In this guide
What Counts as Normal Shrinkage
There is no single normal shrinkage figure for Indian retail, and any number quoted as one should be treated with suspicion. Shrinkage varies by format and by category to a degree that swamps the average: a fresh food operation loses stock to spoilage in ways an electronics store does not, a self-service format is exposed differently from a counter-service one, and high-value small items behave differently from bulky low-value ones in the same store. A single benchmark misleads because it invites a store to be judged against a mix it does not carry, and because published figures are drawn from populations that may share nothing with your estate. What an auditor actually compares against is your own history for that store, the same category across your other stores, and the movement in the figure rather than its level. A store whose shrinkage is stable and explicable is in better shape than one below an industry average and moving.
How Shrinkage Is Measured
Before any figure can be compared, the basis has to be established, and the two common bases produce different numbers from identical events. Shrinkage as a percentage of sales is the retail convention and it is what most published figures use. Shrinkage as a percentage of stock held describes the loss against the inventory at risk and produces a much larger number in a business that turns its stock quickly. Neither is wrong and they are not interchangeable. Known against unknown loss is the second distinction. Known loss is recorded as it happens: wastage logged, damage written off, markdowns applied. Unknown loss is the residue discovered at a count, which cannot be attributed to any particular event. A business with excellent recording will report a high known loss and a small unknown one; a business that records nothing will report all of it as unknown. Why two retailers report different numbers for the same event follows from both. One counts wastage as shrinkage and the other treats it as a cost of trading; one measures against sales and the other against stock; and the two figures diverge by a factor before anybody has lost anything differently.
What Drives It by Format
Loss patterns are properties of the format and the category rather than of the operator, which is why comparing across them is unhelpful. A supermarket carrying fresh produce, dairy and bakery loses stock continuously to spoilage as an inherent feature of selling perishable goods, and a substantial part of what it records as shrinkage is a trading cost rather than a control failure. An apparel retailer loses very little to spoilage and considerably more to markdown and to theft of small high-value items. An electronics retailer loses least in quantity and most in value per incident. High-value low-volume categories are the ones that dominate the value of loss almost everywhere, which is why the shrinkage figure and the shrinkage count point at different places. A handful of missing items can outweigh a great deal of spoiled produce. Self-service is the format variable with the largest single effect. Where customers handle goods directly and unsupervised, the opportunity for loss rises sharply compared with counter service, and formats that moved to self-service to reduce staffing costs generally accepted a higher shrinkage rate as part of that trade.
Separating Loss From Bad Records
A shrinkage figure is a residual, and residuals collect everything nobody accounted for, so the first question about any shrinkage number is how much of it is actually loss. Posting errors that look like shrinkage are the largest contaminant. Goods received and not booked, transfers recorded at one end, markdowns not entered, and returns processed against the wrong line all produce a difference between records and shelf that is identical in appearance to theft. None of it involves anything leaving the building. Receiving and transfer discipline is therefore examined before any conclusion about loss is reached, because a store with weak discipline at either point will report shrinkage it never suffered. Negative stock is the clearest records symptom available and it is worth watching for that reason. A system balance below zero is arithmetically impossible as a physical state, so wherever it appears there is a definite recording error, and its frequency is a direct measure of how much of the store's reported variance is records rather than loss. Zeroing negative balances without tracing them guarantees the same error recurs and the shrinkage figure stays uninterpretable.
What an Auditor Accepts as Explanation
Shrinkage is explained by documents or it is not explained. Documented wastage is the first and largest legitimate component in perishable formats: goods logged as damaged, spoiled or out of date at the moment it happened, with the quantity and the reason recorded by the person who found them. Wastage logged at the end of a shift in round quantities is weaker evidence than the same wastage logged as it occurred, and the difference between the two is visible in the data. Mark-downs are the second component, and they explain a value gap rather than a quantity gap, so a shrinkage figure computed at retail value without adjusting for markdowns will overstate loss substantially in any format that discounts. Approved write-offs are the third, evidenced by the authorisation rather than by the entry. What remains after all three is the unexplained gap, and that is the figure the finding attaches to. A store with high total shrinkage and a complete explanation is in better order than one with lower shrinkage and nothing supporting any of it.
Using a Benchmark Without Being Misled
Compare against your own history first. Your figure for the same store last year, and for the same category across your other stores, are the only comparisons drawn from a population that resembles yours. An external benchmark is drawn from a mix of formats, categories and markets you cannot see, and matching it proves nothing while missing it may only mean your range is different. Use published figures for orientation, never as a target. Separate the store-level and chain-level questions, because they have different answers. A chain-level figure tells you whether the operating model is working; a store-level figure tells you whether one site is. A chain performing well overall can carry stores that are not, and an average conceals exactly the variation worth acting on. Look at the distribution rather than the mean. An independent count settles the argument where a store disputes its figure, where the internal counting method is itself in question, or where the shrinkage number will be reported externally, and retail stock audit work provides the consistent basis that makes stores genuinely comparable.
