the exit sounds an alarm if the tag has not been deactivated at the point of sale. That is the technology most independent retailers are still using in 2025, and its limitations are also unchanged.

Shrinkage is the retail term for inventory loss from all sources: shoplifting, employee theft, administrative errors, and vendor fraud. The National Retail Federation's annual surveys have consistently put total retail shrinkage in the US at around 1.5 to 1.6 percent of sales. For a store doing $1.2 million in annual revenue, that is $18,000 to $20,000 in losses. EAS prevents some portion of opportunistic shoplifting, but it does nothing for mis-scans, sweethearting, or the growing category of skip-scan incidents at self-service checkout.

What EAS actually does and does not do

EAS is good at one thing: deterring impulse theft of tagged items by shoppers who know the system is there. It creates a visible signal that merchandise is protected. When the alarm triggers, it stops a theft in progress or at least creates a confrontation point.

It does not verify whether the correct items were scanned at checkout. It does not detect items that were never presented to any scanner. It does not capture the item type, only that a tag was or was not deactivated. When the alarm sounds, the cause could be anything from a legitimate purchase where the cashier forgot to deactivate the tag to a shopper who concealed an unscanned item. The false-positive rate at many busy stores is high enough that staff have been trained to treat the alarm as background noise rather than a reliable theft signal.

EAS also requires physical tagging of individual items, which carries its own cost. The labor to attach tags, the cost of the tags themselves (typically $0.07 to $0.20 per disposable soft label, more for hard tags), and the space and equipment required for the deactivation station at the register all add up. For high-volume SKUs like individual canned goods or packaged snacks, tagging every item is simply not economic.

Where camera-based exit verification fits

The structural difference with computer-vision exit verification is that it checks the transaction record against what is physically leaving the store. Rather than detecting a passive signal from a tag, it uses cameras at the exit to identify items and match them against the closed receipt. Items that appear in the camera feed but not in the transaction log become a reconciliation flag.

This approach has been deployed at scale by Amazon in its Just Walk Out stores, but the cost structure of their implementation is significant: ceiling-mounted camera grids, weight sensors on every shelf, and a dedicated infrastructure team. That version is not realistic for independent retailers.

The version that is becoming viable for small-format retail works differently. A single wide-angle camera at the exit zone, combined with barcode recognition at the cart-level or on the shopper's phone, creates a cross-check between what was scanned during the shop and what is visible at the point of departure. The camera does not need to identify every product by visual appearance. It identifies whether the number of items being carried appears consistent with what was scanned, and flags sessions where the gap exceeds a threshold for staff review.

Leav's loss prevention layer works within this model. When a shopper completes a mobile checkout session, the exit verification can compare the item count from the session against the visual item count at the door. If a shopper scanned three items and the exit camera detects five objects in their basket, that is flagged for a staff exception review. The staff member can verify in a few seconds without requiring a full item-by-item inspection. Most of the time it is a packaging discrepancy or a legitimate item in a personal bag. Occasionally it is something that warrants closer attention.

Skip-scan and the self-checkout shrinkage question

One of the most-cited criticisms of self-checkout is skip-scan, where a shopper intentionally places an item in the bagging area without scanning it. Traditional unattended kiosk self-checkout systems have documented skip-scan rates, with some retail loss-prevention studies estimating that 5 to 10 percent of self-checkout transactions contain at least one mis-scanned or unscanned item.

Mobile scan-and-go has a different risk profile than unattended kiosk checkout. In kiosk self-checkout, the shopper scans in a fixed location with a dedicated scanner, and the bagging weight sensor is the primary verification layer. In mobile scan-and-go, the shopper scans as they shop, using the phone camera, with a receipt generated before they reach the exit. The exit verification layer is what catches discrepancies.

We are not claiming mobile checkout eliminates shrinkage. It does not. What the data from our pilots shows is that the skip-scan rate in phone-camera-based checkout is lower than at traditional unattended kiosks, likely because the scan action is more deliberate when a shopper controls the camera angle and timing. And the combination of a transaction receipt plus exit image comparison creates two checkpoints where kiosk self-checkout often has only one (the bagging scale, which is easily defeated by weight spoofing or item weight similarity).

Cost comparison for independent retailers

The capital cost of a full EAS installation for a small store, including antenna panels, a deactivation pad, and a basic tag inventory, typically runs $3,000 to $8,000 depending on the store size and system quality. Ongoing costs include replacement tags and occasional service calls when the antenna triggers false positives due to nearby interference.

Camera-based exit verification at the level that integrates with a mobile checkout system is a different cost model: it runs on existing CCTV infrastructure if the camera resolution and placement are adequate, or requires a single camera installation costing $200 to $600. The analysis layer is software, so the marginal cost per additional camera or additional transaction is low.

Neither approach is a complete answer to shrinkage. The realistic picture is that EAS deters some opportunistic theft, camera verification catches some exit discrepancies, and neither system stops a determined and prepared shoplifter. The combination of mobile checkout with exit verification addresses a different segment of the shrinkage problem than EAS alone does, specifically the transaction-integrity gap at the point of payment.

What this means for stores evaluating both

Stores that already have EAS installed are not in a position where mobile checkout makes their existing investment obsolete. EAS continues to work on tagged merchandise. What camera-based transaction verification adds is coverage of the checkout interaction itself, which EAS does not touch.

Stores evaluating their first investment in loss prevention technology face a different question. A camera-based system paired with mobile checkout has a lower upfront cost than EAS, covers the transaction integrity layer that EAS misses, and does not require individual item tagging. For stores with high SKU counts, mixed item sizes, and limited staff bandwidth, that combination may be the more practical starting point in 2025.

The 30-year stability of EAS is partly a testament to its simplicity and partly a sign that nothing better suited to independent retail economics has been available until recently. The software-defined alternative is now mature enough to take seriously.