33% of your shrink isn’t theft, and that money now sits on the receiving dock instead of the front door. Shrink is the trade word for stock that is lost, stolen or damaged before it can be sold.

The study behind that number covers 1,421 retailers in 12 countries [1][2]. It puts outside theft at 25% of total loss, so the other 75% lands in other parts of the business [1]. The largest block beyond the front door is the supply chain. Supplier fraud, warehouse shrink and goods-receiving loss add up to 33% of everything those retailers lose [1].

In the same month, the market selling software for this problem gave its own figure. Spending on order systems is set to climb from $2.2 billion in 2024 to $4.3 billion by 2029 [3]. Cloud software takes $4.0 billion of that total, while the systems a retailer owns, and the cost of keeping them up, are cut by more than half as cloud takes over [3].

The big picture:

The firm behind the study looked at five places a retailer loses money: store theft, the supply chain, returns, online orders and the paperwork behind all of it [1]. My read is that the new name matters more than the percentages. A loss ledger decides who owns the fix, and while the ledger was a store ledger, the leader who runs receiving had no line on it. Now that leader has two lines. Dock loss sits inside the 33% block. It is the same physical event as a shipment that fails the in-full test: a case left a supplier’s dock and never landed on the retailer’s books.

Only the unit differs. The loss team counts it in dollars, and the order desk counts it as a share of orders that turns into a fine once it passes a set line [4]. The quantity half of that score has its own trail, which this desk has followed before in Your Hidden Quantity Gap.

The money runs the other way. Cloud order software rises from $1.4 billion in 2024 to $4.0 billion in 2029. Over the same span owned software falls from $416 million to $180 million, and its upkeep from $358 million to $170 million [3]. Every line of that forecast is a choice about where the order record lives.

Put the two papers side by side and a point appears that neither one makes on its own. The layer that makes a promise is being paid for. The layer that holds the proof of what arrived is being cut. A promise is a sum of what the system thinks it can ship. A receipt is a scan, a shipping document and a name. It is the only thing that settles a short case after the fact [4].

By the numbers

  • 25% — Theft share: outside theft, organized retail crime included, is the largest single source of loss, and 75% of lost dollars occur in other parts of the business [1].
  • 33% — Supply chain block: supplier fraud, warehouse and sales-floor shrink, and goods-receiving loss make up a third of retailers’ total losses [1].
  • $103.40 — Cost per paperwork error: one such error costs $103.40 on average, close to the $107.00 typical cost of staff theft [1].
  • $2.2bn to $4.3bn — Order software: forecast spend on order systems by 2029, with $4.0 billion of that total arriving as cloud software [3].

What I’d watch:

The people closest to this are not arguing about the study. They are arguing about a receiving record, and those fights carry price tags.

  • The budget line that sorts winners: the typical loss prevention budget in the study was about 1.2% of sales. The only group reporting more declines than increases was the group spending above 2% of sales [1]. That is a budget gap more than a software gap, and I would frame it that way in any pitch for a platform.
  • The fine that follows a short case: one published summary of the retailer rules has Walmart setting goals of 90% on-time for prepaid freight, 98% for the freight it collects and 95% in-full. The fine is 3% of the goods’ cost for cases that miss [4]. The same summary has Kroger charging 3% of the invoiced amount on a late order it blames on the supplier, and waiving late fees for a supplier that holds 93% on-time across four weeks [4].
  • The clock on the claim: Kroger allows 180 days from the invoice date for a billing question, while Walmart’s window depends on the contract and on how the fine was processed [4]. A record that takes a quarter to match up can outlive the window that would have reversed the charge.
  • Where the record moves to: with owned order software and its upkeep both falling, the file that settles a receiving claim sits more and more inside a cloud service someone else runs [3]. What a leader keeps, in that setup, is the export.

The catch

The study measures Total Retail Loss, and its own method leaves spoiled goods and extra stock out by agreed scope so the benchmark holds steady [2]. So the 33% is a share of lost dollars, never a rate of failure. It is not a fill rate, and not a stockout rate, and it is not a figure to set against an on-time number in the same sentence. The study also does not split its three supply-chain parts. So I cannot tell you how large goods-receiving loss is on its own [1]. What I can say is that it is named as one of them, which is more than most service dashboards manage.

One wrinkle is worth saying plainly. The study’s own page describes 1,421 retailers in 12 countries, while the wire version of the same release counts a smaller group. I have used the study’s own method statement.

The $4.3 billion is a forecast for 2029, and it covers software spend rather than a market for fixing loss. Nothing in either paper says the buyers are wrong, or that an order platform does not improve a promise. My read is narrower. A better promise does not by itself produce a record of what arrived, and only one of those two things settles an argument at the dock.

At a glance

  • The Big Shift: A study of 1,421 retailers finds outside theft at 25% of total loss. A 33% supply-chain block, dock loss included, sits beyond the front door, while order software spend is set to almost double by 2029.
  • Why It Matters: The order desk’s service share and the loss team’s dollar ledger describe the same failed shipments. A promise that improves without a receipt that improves leaves the loss where it was.
  • What I’d Watch: the signals that decide which side of this gets paid for.
  • The 2% budget line: loss prevention spending above 2% of sales was the only group reporting net declines, against an average budget of 1.2%.
  • The 3% fine: retailer fines land on the case that misses, which makes the receiving record the cheapest piece of proof in a claim.
  • The cloud shift: cloud order software rising to $4.0 billion by 2029 while owned software and upkeep are cut means the record sits in someone else’s system.
  • The Catch: Total Retail Loss leaves spoiled goods and extra stock out by scope, so its figures are not stockout or fill rates. The 2029 spend number is a forecast for software rather than a market for the fix.

Related reading

Sources

[1] IHL Group and Sensormatic Solutions, Beyond Shrink: The 2026 Global Total Retail Loss Study, release 22 October 2026 — https://www.sensormatic.com/resources/pr/2026/beyond-shrink-report-press-release [2] IHL Group, Beyond Shrink: The 2026 Global Total Retail Loss Study product page and method statement — https://www.ihlservices.com/product/beyond-shrink-2026-global-total-retail-loss-wp102026/ [3] IHL Group, 2026 Order Management Systems Market Study, October 2026 — https://www.ihlservices.com/product/2026-order-management-systems-market-study/ [4] SPS Commerce, “Retailer Deductions in Q4: How Walmart and Kroger Compare”, last updated 13 October 2026 (the publisher describes its own cost range as an expert estimate rather than an independent study) — https://www.spscommerce.com/community/articles/retailer-deductions-in-q4-how-walmart-and-kroger-compare