Most restaurants that "lose money somewhere" are not being robbed in one dramatic way. They are leaking in eight or nine small ways at once, each too small to notice in a month-end P&L, and each leaving a trace in records the restaurant already keeps. This guide names the ten places it happens, how to find each from your own records, and what actually stops it.
Restaurant leakage is money that leaves a restaurant without appearing as a legitimate, recorded cost — stock used but not sold, goods paid for but never received, expenses paid without a bill, or cash that never reaches the till. It is the gap between what the records say should have happened and what physically did.
The important word is gap. Leakage is not a line item you can look up; it is a difference between two things that should agree and don't. That is why it survives in restaurants with perfectly good bookkeeping: the accounts add up internally, and nothing in them is being compared to the shelf.
Stock leaves the building without a sale behind it. Rarely dramatic — usually small and regular, and concentrated in items that are valuable, portable and hard to count precisely: oil, cheese, dry fruit, liquor, meat, gas cylinders.
The recipe says 150 g of paneer; the kitchen serves 180 g when it is busy. Nobody is stealing anything, and the dish is sold at the same price. Twenty per cent over-portioning on a dish that sells two hundred plates a week is a serious cost with no villain in it.
Spoilage, over-prep, trimming loss, staff meals, comped dishes and mistakes. Wastage is not leakage if it is recorded — it becomes leakage the moment it happens quietly, because then it is indistinguishable from theft in the numbers.
The rate you agreed drifts. Onions that were ₹28 arrive at ₹35, then ₹40, and because each delivery is small the change is never argued. Price drift is the most common leak in Indian restaurants and the least likely to be noticed, because every individual bill looks normal.
Quantity billed exceeds quantity delivered; a payment goes out twice against the same bill; an old bill is settled again months later. Vendor ledgers are usually reconciled from the vendor's side, which is exactly the wrong direction.
Staff buy directly from a shop and hand over a slip. Sometimes genuine, sometimes inflated, occasionally invented. Without an approval step, "urgent purchase" becomes an unmonitored spending channel.
Cash paid for a repair, a delivery boy, a gas refill — no bill, no entry, and the money is simply gone from the drawer. At month-end this appears as a cash shortage nobody can attribute.
The till does not match the day's sale. Small daily differences are normal; a consistent bias in one direction is not. The signature to look for is direction, not size.
An order is served, the customer pays cash, and the bill is voided or discounted afterwards. The food is gone from stock; the sale is not in the POS. This is the leak most invisible to a POS report, because the POS is the instrument being used to hide it.
The dish that was costed at ₹96 two years ago now costs ₹128 to make, and the menu price never moved. Not theft at all — but the same effect on the bank balance, and it compounds silently across a menu of two hundred dishes.
Published industry percentages exist, but they are a poor guide for any individual restaurant: a bar, a QSR and a 200-cover family restaurant have entirely different exposure, and averages across formats say nothing useful about yours. We are not going to quote a number here that we cannot stand behind for your business.
The honest answer is that leakage is measurable in your own restaurant, this month, from records you already keep. The arithmetic is simple:
Opening stock + purchases − closing stock = what was actually consumed.
Compare that against what your sales should have consumed at recipe quantities. The difference, priced at your own rates, is your leakage figure — not an industry estimate, but yours.
You can run this on one item before you run it on everything. Pick the most expensive thing you buy, count it on a Monday morning, count it the following Monday, and put the week's purchases and sales beside it. Most owners find their first real number within a week of trying this. Our restaurant leakage calculator does the arithmetic for you, and the food cost calculator handles the COGS side.
A note on scale. In the restaurants we work with, the largest single findings are usually not thefts. They are a rate typed wrong that inflated a month's stock valuation, an item recorded under two spellings so half its usage was invisible, or a purchase everybody remembers that nobody wrote down. Leakage control is mostly a records problem wearing a theft costume — which is good news, because records can be fixed.
Detection is a sequence, and the order matters — running the sophisticated checks on unreliable records produces confident nonsense.
