Inventory variance is the gap between what your kitchen should have used, based on sales and recipes, and what it actually used, based on your stock counts. Calculate it as (actual usage − theoretical usage) ÷ theoretical usage × 100. Then break the result down by category and by item, because a single venue-wide number tells you something is wrong but never what.
This guide shows you the formula, a worked example, the benchmarks Australian operators use, the four usual causes, and a weekly routine that keeps the number honest.
TL;DR
- Variance = actual usage minus theoretical usage. Theoretical comes from POS sales and recipe costs. Actual comes from opening stock plus purchases minus closing stock.
- A common rule of thumb: under 3% is tight, 3–7% is normal, over 7% needs urgent investigation.
- Almost every variance traces back to one of four causes: portioning, untracked waste, shrinkage or theft, and bad data (recipes, units, deliveries).
- Count the high-value and high-risk items weekly. Count everything monthly. Investigate by item, never by venue total.
Table of Contents
- What inventory variance actually measures
- How to calculate it, step by step
- What is a good inventory variance percentage?
- The four causes behind most variance
- Investigate by item, not by total
- A weekly routine that keeps variance in check
- Keep the records the ATO expects
- Our perspective
- See your variance weekly with PantryHub 📦
- FAQ
- Sources
What inventory variance actually measures
Your till tells you what you sold. Your recipes tell you what that should have cost. Your stocktake tells you what really left the shelves. Variance is the distance between the last two.
Two terms get mixed up, so pin them down:
- Theoretical usage is what you should have used if every plate matched the recipe card and nothing was wasted, spoiled or taken.
- Actual usage is opening inventory plus purchases minus closing inventory, in units or dollars.
If you sold 120 chicken parmas and each uses 180 g of chicken breast, theoretical usage is 21.6 kg. If the count says you used 24.5 kg, you have 2.9 kg of unexplained chicken. That is variance.
The same idea shows up in accounting as food cost variance: the gap between the food cost a restaurant should have incurred for its sales and what it actually spent, expressed as a percentage.
How to calculate it, step by step
You need four inputs for the period: opening count, purchases, closing count and sales by menu item. A weekly period is the most useful for fresh produce, proteins and dairy.
- Count opening stock. This is last week's closing count. Same units, same storage areas.
- Add purchases. Use invoices received, not orders placed. Credits for short or damaged deliveries come off.
- Count closing stock. Count every location: coolroom, freezer, dry store, bar, the shelf above the pass.
- Work out actual usage. Opening + purchases − closing.
- Work out theoretical usage. Multiply units sold of each dish by the recipe quantity of each ingredient, then add them up per ingredient.
- Compare. Variance % = (actual − theoretical) ÷ theoretical × 100.
A worked example
This is an illustration with made-up numbers, not data from a real venue. A café tracks flat white milk for one week.
| Line | Litres |
|---|---|
| Opening stock | 48 |
| Purchases | 240 |
| Closing stock | 36 |
| Actual usage | 252 |
| Theoretical usage (from sales × recipe) | 228 |
Variance is (252 − 228) ÷ 228 × 100, which is 10.5%. That sits above the 7% line and is worth chasing. The likely suspects: over-pouring in the jugs, milk wasted on remakes, a recipe card that assumes 200 ml when the baristas use a larger cup size, or a delivery that was short and never credited.
Note the sign. Some teams show a positive number when actual exceeds theoretical, others the reverse. Pick one convention and put it at the top of the report so everyone reads it the same way.
What is a good inventory variance percentage?
There is no single legal or regulatory benchmark. The most quoted rule of thumb, from a restaurant inventory guide, is:
| Variance | Reading |
|---|---|
| Under 3% | Tight control |
| 3–7% | Normal for most venues |
| Over 7% | Structural issue, investigate now |
Treat those as starting points, not targets. A few things move the number:
- Item type. Bottled spirits can be counted to the millilitre. Fresh herbs, fries and cooking oil cannot. Expect higher variance where yield is hard to measure.
