Restaurant Staffing and Scheduling Basics

Front of house (FOH) is everyone the guest interacts with directly: hosts, servers, bartenders, bussers, and often an expo person calling out tickets at the pass. Back of house (BOH) is the kitchen: line cooks, prep cooks, dishwashers, and whoever is running the pass when that role isn't split out separately. The split matters because the two teams scale on different drivers. FOH scales with covers, meaning how many guests you are seating and turning through the dining room in a given period. BOH scales more with menu complexity and volume together, meaning how much of the menu is made from scratch and how many of those items are moving through the kitchen at once.

Priya opens a 60 seat New American restaurant with a six item entree list, most of it built from a handful of proteins and a rotating vegetable side. Her BOH is lean: two line cooks and one prep cook can run a full Friday night. Two miles away, a 45 seat brunch spot with a 20 item menu of scratch made hollandaise, pastries baked in house, and a dozen egg preparations needs three line cooks and two prep cooks to hit that same 60 covers, because the kitchen's workload is set by the menu, not the seat count. Both restaurants need roughly the same FOH headcount for that many covers, because a table is a table no matter what's coming out of the kitchen. This distinction earns its keep the moment you sit down to build a labor budget or write a job posting. Treating "restaurant staff" as one undifferentiated line item hides where the money is actually going and makes it hard to answer a simple question: is this Friday's overtime a FOH problem or a BOH problem. It also shapes how you think about cross training. A server picking up bussing shifts is a small stretch. A server picking up prep cook shifts is a different job entirely, and scheduling as if it isn't sets someone up to fail.

These are starting points, not rules. A scratch kitchen with a long prep list needs more BOH labor than a menu built around a few high quality proteins with simple sides, even at identical covers. Track how long your actual team takes to prep for a normal night, and let that set your real ratios.

Put numbers on it. An 80 seat full service Italian concept with a wood fired oven typically runs four line stations (grill, saute, pasta, pizza) plus two prep cooks and a dishwasher on a Friday night. A 40 seat fast casual sandwich shop covering that same 80 guests over a lunch rush can run on one line cook and one prep cook, because there is only one real station and most of the menu is assembled rather than cooked to order. Same covers, very different kitchens.

The single biggest scheduling mistake in restaurants is treating every day like it has the same demand curve. It doesn't. Most restaurants have a small number of shifts that generate a disproportionate share of weekly revenue, and a schedule that doesn't flex around that guarantees a bad labor percentage on the slow shifts.

Marisol runs a 50 seat Mexican restaurant and pulls ten weeks of hourly POS data before building her next schedule. Tuesday lunch averages $850 in sales across that window. Saturday lunch averages $2,150. Tuesday is running at about 40 percent of Saturday's volume, so instead of scheduling the same five servers on both shifts out of habit, she schedules two servers and a food runner on Tuesday and keeps the full five person crew for Saturday. Her Tuesday labor cost as a percentage of sales drops from an uncomfortable 38 percent to a much healthier 24 percent, without cutting a single shift she actually needed.

Redo this exercise every few months, not once and never again. A slow Tuesday lunch that turns into a busy one after a new office building opens down the street will punish a schedule built on outdated historical data just as surely as one built on no data at all.

Everything above assumes you already have weeks of sales history to schedule against. A lot of real decisions happen before that data exists: a new restaurant's first month, a new shift you've never run before, or a concept that simply doesn't have eight clean weeks of numbers yet. Use the branches below to find a reasonable starting point, then correct it against real sales as soon as you have any.

Most of the labor problems above show up as one of a small handful of recurring mistakes. None of them are exotic. They're habits that feel reasonable in the moment and only look wrong once you compare them to what the sales data or the schedule itself is actually saying.