In this article
A few euros' cash-register variance never feels like a problem — it feels like rounding, a forgotten receipt, something that will sort itself out tomorrow. That exact feeling is why it almost never gets investigated, and why, when something genuinely is wrong, it takes so long to surface.
Every night, someone counts: the float plus card takings plus cash takings should match what's actually sitting in the drawer. Most nights it does, give or take a few euros. Those few euros are rarely tracked on their own — they disappear into "small variance," a category nobody ever totals up over a month, let alone a year.
That's exactly the problem. A cash shortage isn't one event you can investigate — it's a habit that only shows itself if you track it. Most venues don't, which means the difference between "a colleague fumbled the change once" and "a pattern that's been running for months" stays invisible until someone happens to add it up.
This isn't an article that assumes every cash variance is theft. Most aren't — they're keying errors, wrong change handed back, a tip that never got counted separately. But the numbers below explain why you can never tell the difference without counting, and what it costs you if you don't.
Below is a calculator that runs your own average daily variance out to a yearly figure — and to what that figure becomes by the time a pattern like this typically gets noticed. Everything runs in your own browser: nothing is sent anywhere and nothing is stored.
Why a cash shortage is so easy to wave off
Every other cost on this site has a clear, recognisable shape: an invoice, a payslip, a lease. A cash-register variance doesn't. It's a number that's a little different every night, sometimes positive, sometimes negative, and that averages out close to zero over a month — which makes it tempting to conclude it cancels itself out and therefore isn't a problem.
That average hides exactly what matters. A till that's €20 over one night and €18 short the next adds up to almost nothing, but that's a completely different signal from a till that's consistently €2 short every single night. The first pattern is noise. The second is a habit — and habits are exactly what the numbers below describe.
And unlike a VAT audit or a supplier invoice, a cash shortage rarely produces one moment that forces you to look into it. It's just there, night after night, until someone — often by accident — adds it up over a longer stretch.
The Ultimate Guide Restaurant Finances: The Complete Guide From financing to cash flow to daily cash control: everything to do with your venue's money, in one guide. Open the guideThe 7 numbers, and what each one tells you
In the order they build the case: first how often a variance is actually something other than a rounding error, then how it usually comes to light, and finally what it costs you when it doesn't.
1. 75% of restaurant shrinkage is attributed to internal theft, not spoilage or error
Industry research into restaurant inventory and cash loss keeps landing on the same picture: the majority of what "goes missing" — cash, stock, recorded sales that don't match what came in — is attributed to internal theft rather than spoilage, breakage or simple arithmetic mistakes. That figure flips most owners' instinct on its head: the reflex is to explain a cash shortage as innocent first, while the data says the majority of shrinkage isn't.
That's not a reason to view every staff member with suspicion — most individual cash variances genuinely are innocent. It's a reason to look at the pattern over time rather than judging each night on its own, which is exactly what the calculator further down does.
2. Roughly 4% of revenue, up to $3–6 billion a year — what internal theft costs the US restaurant industry
Independent industry estimates put the annual cost of internal theft in the US restaurant sector at roughly 4% of revenue, totalling an estimated $3 to $6 billion a year. No equally large-scale, comparable European figure exists — the sector here is more fragmented and less centrally surveyed — but the order of magnitude is what matters: this isn't a story about isolated incidents, it's a structural share of revenue that almost no bookkeeping line item ever captures on its own.
For most independent venues, 4% of revenue is bigger than the net margin itself. That's exactly why a cash shortage shouldn't be treated as a rounding curiosity: structurally, it sits in the same order of magnitude as the difference between a venue that's profitable and one that just isn't.
3. 75% of employees admit to having stolen from an employer at least once
A California Restaurant Association survey of hospitality workers turned up a strikingly high figure: three in four admitted to having taken something from an employer at least once during their career — cash from the till, free food, unrung drinks for themselves or friends. That's not a figure about how often it happens at your venue specifically, but it does correct the baseline assumption: it isn't rare, it's closer to the norm that it happens at some point.
The point of this figure isn't to distrust your current team — most people who once took something did it once, years ago, somewhere else, and are entirely trustworthy today. The point is that a system built on "I'd notice" is a system relying on exactly what the numbers contradict.
4. 43% is caught by a tip, 14% by review, only 5% by accident
The Association of Certified Fraud Examiners' (ACFE) international fraud research has tracked, for years, how fraud — including cash theft — actually comes to light. The result is strikingly consistent: 43% of cases are uncovered by a tip (usually from a colleague), 14% through internal audit or account review, and only 5% by pure accident.
That means "it'll turn up eventually" is the least likely way a pattern actually surfaces. Luck almost never does the job. What does work is a standing habit of counting and comparing — exactly the mechanism behind a checklist — and a culture where a colleague who notices something feels able to say so.
