How to control payroll costs without losing good staff

Cutting hours is not a strategy

When payroll gets uncomfortable, the standard move is to take hours out of the rota. It is fast, it is visible, and it produces a number that looks better next week.

It is also the move most likely to make the problem worse. Hours cut below what the room needs show up first as slower service, then as smaller checks, then as covers that do not come back, then as the two people who were carrying the section handing in their notice because every shift is now short-staffed. By the time payroll looks healthy as a percentage, the revenue underneath it has moved too.

Cutting is not wrong. Cutting blind is wrong. The difference between the two is whether you can see where the hours are actually going.

Payroll is a ratio, not a total

The total on the payroll run is almost meaningless on its own. A large number in a busy month can be excellent and a small number in a quiet month can be ruinous. What matters is payroll as a percentage of the revenue it produced, tracked over time.

That reframing changes the question you are asking. It stops being “how do we spend less on staff?” and becomes “where are we paying for hours that did not produce anything?” Those are different questions with different answers, and only the second one has answers that do not cost you people.

It also lets you separate two costs that behave nothing alike: salaried management, which is fixed and sits in your break-even, and hourly staff, which is variable and should move with covers. Blending them into one payroll number hides the only part you can actually steer week to week.

The schedule is where the money is decided

Almost all of your controllable labour cost is committed at the moment the rota is published, several days before a single euro is earned. Everything after that is damage control.

Most schedules are built from habit. Same shape as last week, same shape as last year, adjusted for holidays and whoever asked for a Saturday off. That is not a plan, it is a copy, and it carries forward every staffing decision anyone has ever made — including the ones that were wrong.

A schedule built against expected covers looks different. It is not more aggressive; it is differently shaped. It puts people where the volume is and takes them out of the hours that never had any, which is usually a quiet Tuesday opening and the last hour of a Wednesday nobody wanted to work anyway.

• Compare payroll to revenue by day, not just by week — the loss is rarely spread evenly
• Split salaried from hourly before you read the number, or the fixed half hides the variable half
• Watch overtime separately: it is the most expensive hour you buy and the easiest to normalise
• Look at what a shift produced, not only what it cost

The costs that are not on the rota

The rota shows what you paid. It does not show what turnover cost you, and turnover is where payroll quietly becomes expensive.

Replacing someone is never one cost. It is the recruiting, the interviews, the weeks of paying two people for one competent one, the errors during training, the slower service while the section learns a new person, and the load carried by the staff who stayed — some of whom will leave because of it. None of that appears in a payroll percentage. All of it appears in your year.

This is why the crude version of payroll control is such a bad trade. Trimming hours from your best people to protect a percentage is one of the more reliable ways to buy a resignation, and a resignation costs multiples of the hours you saved.

What to look at before you cut anyone

Before touching a single shift, three things are worth checking, and all three are cheaper to fix than a departure.

First, whether the schedule matches the covers, day by day. Second, whether overtime is structural — the same people, the same nights, every week — because structural overtime is a scheduling problem being paid for at a premium. Third, whether the shift shapes still fit the business you have now rather than the one you had two years ago, which is where the quietly unnecessary opening hour usually lives.

In our experience most of the recoverable payroll cost is in those three places, not in anyone’s hourly rate. They are also the only fixes that leave the team intact, which matters more than the percentage does.

Good people are the cheapest thing in the building

A strong server sells more, comps less, and turns a section faster. A strong line cook wastes less product, portions consistently, and produces a lower food cost than a weak one on the same recipes. These effects do not show up in the payroll line. They show up in every other line.

Which means the goal was never a lower payroll number. The goal is a payroll number that is earning what it costs. Those are very different targets, and only one of them still has a restaurant at the end of it.

Make it visible before it becomes a decision

Payroll cost that is only reviewed when it hurts is payroll cost you will always address in a panic, and panic produces the cutting move every time. Seen weekly, alongside revenue and covers, it stops being an emergency and becomes a small adjustment — a shift moved, an opening hour dropped, one Tuesday staffed for the Tuesday you actually get.

That is what the payroll view in master.restaurant is for: your labour cost against the revenue that paid for it, week by week, next to food cost and break-even, so you can see which one is actually moving. You can run it against your own figures in the free three-day demo before deciding anything, and if it belongs in your week, it is €499 once for lifetime access rather than another monthly bill.

The number itself will not fix your rota. It will tell you which week to look at, which is the part nobody has time to work out during service.