Kitchen Manager Daily KPIs: The Eight Numbers Worth Checking

A kitchen manager can measure almost anything. The POS reports dozens of numbers, the scheduling tool has its own, the inventory system has more, and every one of them looks important in a meeting. The trouble is that nobody can act on forty numbers before lunch service. In practice, a manager who tries to watch everything ends up watching nothing, and the only number that gets attention is the one the owner asked about yesterday.

We think about eight numbers are enough for the daily look. Each one should answer a question you can do something about today. The rest belong in the weekly review.

The eight

Here is how we read each one.

Waste

Look at the value, not the weight, and break it into reasons: overproduction, spoilage, trim, mistakes and returned plates. A $90 day where $70 is overproduced rice is a prep problem. A $90 day where $70 is a dropped tray of steaks is bad luck.

The common mistake is reading a low waste number as good news. If the waste log says $8 on a Saturday, it almost certainly means nobody logged anything. In a busy kitchen, a day with no recorded waste is a data problem, not a result.

What to do with it: pick the biggest single reason from yesterday and ask about it at the pre-shift. One question a day is enough.

Prep vs par

At the end of service, compare what was prepped with what was used and what is left. If you prepped 6 quarts of pico, used 3 and threw away 2 because it was past its time, the par is too high for that day of the week. If you ran out at 8:15, it is too low or the prep list was not followed.

The useful daily number is the count of items that ran over or under by a meaningful margin, say 25% either way. Five items over and none under means you are overprepping across the board. One item under every Friday means the Friday par needs to change. Setting par levels covers how to adjust them.

Ticket times

Average ticket time for the whole day is close to useless, because a quiet 3 pm pulls the average down and hides the 7:30 rush. Look at your peak window only (usually the busiest hour of lunch and of dinner) and look at the slowest 10% of tickets, not just the average.

If the average at peak is 14 minutes and the slowest tickets are 18, that is a kitchen running steady. If the average is 14 and the slowest are 32, a few tables waited half an hour, and you need to find out what was on those tickets. Usually it is one station or one dish.

Targets depend entirely on the concept. A quick-service counter might aim for under 5 minutes; a full-service dinner menu with grilled proteins might accept 15 to 20 for mains. Set your own target, and watch the trend rather than any one night.

Checklist completion

This is the share of scheduled checklist tasks (opening, closing, line checks, cleaning, receiving) completed on time, by the person assigned. It is the only number on this list that measures whether the routine work happened, and most of the others depend on it. A missed line check is how a wrong temperature goes unnoticed; a missed closing checklist is how the hood filters do not get washed.

Read it as a percentage but act on the specific misses. A day at 94% is fine if the missed 6% was a deep-clean task that moved to tomorrow. It is not fine if it was the opening temperature round. Look at what was missed, not just how much.

Watch for completion that is too perfect. 100% every day, every task, signed at the same minute at close, usually means the checklist is being filled in from memory.

Temperature exceptions

Count how many readings yesterday were out of range: cold holding above 41°F (5°C), hot holding below 135°F (57°C) under the US Food Code (other jurisdictions use slightly different numbers, such as 60°C for hot holding in some), cooling that missed a checkpoint, or a cooler that read high at opening.

The right number of exceptions is not zero, because equipment drifts and a door left open will show up. The right number of unresolved exceptions is zero. Every out-of-range reading should have a note: what was done (discarded, reheated, moved, door fixed), by whom, and when. If you see three exceptions on the same unit in a week, call the refrigeration technician before it fails on a Saturday.

Labor percentage

Labor cost as a percentage of sales for the day, using actual clocked hours, not scheduled ones. There is no universal right number: it depends on your concept, local wage levels and how much prep you buy in rather than make. Use your own last eight weeks as the baseline.

A single day's labor percentage swings a lot with sales, so do not overreact. A rainy Tuesday with sales down 30% will show high labor even if the schedule was right. What you can act on is the gap between scheduled and actual hours. If people consistently clock out 40 minutes after their scheduled end, either the closing work is too much for the time given or the close is not organized.

Food cost variance trend

The real food cost number comes from the weekly stock count and the food cost calculation that follows it. Day to day, you are only watching the direction: is the gap between theoretical cost (what recipes and sales say you should have used) and actual cost narrowing or widening?

Some kitchens estimate daily food cost from purchases divided by sales. That number jumps around with delivery days and means very little on its own. If you use it, use a rolling seven-day figure, and only to spot a trend early. The decision gets made weekly, with a proper count behind it.

86'd items

Every item that ran out during service, with the time it ran out. One 86 at 9:45 pm on a busy night is normal and arguably good planning. The same item 86'd at 7 pm three days in a row is a par problem or an ordering problem. An item 86'd because a delivery did not arrive is a supplier issue, and it should be written down so you can raise it with them.

Also look at items that should have been 86'd and were not: dishes sent out with a substitute ingredient nobody approved, or a special that kept selling with the last portions stretched thin. Those do not show up in the POS, which is why the chef's word at the morning check matters.

Reading them without drowning

Eight numbers is still a lot if you look at all of them with equal weight every morning. A few habits make it manageable.

Take ten minutes, the same ten minutes each day, before the first deliveries or right after them. Look at yesterday only. Write down at most two things to act on. If nothing stands out, nothing needs to happen, and that is a good morning.

Separate safety numbers from money numbers. Temperature exceptions and checklist misses on critical tasks need action today, every time. Waste, labor and food cost need action when a trend appears, which usually takes a week of data to see.

Compare against the same day last week, not against yesterday. Tuesday looks nothing like Saturday in any kitchen, so comparing them only produces noise.

Share the numbers with the team, briefly. A one-line note at pre-shift ("waste was mostly rice yesterday, let's halve the 4 pm batch") gets more done than a report nobody reads. If a dashboard pulls checklist completion, temperature exceptions and open issues into one view, as TenorOps does for each location, use it; otherwise a single whiteboard by the office with yesterday's eight numbers works fine.

And when you have a bad day, short-staffed, a cooler down and a bus tour that showed up unannounced, ignore the money numbers entirely. Check that the safety numbers are clean, and look at the rest tomorrow.