Shrinkage in supermarkets: where the margin goes and how to find it

What shrinkage is, and why it hides

Shrinkage is the gap between the stock your system says you should have and the stock you actually have, after allowing for sales. Because grocery margins are thin, even a small percentage of lost stock can wipe out a large share of profit. It is also hard to see, because the causes are spread across the store and many of them look like normal business.

Part of the confusion comes from mixing different kinds of loss. It helps to separate them.

• External theft. Shoplifting and organised theft.
• Internal theft. Staff taking stock or cash, or helping others to.
• Process and administrative errors. Wrong prices, missed scans, unrecorded damage, receiving mistakes and incorrect stock adjustments.
• Supplier issues. Short deliveries, wrong items and unreturned credits.
• Spoilage and expiry. Fresh food and short-dated products that cannot be sold.

Some of these are visible and recorded, such as a waste log. The dangerous ones are unrecorded, because they only show up as a mystery gap at the next stock count.

Measure first

Calculate shrinkage by store, department and category, and track it over time. Two measures make it manageable:

• The shrink rate. The value of unexplained stock loss as a share of sales.
• The recorded loss rate. Waste, damage and expiry that were logged, tracked separately so you can see whether recording is improving.

Move from annual counts to regular blind cycle counts, where counters do not see the system quantity. A little counting every week finds problems when they are small and fresh, and tells you which category or location is the source. Compare the results by store: a store with a persistently higher gap than its peers is a lead worth following.

Where the margin goes

At the front end

The till is where most internal and many process losses show up.

• Voids after the sale is nearly finished
• Refunds without a receipt or for items never sold
• Discounts and price overrides outside policy
• No-sale drawer openings
• Items passed over the scanner without scanning
• Till shortages that repeat on the same shift or cashier

At receiving and in the stockroom

• Short deliveries that are signed for without a count
• Goods received without a purchase order
• Unrecorded damage or breakage
• Transfers between stores that are sent but never received
• Stock adjustments without a reason or approval

On the shop floor

• Shelf prices that do not match the till price
• Expired or damaged items that are not removed or marked down in time
• Over-ordering of fresh products
• Display and packing losses

Controls that work at the till

The most effective controls are built into the sale itself.

• Role-based access. Cashiers, supervisors, managers and administrators should have different permissions.
• Supervisor approval. Voids, refunds, discounts and price overrides should require approval, with dual approval for high-value actions.
• Audit trails. Every override and master-data change should be logged with who, when and why.
• Cash drawer tracking. Record counts and variances at till close and track them by cashier and shift.
• Anomaly detection. Flag unusual refund spikes, void rates, discount use, till shortages and concentration of events around a single cashier.

The point is not to catch people out. It is to make normal behaviour easy and abnormal behaviour visible.

Controls in stock and receiving

• Three-way match. Pay suppliers only when the purchase order, the goods receipt and the invoice agree.
• Count at the dock. Blind receiving, where the receiver counts without seeing the order quantity, is more accurate.
• Batch and expiry discipline. Sell the oldest stock first and review short-dated lines early so they can be marked down before they are lost.
• In-transit tracking for transfers. Stock leaving one store and not yet received by another should be visible.
• Adjustment rules. Every stock adjustment needs a reason code and the right approval.

Use analytics to find patterns

Single events rarely tell you much. Patterns do. Useful exception views include:

• Voids and refunds per cashier per hundred transactions, compared with peers
• Refunds and voids at unusual times or just after tender
• Discount use by cashier and by item
• Till variances by shift
• Items with the biggest gap between expected and counted stock
• Stores whose shrink rate moves against the group trend

An AI agent can do this scanning continuously and propose a short list for a human to review. Keep a person in charge of the decision.

Culture and fairness

Controls work best when staff understand them. Explain the purpose, apply the rules evenly, protect honest staff from suspicion and give people a safe way to report concerns. Investigate patterns, speak to the person involved before drawing conclusions and document the outcome.

Quick wins

A 30-60-90 day plan

First 30 days. Establish the baseline: shrink rate by store and category, recorded versus unrecorded loss, and the top five exception patterns at the till.

Days 31 to 60. Turn on approval rules and audit logs, begin weekly blind counts in the worst categories and fix the obvious receiving gaps.

Days 61 to 90. Add automated exception reports or AI-driven flags, set targets by store and review results with store managers each month.

What investors should look at

Ask for shrinkage by store and category over time, how it is measured, and what controls are in place at the till and the dock. A chain that can show a falling trend and explain why usually has control of its operations. A chain that cannot measure it is likely losing more than it knows.