Insights · Loss Prevention

Shrinkage assumed to be normal

Distinguishing between internal, external and process-related losses completely changes the loss prevention plan. And it is worth looking at the absolute figures, not just the percentage of sales.

At the June committee meeting, the operations director presented the figure with the calm of someone delivering a routine report: shrinkage stood at 1.8 per cent, as always. No one asked any questions. 1.8 per cent was this year's figure, last year's and the year before's, and through sheer repetition it had stopped being a statistic and become a hallmark of the company, almost a tradition. Until someone from finance worked the figure out in euros and dropped it on the table: that usual percentage came to six hundred thousand euros. And as sales that year had been higher than ever, the usual shrinkage was also the most expensive in the company's history.

The percentage of sales is a good thermometer and a guide that does not tell the whole story. It serves to compare you with the sector and little else, because inside that 1.8 per cent there is room for three different companies. One where goods are stolen from outside, another where they leak from within, and another where nobody steals anything and yet the stock vanishes on paper. Stock counted wrong, prices entered wrong, breakages that go unrecorded, returns that nobody reconciles. All three show the same figure on the shrinkage line, and all three need completely different plans. That is why the first question is not how much are you losing, but what are you actually losing.

Separating internal, external and process loss does not call for a mammoth project. It calls for looking at the data that already exists with the right question in mind, and accepting from the start that part of the answer will not be popular. Internal loss is uncomfortable because it forces you to talk about controls and trusted people, and process loss is uncomfortable because it spreads responsibility across departments that saw themselves as bystanders. I remember a retail chain that discovered, in its own data, a pattern of cash refunds concentrated in a few tills and a few shifts. They had had all the information in front of them for months. They just needed to look at it.

The online channel deserves a paragraph of its own, because it tends to grow faster than its controls. Claims for orders that never arrived, returns that come back incomplete or with something else inside, and the abuse of refund policies are no longer the preserve of habitual offenders. Cheating a multinational has become an everyday temptation that thrives among groups of friends and family. It is a game of let's see who's the cleverest, and it feeds on the care many companies take of their customers and their reputation.

If the physical shop has stock counts, till reconciliations, CCTV and process reviews, the online basket cannot live without its equivalents. And how do you do all this without growing the team? Pick a few battles and hold them. A loss prevention plan that works can and should fit on a few pages. The three or four indicators that a named person checks every week, the thresholds that trigger a review and the specific action that follows each alarm. The rest is consistency. We come from years of security and loss prevention in retail, and the difference between the organisations that reduce their shrinkage and the ones that absorb it is almost never in the technology. It is in recognising that loss can be reduced, not accepting it as part of the business plan. It is in whether someone has the brief, the time and the backing to look. And in good software that tells you about the anomalies in seconds.

What is your shrinkage made of? If the answer is a percentage, you still do not know. Turn it into euros, split it into its three families and ask yourself which part you have been calling normal for years. In losses, normal is simply what has not yet been looked at closely.

If this resonates with something you are seeing in your organisation, let's talk.

Let's talk