Why Don't Your Restaurant Numbers Match?

Why do restaurant numbers stop matching across reports?

Restaurant numbers stop matching because they come from three separate systems, POS, bookkeeping, and payroll, and if any one of those systems is dirty, every number that depends on it is wrong before an owner even opens the report.

Most owners treat their P&L as a single source of truth. It is not. It is an output. The POS records what was sold. The bookkeeper records how that gets categorized. Payroll records what people were actually paid for the hours they actually worked. Three different systems, three different chances for the data to drift from reality. When the sales report says one thing and the bank balance says another, the problem is rarely one bad number. It is usually a gap between two or three of these layers that nobody has lined up.

Clean data is not a nice-to-have before the real decisions get made. It is the prerequisite for every decision a restaurant owner makes, from pricing to staffing to whether a location is actually worth keeping open.

What is a Data Integrity Audit for a restaurant?

A Data Integrity Audit is a structured check of the POS, the bookkeeping, and the payroll, run in that order, to find where the data breaks down before it ever reaches a P&L.

The audit has three layers. Layer one is the POS, where sales are first recorded. Layer two is bookkeeping, where those sales and expenses get categorized into a chart of accounts. Layer three is payroll, where scheduled hours are supposed to match what actually got clocked and paid. Each layer can be checked on its own, but the point of running all three is that a clean Layer 1 means nothing if Layer 2 is a mess, and a clean Layer 2 means nothing if Layer 3 doesn't reflect what actually happened on the floor.

Most owners stop at the POS because it is the layer they look at every day. That is also why it is the least informative layer to audit in isolation. A P&L can look fine with perfect POS data and still be wrong underneath it, because the bookkeeping or the payroll never got checked.

Why can't managers be trusted to self-report their own numbers?

Managers can't be trusted to self-report because they are graded on the same numbers they are reporting, which creates a built-in incentive to round everything up to "excellent."

This is not a question of honesty. It is a question of incentive structure. If the person being measured is also the person producing the measurement, the measurement is compromised, regardless of how trustworthy that person is on every other front.

“If you don't check, I guarantee it's totally a lie. From my experience, if you don't check, they give you wrong information.”

One multi-location operator in the program described his team filing weekly reports that said "excellent" across hospitality, cleanliness, and food quality, every category, every week. There was a verification step in place, but it only kicked in reactively, after something had already gone wrong. That is not a system. That is hope dressed up as a report.

The fix is not to distrust the team personally. It is to build a system that produces the numbers independently of the people being graded on them, so pressure can actually be applied where it is needed.

Results are not typical and will vary with your business, your market, your team and how much of the work you actually do. Client figures on this site come from recorded interviews and are dated. Nothing here is a guarantee of revenue, profit or growth.

What should an owner check first in the POS system?

The first check is whether menu items, modifiers, and sales data are categorized cleanly enough to trust at all, because dead items and miscategorized items turn every downstream report into noise.

• Are items categorized correctly, with appetizers filed as appetizers rather than dumped into "misc food"?
• Are modifiers like upcharges and protein swaps actually tracked?
• Can sales be pulled by item, by daypart, and by server?
• Are there dead items still active in the system that are no longer on the menu?
• Do POS categories match the chart of accounts used in bookkeeping?

A menu item that was discontinued but is still active in the POS does not just sit there harmlessly. It shows up in reports, confuses comparisons month to month, and makes it harder to see what is actually moving. If it is not on the menu, it should not be in the system.

How much miscellaneous expense is too much in a restaurant's books?

Any "miscellaneous" or "other" category above 2% of total expenses is a sign the bookkeeping is hiding real cost drivers instead of revealing them.

Bookkeepers are not trying to deceive anyone. But a bookkeeper who does not understand restaurant operations will lump things together out of convenience. "Supplies" might mean cleaning chemicals one month and smallwares the next. Once categories start absorbing things inconsistently, the P&L stops being a tool for decisions and starts being a guess with a dollar sign in front of it.

Checking Layer 2 means pulling the chart of accounts and looking for three things: whether any miscellaneous bucket has crossed that 2% line, whether food and beverage are separated rather than combined, and whether labor is broken into hourly FOH, hourly BOH, salaried, taxes, and benefits rather than one lump figure. It also means checking whether the bookkeeper categorizes consistently from month to month, because a category that means one thing in January and another in March makes trend analysis useless.

How do you know if payroll numbers can be trusted?

Payroll can be trusted when scheduled hours and actual clocked hours for the week match within 3%, and anything wider than that needs to be flagged and explained.

Layer 3 is the payroll check. Pull the scheduled hours against the actual clocked hours for the last week and look for a variance over 3%. Then look at why. Early clock-ins and late clock-outs add up fast and rarely get questioned in the moment. Overtime is either pre-approved on purpose or it is just happening, and those are two very different situations for a P&L. Shift swaps should be entered into the system before they happen, not reconstructed afterward to match what was actually worked.

Labor is usually one of the largest line items on a restaurant P&L, so a payroll layer that does not match reality quietly distorts one of the most important numbers an owner has.

What mistakes keep owners from trusting their own numbers?

The three most common mistakes are trusting the bookkeeper without checking, ignoring dead POS items, and stopping the audit at Layer 1 instead of carrying it through bookkeeping and payroll.

Trusting the bookkeeper without checking assumes they understand the restaurant the way the owner does. They usually do not, and the miscategorization that follows is not malicious, it is just a gap in context. Ignoring dead POS items seems harmless because they do not generate sales, but they generate confusion in every report that runs off that menu data. And auditing only Layer 1 gives a false sense of security. Perfect POS data sitting on top of messy bookkeeping or unverified payroll still produces a P&L that cannot be trusted.

Success in this part of the business comes down to the quality of the information, the quality of the analysis applied to it, and the quality of the execution that follows. Information is the layer that can be fixed fastest, sometimes within months, once someone commits to checking it rather than assuming it.

Owners who want a structured way to run this audit, along with the rest of the systems built to run a restaurant without being dependent on any one person's reporting, can look at how it works inside the Founders Board or see how operators have applied it in the case studies. Related reading on building the reporting structure behind this audit is covered in the two pillars of data and restaurant P&L line items.