Controls that survive contact with a busy kitchen have three properties: they take seconds, they belong to one named person, and they leave a record that someone else reads.
A working set of daily, weekly and monthly controls:
| Frequency | Control | What it catches |
|---|---|---|
| Daily | Closing cash counted against day's sale | Cash leakage, till bias |
| Daily | High-value stock count (10–15 items) | Theft, over-portioning |
| Daily | Every bill photographed at entry | Missing bills, unrecorded expenses |
| Daily | Wastage logged with reason | Wastage disguised as shrinkage |
| Weekly | Full store count | Slow leaks, count drift |
| Weekly | Purchase rate review per item | Price drift, vendor overbilling |
| Weekly | Void/discount report by staff member | POS manipulation |
| Monthly | Actual vs theoretical food cost | Recipe variance, systematic loss |
| Monthly | Vendor ledger reconciled from your side | Double payments, stale dues |
| Monthly | Closing stock valued and carried forward | Valuation errors, phantom stock |
Manual control works. Plenty of well-run restaurants do this on paper and in spreadsheets, and a diligent owner with a register will out-perform a badly used app. The difference is not accuracy — it is latency and attribution.
| Question | Register / spreadsheet | Accountant only | POS only | Control software |
|---|---|---|---|---|
| When do you learn about a problem? | When you look | Month-end or later | Not at all (sales only) | Same day |
| Stock vs sales comparison | Manual, slow | Rarely done | No stock side | Automatic |
| Proof for an expense | Paper, if kept | Bill copies | Not covered | Photo at entry |
| Who recorded what | Handwriting | Not tracked | POS user | Named, per role |
| Rate drift per item | Possible, tedious | Not typically | Not covered | Flagged |
| Cost as it scales | Your time | Fees | Already paid | Subscription |
The honest summary: a POS tells you what you sold. An accountant tells you what it added up to, after the month is over. Leakage control is about the interval in between — and about the physical side of the business that neither instrument touches.
Leakage software does not detect theft. It removes the places a discrepancy can hide, so that what remains is small enough to ask a person about. In practice that means four things:
RestroMatrix is built for exactly this interval: purchases, inventory, wastage, cash and expenses recorded by the people already doing the work, reconciled daily against the sales your POS already records. It does not replace your POS, and it does not claim to identify theft — it shows you the gap, on the day, with the day and the person attached.
Before evaluating any software, measure the gap in your own restaurant. The calculator uses the same arithmetic described above, on figures you already have.
Calculate your leakage Food cost calculatorMoney that leaves the restaurant without appearing as a legitimate recorded cost — stock used but not sold, goods paid for but not received, expenses without bills, or cash that never reaches the till.
Count one high-value item at the start and end of a week, and compare opening + purchases − closing against what your sales should have consumed. A persistent gap in one direction, week after week, is leakage rather than counting error.
It depends heavily on format — a bar, a QSR and a full-service restaurant are not comparable, and rent and staffing change what a business can carry. The number that matters is your own trend on a consistent method, not a published average.
No. Software can show that 8 kg left the shelf while sales account for 5 kg, on a particular day and shift. Whether that is theft, over-portioning, unrecorded wastage or a missed entry is a question for a person standing in the kitchen.
High-value fast-moving items daily; the full store weekly or fortnightly. Monthly-only counting means a discrepancy found on the 30th could have begun on any of thirty days.
Usually purchase price drift, portions growing quietly, recipes never re-costed after rate changes, unrecorded wastage, or stock counted less often than it moves. Each leaves a different signature, which is why the cause is findable.
No. Recorded wastage is a known, explainable cost. Wastage becomes leakage only when it goes unrecorded — at which point it is indistinguishable from theft in the numbers.
With the item list. One item, one name, one rate. Every other control is built on it, and it is the cheapest fix on this page.
RestroMatrix records purchases, stock, wastage, cash and expenses as your team works, and shows the gaps the same day. Keep your POS.
How RestroMatrix works Pricing