- Menu complexity. A fish and chip shop with ten items has an easier job than a restaurant with 80 and daily specials.
- Count frequency. Monthly counts hide problems inside a big average. Weekly counts on key items expose them while you can still act.
Why does a small percentage matter? The ATO's small business benchmarks for restaurants put cost of sales at roughly 31% to 39% of turnover, depending on the turnover band. With cost of sales that large, a 5% unexplained variance on food is a real chunk of your gross profit leaking out, quietly eating into margin every week.
The four causes behind most variance
After a number appears, sort the gap into causes. Each one needs a different fix.
1. Portioning
The chef who plates 220 g instead of 180 g isn't stealing. They're being generous, or rushed, or working from memory. Across hundreds of covers, it adds up fast.
Signs: variance concentrated in high-volume proteins, sides and sauces; it gets worse on busy shifts or with particular cooks.
Fix: scales and portion tools at the pass, plated photos, and recipe cards that match what the kitchen really does. Retrain, then recount.
2. Untracked waste
Spoilage, trim, dropped plates, remakes, staff meals and over-prep all consume stock that your sales never explain. Many operators record waste loosely. Research from End Food Waste Australia found that although digital tools are widely used in quick service restaurants, only 7% use scales and most rely on estimates, which leaves operators unaware of the true scale of their losses.
Signs: variance spread across many perishables, worst after quiet days or menu changes.
Fix: log waste at the moment it happens, with a reason code. Our guides on measuring prep versus plate waste in two weeks and restaurant waste tracking walk through setting it up without slowing service.
3. Shrinkage and theft
Shrinkage covers stock that disappears for reasons other than waste: theft, unrecorded giveaways, unrecorded comps, deliveries signed for but not received. Spirits, steak, seafood and anything easy to carry are the usual targets.
Signs: high-value items are short while low-value items reconcile; losses cluster around particular shifts, delivery days or storerooms.
Fix: lock and limit access, check deliveries against invoices at the door, and use team logins so every adjustment has a name on it. For the full picture, see what to measure for stock shrinkage in Australian kitchens. Pubs and clubs with large liquor ranges sometimes bring in an independent service for beverage counts. Associated Stocktaking Services is one Australian example that lists liquor and beverage stocktakes for hospitality.
4. Bad data
This is the cause nobody wants to hear about, and often the biggest one. If the recipe says 150 g but the cook uses 200 g, the recipe is wrong. If the invoice is in cartons but the count is in kilos, the conversion is wrong. If a menu item was sold through the POS under a modifier that isn't linked to a recipe, theoretical usage is wrong.
Signs: a single item shows variance every single week, in the same direction; a new menu item or supplier change lines up with the jump.
Fix: audit the recipe, unit of measure and POS mapping for that item before you blame anyone.
Investigate by item, not by total
A venue-wide variance of 4% can hide a 20% problem on beef and a negative variance on cheese that's cancelling it out. Always work down the list:
- Sort items by dollar variance, not percentage. A 30% variance on parsley is trivia. A 6% variance on eye fillet is not.
- Take the top five and check data first: recipe, units, POS mapping, last delivery.
- Recount the item if the number looks implausible. Counting errors are common.
- Only then look at behavior: portioning, waste, access.
- Write down what you found and what you changed. Next week's number will show if it worked.
The 80/20 rule helps here. A small share of your items usually accounts for most of your spend, so those are the ones to count weekly. Everything else can wait for the monthly count.
A weekly routine that keeps variance in check
You don't need a full-day stocktake every Monday. A rhythm like this works for most cafés, pubs and restaurants:
- Daily: log waste and staff meals as they happen. Check deliveries against invoices before signing.
- Weekly (30 to 45 minutes): count your top 20 to 30 items by value. Run the variance report. Pick the top three items and assign someone to explain each one.
- Monthly: full count, reconcile against your accounting, review recipe costs against current supplier prices.
- Quarterly: review par levels, menu mix and supplier pricing.
