The Bottom Line: A variance report compares what your POS says you sold against what your inventory says you actually used, and the gap between those two numbers is your shrinkage in dollars. WISK.ai runs that comparison automatically after every count and breaks the result down by category, revenue center, and recipe, so restaurant owners, kitchen managers, and multi-unit operators know which shelf, station, or shift to investigate instead of guessing. Shrinkage stops being a month-end surprise and becomes a number you can act on this week.
What Is a Variance Report, and What Is It Actually Measuring?
A variance report measures the difference between theoretical usage (what your recipes and POS sales say you should have consumed) and actual usage (what your counts and invoices say you did consume). Every dollar in that gap is shrinkage, and a one-point gap on $85,000 in monthly sales is $850 gone.
The math is simple, which is exactly why it works as a detection tool:
- Theoretical cost % = theoretical cost of goods sold ÷ sales. It assumes every pour, portion, and plate matched the recipe.
- Actual cost % = opening inventory + purchases − closing inventory, divided by sales. It reflects what really left your shelves.
- Variance is the difference. Theoretical says you should have used 38 bottles of well vodka. You used 44. That is six bottles nobody sold.
The number itself is not the point. The point is that variance is a pointer. It tells you where in the building to go look.
One clarification worth making, because it trips up a lot of operators: variance is not the same as food cost. Your food cost can sit perfectly on target while you are bleeding product, if your menu mix happens to be covering for it. Variance ignores menu mix entirely and asks one question. Did the product leave the shelf for a reason you can name?
Why Does Finding Shrinkage a Month Late Cost More Than the Shrinkage Itself?
Restaurants run on 3% to 5% net margins, and the National Restaurant Association estimates that 75% of restaurant inventory shrinkage comes from theft, most of it internal. A loss you catch in week one costs you one week. A loss you catch on a month-end P&L has already run for four to five weeks.
Here is what the delay actually does to you:
- The evidence is gone. Nobody remembers which shift was short on Tuesday three weeks ago, so the conversation with your team becomes accusation instead of correction.
- You reorder against the bad number. Inflated usage inflates your pars, so you buy more of the thing that is already walking out the door.
- The behavior sets. Industry estimates put employee theft at roughly 4% of restaurant sales, and the schemes that go undetected longest are the ones that started small and were never questioned.
Speed is the whole advantage. A variance report you read weekly is a control. A variance report you read quarterly is a post-mortem.
Where Should You Look First When a Variance Report Flags a Loss?
Start with the items carrying the most dollars, not the biggest percentage. In most operations, 10 to 15 key items account for around half of total food or beverage purchases, so a 4% variance on your house vodka matters more than a 40% variance on cinnamon sticks.
Work this order. It resolves most variance before you ever reach the uncomfortable conversations:
- Recipe mapping. A cocktail built at 2 oz but mapped in the system at 1.5 oz produces variance every single time it is sold. Check anything added to the menu in the last 60 days first.
- Portion and pour discipline. Free pouring, heavy hands on well spirits, and unweighed proteins at the line are the most common causes and the easiest to fix with a jigger and a scale.
- Receiving. You were invoiced for 12 cases and six arrived. That is variance, and it is your supplier's problem, not your bartender's.
- Unlogged waste and spoilage. Restaurants and foodservice operations generate 12.5 million tons of surplus food a year, and the portion that never gets logged shows up as unexplained variance.
- Transfers. Product moved from the bar to the kitchen, or between outlets, without a record will read as a loss in one place and a gain in the other.
Only after those five come up clean should you treat variance as a theft signal. That sequence protects your team from being blamed for a spreadsheet error, and it protects you from ignoring a real one.
How Much Is Your Current Variance Costing You?
Enter one period of sales and your two cost percentages below. A gap of 1 point or less is controlled, 1 to 2.5 points is a habit problem, and anything above 4 points is structural and needs investigating before your next order.
Run your own numbers before you read further. Most operators are surprised less by the monthly figure than by what it annualizes to.
How Does Manual Variance Tracking Compare to WISK's Automated Analysis?
Manual variance tracking gives you one blended cost percentage, three to five weeks late. WISK's automated actual vs. theoretical analysis gives you variance by item, category, revenue center, and recipe, generated after every count with sales pulled automatically from 60+ POS integrations.
The difference that matters most is the third row. A single blended number tells you that you have a problem. A breakdown by revenue center tells you the problem is in the upstairs bar on weekends, which is something a kitchen manager can act on before the next shift.
The second row matters almost as much. Most spreadsheet variance is wrong before anyone looks at it, because the inputs were typed twice, once into a count sheet and once into Excel, and because partial bottles were estimated by eye. Variance groups in WISK also solve a quieter version of this problem, matching the 750 ml and 1 L formats of the same product so a format switch mid-period does not read as a loss.
How Do Multi-Unit Operators Find Which Location Is Leaking?
WISK's location-based reporting puts every venue and every stock area on one dashboard, so a group operator can rank locations by variance and see whether a cost spike is a company-wide supplier issue or one site behaving differently from the rest.
For multi-unit operators, the useful move is comparison rather than absolutes:
- Four venues running 3% variance is a purchasing or recipe problem at the group level. One venue at 9% while the others sit at 2% is a site problem.
- Central stockroom transfers are a frequent culprit. Product leaves the warehouse, arrives at a venue, and never gets received into that venue's inventory.
- WISK shrinkage alerts push the signal to you rather than waiting for you to open a report. Set your cost thresholds once and get notified by email, in the desktop app, or on your phone when a venue crosses them.
The goal is not to rank your managers. It is to know within a week which one needs help, and with what.
How Does WISK Help You Eliminate Shrinkage Before It Reaches Your P&L?
WISK turns shrinkage detection into a routine instead of an investigation. Counts go in on a phone, sales sync from your POS, invoices are scanned into your costs, and the actual vs. theoretical analysis runs itself, with the WISK theoretical cost calculator pricing every recipe so the comparison is accurate down to the portion.
Operators using it report the practical version of that: counts dropping from two hours to twenty minutes, and cost of goods falling by thousands of dollars in a period once the leaks are visible. Partial bottles get weighed on a Bluetooth scale instead of guessed at. Variance reports arrive broken out by category and recipe. Alerts reach you while you can still do something about it.
If you are running your variance on a spreadsheet right now, you already know the number is late and probably wrong. Book a free WISK demo and we will run your actual vs. theoretical on your own last period, so you can see exactly where your money is going before you place another order.