Three detection methods, from the same international fraud research — and why "it'll turn up" is the weakest of the three.
Source: Association of Certified Fraud Examiners, Occupational Fraud 2024: A Report to the Nations. Full citation in the article text above.
5. 14 to 24 months — how long a pattern typically runs before anyone notices
Research into how long internal fraud goes undetected consistently lands on a window of roughly 14 to 24 months — averaging around 18 months — before someone catches it, and seven in ten cases run longer than a year. That's not an outlier; it's the average. A small, consistent cash variance is exactly the kind of pattern that stays under the radar long enough to compound into an amount that no longer explains itself as "rounding."
The reason is simple: a variance of a few euros a night never trips an alarm. There's no threshold it crosses, no red flag that lights up. The only thing that makes it visible is laying several weeks or months side by side — something almost no venue does as a matter of routine.
A small daily variance is the visible tip. What it becomes by the time a pattern is typically noticed is not the smaller half.
Based on this article's default figures (average daily variance, days open, the 14–24 month average detection window). Use the calculator below to run the same math with your own numbers.
6. $141,000 — the median loss at small businesses once a case is finally caught
The ACFE's research also breaks losses down by organisation size, and the result is counter-intuitive: at organisations with fewer than 100 employees — essentially every independent restaurant — the median loss per detected case is HIGHER than at large organisations, not lower. The explanation isn't that small venues face a bigger theft risk itself; it's that they run structurally fewer controls — no separate accounting department, no dual counting, no system that catches a pattern early — which lets it run longer before anyone notices.
If a pattern like the one in the calculator above ran for years without a check, €2,808 — the share the 75% industry figure would attribute to theft rather than error — is exactly the kind of amount a small business only discovers late, and then all at once.
7. Roughly half — how much smaller the loss is where a standing counting habit exists
That same ACFE research compares organisations with a basic reporting or verification mechanism to those without one: where that mechanism is missing, the median loss runs at roughly double what it does where one exists. For a till, that mechanism isn't complicated — it's a fixed, daily count that gets written down rather than left in whoever's closing that night's memory.
At an average daily variance of €12.00, open six days a week, that adds up to €3,744 a year — an amount nobody would call "rounding" if it appeared as a single line. That's exactly what a habitual daily close does: it puts the number on paper every night, so it never gets the chance to build up unnoticed for twelve or eighteen months.
What your own cash variance actually adds up to
Enter your own average daily variance, how many days you're open, and how many staff members have access to the till — the calculator does the rest.
The 75% theft-share estimate is illustrative, not a guarantee: it's the widely-used industry figure from the numbers above, applied to your own input rather than invented.
The calculator: what your cash variance actually adds up to
Three numbers about your own till — see what a small daily variance becomes over a year.
—
Illustrative, built entirely from your own input — not a measurement of theft at your venue. The 75% estimate and the 18-month detection window are widely-used industry figures, not a guarantee for any individual case.
Close out every night properly with the free Daily Close toolThe point isn't that every cash variance is theft — most nights balance fine, give or take a few euros of rounding. The point is you can never tell the difference if the variance isn't tracked.
The Daily Close tool counts the till, compares it to what's expected, and tracks the variance night by night — so a pattern is visible after two weeks, not after eighteen months, by accident.
What to do with this, this week and after
You don't need a fraud investigation. You need a habit that fits into two minutes a night.
This week
- Count the till properly, deliberately, once: float, card takings and cash takings tracked separately, against what's actually in the drawer.
- Write down the variance — not just whether it balanced, but the exact amount and direction (over or short).
- Do this for two weeks straight before drawing any conclusions. One night tells you nothing; two weeks shows the first pattern.
If a pattern shows up
- Look at the process first, not a person: is there a dual count, a fixed counting moment, are tips kept separate from till takings?
- Where possible, limit how many people work the till unsupervised during the count, or rotate a fixed person responsible per shift.
- Talk about it openly with the team — a team that knows counting happens also counts along, which is exactly what the {tipPct}% figure above says about how most cases actually come to light.
To make it structural
- Make the daily close a standing part of closing up, not an extra task that gets skipped on a busy night.
- Keep the history somewhere outside the memory of whoever's closing — on paper or digitally — so a pattern becomes visible over weeks, not years.
- Use the free Daily Close tool to record this without extra work: counting, comparing and tracking the variance already sit in one template.
The variance nobody adds up
None of the seven numbers above say your team can't be trusted. They say the opposite of what most owners assume: the problem is rarely one dramatic moment — it's a small, repeated variance that never gets totalled, inside a system that relies on luck to notice it.
The fix isn't suspicion either. It's a habit: count, compare, write it down, every night — the same kind of discipline you already apply to stock or to reservations, now applied to the till.
Start this week with the calculator above to see what your own variance adds up to, and let the Daily Close tool take over once you know where you stand.