The Store Has It. The System Says It Doesn't.

When people talk about inventory records going wrong, they usually picture a familiar problem: the system promises a product, someone looks for it, and the product is not there.

That discrepancy matters. But it has a counterpart that gets far less attention.

Sometimes a store has more physical stock than its system records. The units are present. The recorded quantity is too low. In some cases, the system may even say zero.

It is tempting to call that the harmless kind of inventory error. At least the product is not missing.

But the record does more than describe the store. It helps the store decide what to do next.

An error that can create the next order

In February 2026, ECR Retail Loss published Defining and assessing inventory record inaccuracy (IRI) metrics, research by Christoph Glock, Yacine Rekik and Aris Syntetos. The paper explains how inaccurate records can affect decisions in both directions. When a record overstates what is physically available, replenishment can be delayed. When it understates what is there, ordering can happen too early. [1]

Consider the second case. A replenishment process sees a recorded quantity below its ordering threshold. It has no way to infer, from that number alone, that more units are already in the building. Depending on the retailer's processes, another order can follow.

This is not an automatic result of every inventory mismatch. But when it happens, the consequences extend beyond an inaccurate count. Capital can be tied up in unnecessary inventory. Receiving and storage work can increase. A product that should have been available for sale may be treated as unavailable if a customer-facing system relies on the same incorrect record.

The product exists. The information needed to make the right decision does not match it.

Why one reassuring number can be misleading

The ECR paper makes a distinction every retail leader should hold onto: different inventory metrics answer different questions.

As the paper recommends defining it, shrink counts missing units, not unrecorded overages. ECR also notes that retailers do not all use the word shrink consistently. Its recommendation is to keep loss reporting separate from measures that cover errors in both directions. [1]

A net inventory difference creates another problem. Large shortages in some product records and large overages in others can offset one another, leaving a total that looks reassuringly small.

The paper demonstrates this with a worked illustration, not a measurement of a real store. Across 12 example products, 11 have an incorrect record. That is about 92% of the records in the example, even though the net difference across them is close to zero. [1]

Put those two numbers side by side. A nearly balanced total does not tell you whether the underlying product records are reliable.

For replenishment decisions, the researchers recommend paying attention to the direction of error, particularly at the individual-product level. They also warn that signed differences can cancel out and suggest pairing direction-aware measures with another view of how widespread inaccuracies are. [1]

That is a more useful question for a retail leader than asking whether one headline percentage looks good.

Inventory truth has to follow the physical store

At Aethreallegence, we use the term Inventory Truth Decay (ITD) to describe the continuous process by which inventory records drift away from what is physically present as store conditions change.

It is easy to picture that separation in one direction: records show items that cannot be found. The opposite direction belongs in the same conversation. The physical inventory is there, but the system understates it.

The ECR study is about measuring these discrepancies, not identifying every root cause. Its lesson is straightforward: a business needs measures suited to the decisions it is making. A shrink number, a net difference and an item-level count are not interchangeable.

When the store and the record disagree, treating the record as complete truth can create avoidable uncertainty for employees, replenishment teams and customers. In the worst case, the store pays twice: once for the units it already holds, and again for the ones it reordered because the record could not see them.

The store has it. The system says it doesn't. And the next decision may be made using the wrong answer.

What inventory is your store already holding that its own records are failing to recognize?

---

Sources

1. Glock, C. H., Rekik, Y. and Syntetos, A. A. Defining and assessing inventory record inaccuracy (IRI) metrics. ECR Retail Loss, February 9, 2026. ecrloss.com