Summary
- According to an NRF survey (177 brands, more than 97,000 stores), shoplifting caused 36% of US store losses, employee theft 29%, and process errors 27%.
- Together that came to 1.6% of sales, or $112.1 billion, in fiscal 2022.
- From 2019 to 2023, the number of theft incidents at the average US store grew by 93%, and the UK recorded more than 20 million shop thefts in a year.
- Store footage is usually watched after a stock count; analytics on the same cameras can flag moments to check as they happen, once it is configured for the store and tested in shadow mode first.
For: owners of stores and retail chains, heads of security and loss prevention
What losses are made of
The US National Retail Federation (NRF) surveys retail chains about losses every year. The 2023 report covered 177 brands, together more than 97,000 stores and $1.6 trillion in sales. Losses came to 1.6% of sales, or $112.1 billion.
| Source of loss | Share |
|---|---|
| Shoplifting by customers | 36% |
| Employee theft | 29% |
| Process errors: receiving, inventory records, price labels | 27% |
| Other | 8% |
Behind the second row are the people trusted with the keys, the till, and the stockroom: almost a third of losses. In a store with a meat counter or its own cutting, the risk here is higher, because the goods are expensive, and the same staff weigh them and write them off.
Theft is increasing
A separate NRF study from 2024 covered 164 brands, about 30% of the US retail market. From 2019 to 2023, the number of theft incidents at the average retailer grew by 93%, and the money lost to them by 90%.
In the UK, the British Retail Consortium (BRC) counted more than 20 million thefts in a year, over 55,000 a day, worth £2.2 billion. Incidents of violence or abuse against staff rose to more than 2,000 a day, compared with 1,300 a year earlier.
We did not find Ukrainian data on this scale in the public domain. The schemes themselves, from an item in a pocket to one "forgotten" at the bottom of the cart, are unlikely to differ much.
Why the cameras on the floor help so little
In most stores, cameras hang over the tills, on the sales floor, and in the stockroom. The footage is reviewed after a stock count has already found a shortage. Then you need to know which day and which hour to watch, and usually nobody knows, so even well-placed cameras end up as an archive.
Video analytics can be configured to read the same cameras as the footage comes in and flag the moments worth checking: a tray in the display is empty and has not been restocked; a queue has formed at the till; someone enters the back room outside working hours. Each such moment becomes an event card with a frame and a time, and the store manager or the security team makes the decision. We configure these scenarios for the specific store and run them in shadow mode first: the system collects event cards, and you see how many of them are real before any go to staff.
We deliberately give no "how much this will save" figures here. It depends on the store, and an honest estimate is only possible after a few weeks of running on your cameras.
Where to start
- Look at the results of your last stock counts: which goods go missing most often, and in which departments.
- Check whether the cameras see exactly these places: the counter, the scales, the till, the stockroom entrance.
- Decide what matters most to you: theft, empty shelves, or queues.
- Ask your vendor whether you can start in shadow mode on your own cameras, and who in the store will receive the event cards.
We connect to the cameras already installed in the store and start in shadow mode.