If your numbers don't reconcile at all yet, start with five steps to reconcile inventory in a week. For a deeper method, our inventory variance playbook for hospitality covers moving accuracy up toward the high 90s. And if your reports aren't telling you much, check you aren't making any of the common inventory mistakes restaurants make.
Keep the records the ATO expects
Variance work doubles as good record keeping. The ATO says you need records of all transactions for buying, maintaining and selling stock so you can substantiate the amounts in your tax return. That includes stocktake records. Keep them for at least five years, as the same page explains. Its BAS and GST tips also list stocktake records among the documents to keep. For questions about how your stock is valued or reported, your accountant or registered tax agent is the right person to ask.
Our perspective
PantryHub was built by a chef, for chefs, and the pattern we see is always the same. Operators know their food cost is off, but they only see it at month end, when the money has already gone. The spreadsheet is the first thing to fail. Someone has to key in every invoice, someone else counts, and nobody has time to compare the count with sales.
Our view is that variance is a habit, not a report. If counting is slow, it happens monthly, and monthly numbers can't tell you which shift or which delivery caused the gap. If recipes live in a binder, they drift from what the kitchen cooks. If waste goes unlogged, it shows up as mystery variance and gets blamed on theft.
So we build around making the habit cheap: counting on a phone, recipes tied to stock, supplier ordering in the same place, low-stock alerts, team access controls so adjustments have a name, and waste and cost reporting for one venue or many. We price in AUD and report with GST in mind because Australian operators shouldn't have to adapt overseas software to their own books.
The tool won't cook for you. The wins come from the routine: weekly counts, one person owning the top-five items, recipes kept honest. Software just makes that routine take 30 minutes instead of a day.
— Admin
See your variance weekly with PantryHub 📦
If you're still running counts on paper or spreadsheets, try doing it where your stock, recipes and suppliers already sit together.
- 📦 Real-time stock tracking across one or many venues
- 🔔 Low-stock alerts so you stop over-ordering
- 🛒 Supplier ordering with Australian suppliers
- 📊 Waste and cost reporting to find the gap
Plans are Starter at $39 per month, Regular at $69 per month and Pro at $99 per month (AUD). You can start your free trial and run your first weekly count on your own menu. Moving from another platform? See how we compare as a MarketMan alternative in Australia. Based in Western Australia? We have a page for restaurant inventory software in Perth.
FAQ
What is a variance in a restaurant?
It is the difference between what you should have used and what you did use. "Should have used" comes from sales and recipes. "Did use" comes from your stock counts and purchases. The gap can be shown in units, in dollars or as a percentage.
How is inventory variance calculated?
Work out actual usage as opening stock plus purchases minus closing stock. Work out theoretical usage from units sold multiplied by recipe quantities. Then divide the difference by theoretical usage and multiply by 100. Do it per item first and add up the dollar value afterwards.
What is a good inventory variance percentage?
A widely used guide treats under 3% as excellent and 3–7% as normal, with anything above 7% calling for investigation. Your own target should reflect your menu and how often you count. Aim to see the number fall over time rather than hitting a universal figure.
What is the 80/20 rule in inventory?
It is the idea that roughly 20% of your items account for about 80% of your spend. In a kitchen that usually means proteins, seafood, dairy, produce staples and spirits. Count those weekly and the rest monthly, and you catch most of the money with a fraction of the effort.
How often should I check variance?
Weekly for high-value and perishable items, monthly for a full count. Longer gaps blur the cause, because one bad delivery or one new cook is averaged away. Short cycles let you tie a number to a shift, a supplier or a menu change.
Sources
- Restaurant Inventory Management: 7 Steps to Less Waste (HappyChef)
- Food Cost Variance Percentage: Meaning and Example (Patron Accounting)
- Restaurants small business benchmarks (Australian Taxation Office)
- Quick Service Restaurant Sector Action Plan Report 2026 (End Food Waste Australia)
- Stock and asset records (Australian Taxation Office)
- BAS and GST tips (Australian Taxation Office)
- Hospitality Stock Control (Associated Stocktaking Services)