Before you make a single decision based on your sales reports, you need to know whether the reports are telling you the truth. In my experience they frequently aren’t — not because the system is broken, but because of how it counts things, and because nobody ever checks.
Here are three specific errors found in one restaurant’s own data. All three are common. All three change the decision you’d make.
Error one: your “net sales” may include tips
This one is worth checking today.
Most point-of-sale systems give you a total amount and a tax amount, and the obvious move is to subtract one from the other and call it net sales. That number is wrong if your system folds tips into the total.
On one specific day, the two calculation methods differed by $491.36. That figure was exactly that day’s tips. Same pattern on every other day tested.
On its own, a few hundred dollars a day sounds survivable. Here’s why it isn’t. That restaurant started prompting for tips in late 2025. In January 2025, tips ran about 0.5% of net sales. A year later, they were 8%.
So when the year-over-year comparison got run using the wrong formula, July looked up 13%. Corrected, the real figure was closer to 2%.
That’s the difference between “we’re recovering, stay the course” and “we’re flat, something needs to change.” Same restaurant, same month, same raw data. One arithmetic choice.
Check yours: pull one day. Add up the tips separately. Compare your reported net sales against your total-minus-tax figure. If they differ by the tip total, every year-over-year number you’ve looked at is inflated — and inflated by more each year if your tip capture has been growing.
Error two: the same item rung up two different ways
This one hides in your item counts rather than your dollars.
That restaurant’s system had two valid paths to ring up a specialty pizza. A server could select the specialty item and then choose a size modifier, or select the size item and then choose a specialty modifier. Both produce a correct ticket. Both charge the customer correctly. The kitchen makes the same pizza.
But if you pull an item report and count only one path — which is what happens by default — you undercount by about 34%.
Think about what that does downstream. Your prep pars are built on those counts, so you under-prep. Your menu engineering says an item is a slow mover, so you consider cutting something that’s actually selling a third more than you think. Your food ordering is off.
Check yours: pick your three best sellers. Count them in your item report. Then count them from raw order data, catching every path they can be rung. If the numbers don’t match, your item-level reporting is unreliable and so is everything built on it.
Error three: your wage data is stale in at least one system
If you run a point-of-sale system and a separate scheduling system, they probably talk to each other. It’s worth knowing exactly what they share.
In that restaurant’s case, the two systems synced employee records, clock-ins, and daily sales. They did not sync wages. Not on a delay — not at all. The scheduling system’s wages had been typed in by hand once, when each person was hired, and never touched again.
The result, verified against payroll: three employees who’d received raises were still priced at their old rates in the scheduler. One had been at the higher rate for twelve days.
| Employee | Scheduler said | Actually paid |
|---|---|---|
| Employee A | $15 | $18 |
| Employee B | $15 | $18 |
| Employee C | $18 | $20 |
And neither system modeled salary at all. Two salaried managers were being priced at made-up hourly rates, which meant the scheduler “saved money” every time it cut their hours — money that didn’t exist, because their pay didn’t change.
The practical rule: a wage change has to land in three places — payroll, your point-of-sale system, and your scheduler. Payroll is the only one that’s authoritative. Never take a labor dollar figure from your scheduling software without checking it against actual payroll.
Why this chapter comes before the others
Every recommendation in this book depends on your numbers being real. An owner making cuts based on a labor percentage that’s wrong, or killing a menu item that’s actually selling, or believing sales are up 13% when they’re up 2%, is making confident decisions on fiction.
Spend an afternoon verifying before you spend a month acting.
If you want a second set of eyes on whether your reports are telling you the truth, book a time.