Opening and closing checklists: what the data says and what the myth says

Why a signed checklist proves nothing about the closing shift?

A ticked box proves a pen was present, and that is all it proves.

The night that 180-seat restaurant in Bogotá lost 4,100 USD of product because a walk-in freezer door sat ajar, all fourteen boxes carried a signature timed at 11:40 p.m., and the thermometer still read −4 °C by morning. Paper records INTENT; your accounting needs evidence of execution, which is a different animal. Hold it against prime cost: the National Restaurant Association sets the target at 55-65% of sales, and the food block inside that percentage is largely decided between eleven at night and seven in the morning, when nobody is watching. An item that demands a time-stamped photo of the thermometer turns an opinion into data, and that data is the only thing that later lets you argue with your head chef instead of guessing. Fourteen items control more than forty, and this paradox is the hardest one to sell to an owner proud of his sixty-page manual.

List length is a cost variable, not a measure of rigor

Long columns invite the thumb to run down the whole thing in eight seconds, and what you bought was automatic ticking; three or four items covering what genuinely burns money —walk-ins, gas, cash drawer, bar— buy you verification instead. The bar makes it obvious: Sculpture Hospitality measures monthly bar inventory loss between 10% and 20% from overpouring, theft or spoilage, a spread so wide that the width itself tells you where you stand. If your closing list spends an item on mopping the dining room and none on counting open bottles, you do not have a staff problem: you have a checklist that measures the cheap things and leaves the expensive ones unattended. Cut it down, do not extend it. Food inventory turnover benchmarks run from 4 to 8 times per month under the industry rule Sculpture Hospitality publishes, and the bottom half of that range almost always smells of product that moves slowly because it gets lost, not because it sells slowly.

Inventory turnover: the number your closing shift decides for you

That is where closing either pays or costs. A restaurant turning inventory 4 times with the same menu and volume as one turning it 7 has nearly double the capital frozen in the walk-in, and overnight spoilage is one of the three usual culprits. The operational reading is uncomfortable but direct: before you redesign the menu or renegotiate with suppliers, measure two weeks of closings with photographic evidence of temperature and a mise en place count. If turnover climbs on its own, purchasing was never the problem. A sloppy opening gets billed to your tips, and tips are your real recruiting cost. Toast measured a 19,4% average in full-service restaurants during the first quarter in its Tipping in America 2024 report, against 16% in quick service, while Square reports 15,4% across restaurant transactions in 2024 —one tenth below the 15,5% of 2023—.

Opening: the shift that decides your tips, and with them your turnover

Those points are not earned at night: they are earned at ten in the morning, when somebody confirms the ice machine runs, both stations have their mise en place complete, and the POS carries the day's 86 list before the first guest walks in. A server who starts the shift chasing missing items serves the first two hours worse, and those two hours drag the daily average. Measure the opening against that service's tip percentage, never against a feeling. No benchmark lands the same way across the three sizes, so separate before you compare. Small site, up to 40 tables and a single closing shift: keep 8 items maximum, mandatory photo at two points —freezer and gas shutoff— and drop everything else; your risk is concentrated and your people are the same faces every night.

How to read these numbers in YOUR operation?

Mid-size operation, 80 to 150 covers with two or three shift leads rotating:

now 12 to 14 items make sense, plus a geofence so nobody signs off from the parking lot and a bar count three times a week, because that 10-20% bar shrinkage range Sculpture Hospitality reports is precisely your exposure. Group of five sites or more: stop comparing site against site and compare each site against its own twelve-week average; a forty-site benchmark will tell you everything looks fine while one of yours quietly bleeds. Methodological honesty first: the tipping figures come from aggregated transactional data at Toast and Square covering the US market in 2024, the 55-65% prime cost target is published by the National Restaurant Association as a management range rather than an observed average, and both the 10-20% bar shrinkage and the 4-to-8 monthly turnover are Sculpture Hospitality industry rules built on their client base, which overrepresents operations that already bought inventory control.

Where these benchmarks come from and what you cannot ask of them?

Translation: they fix the order of magnitude and the direction of travel, never the verdict on a single loose month at your site in Medellín or Lima.

Diego F. Parra and the Masterestaurant method use these ranges as an external thermometer, always checked against the restaurant's own series, because a borrowed benchmark applied badly justifies the very mistakes it was meant to correct. Push the counterfactual all the way and you will see why review frequency matters more than the tool. Say you digitize the closing shift, switch on photos and geofencing, and then nobody opens the dashboard until the 30th: your shift lead learns within two weeks that the photo can be of the neighboring cooler, within three that nobody looks at it, and by month two you are paying a subscription to reproduce the paper ritual with better typography. The daily three-minute review —open it, look at four photos, write one line to the group chat— is what holds the whole system up.

What happens if you only review the digital log once a month?

Toast estimates a 4% to 6% annual labor cost reduction from predictive scheduling, and that saving only materializes when shift data is faithful; falsified closing data poisons the scheduling model upstream.

Start tomorrow: pick the four items that move money and demand evidence on those alone. Paper measures intent; evidence-backed records measure execution. That distinction explains why two restaurants running the same printed list end up with night shrink that differs by a factor of three. When the item «freezer closed and at −18 °C» demands a photo of the thermometer with the hour on it, the item stops being an opinion. List length is a cost variable, not a rigor variable. A 40-item checklist produces autopilot ticking — the operator runs a thumb straight down the column — while a 14-item list built around the three or four points that hold the real risk produces verification. Fewer boxes, more control: this is the paradox I find hardest to sell to an owner who is proud of his 60-page manual.

The four differences that move the till

Closing is the shift that decides the most money and gets the least supervision. Between 23:00 and 01:00 you settle tomorrow's shrink, the food safety of everything left in the walk-in and the start of the morning shift, and that is precisely the window when the owner has already gone home. Automation pays for itself right there. AI does not replace the manager's judgment: it clears 22 sheets off the desk so the manager lands straight on the three exceptions. A dashboard that only surfaces what failed turns supervision into a four-minute decision, and that is exactly what lets the operation run without the owner on the floor.