How to Cut Labor Costs Without Cutting Service

Labor is usually the second biggest number on a restaurant's P&L, and it’s one of the few metrics you have real control over. So when it starts creeping, the instinct is to act. The response is close to automatic: pull a shift off the schedule, cut people earlier, tighten the hours up. It feels like the responsible move, and it's the one nearly everybody makes.

The trouble is this urgent, corrective action is almost never worth it. You cut a shift, service slows, turns take a little longer, checks come in lighter because nobody has time to sell anything, and suddenly anything you’re saving on labor is offset by lower revenue and returning guests. Sales dropped about as fast as the hours did, The restaurant got worse, and now you've got a crew that feels overworked and taken advantage of.

What's going on in most of these situations isn't really a "cut labor cost" problem, though that's exactly what it looks like on the P&L. It's a deployment problem. The hours usually aren't wrong in total, they're wrong in placement, and no amount of cutting fixes a placement problem. The restaurants that get labor under control without wrecking the guest experience figured this out at some point and stopped asking how many people they could take off the schedule. They started asking which hours were actually producing anything.

Why is my labor cost so high?

Labor percentage is labor divided by sales, which means there are two ways to move it and most operators only ever use one.

Cutting hours shrinks the top of that fraction. It works on paper, which is why it's the reflex. But sales sit on the bottom of the same fraction, and anything that raises sales per hour worked moves the percentage just as well without costing you a thing at the table. That second lever gets ignored almost universally, mostly because it's slower and harder to see on a weekly report.

It helps to think about it one hour at a time. Every hour somebody's on the clock, that hour either produces something or it doesn't. An hour where a server's running four tables and selling a round of after-dinner drinks produces. An hour where that same server is wiping down a section that's already clean, because there's nobody sitting in it, produces nothing at all and costs you exactly the same in wage and payroll tax. Same money, completely different return.

Thus, the useful question isn't how many hours you're paying for. It's how many of those hours are actually spent productively. 

Fixed labor and variable labor aren't the same problem

Before you touch the schedule, it's worth splitting labor into two piles, because they behave differently and they need different fixes.

Fixed labor is what you're paying regardless of how the night goes. The manager on duty, or the lead cook who has to be on that line whether you do forty covers or a hundred and forty. Somebody to open the building and take deliveries. None of that scales down when business is slow, which is precisely why a dead Tuesday does so much damage to your percentage.

Variable labor moves with the volume. The third server, the second bartender on a Friday, the busser you add when the patio opens are all added hours to accommodate higher sales. This is the part of the schedule that's supposed to flex.

Here's where it goes sideways. When the percentage runs high, the cuts land on the variable side, because that's the part that's easy to see on a schedule and easy to change. But if your fixed labor is too heavy for the volume you're actually doing, trimming servers barely touches the number and does real damage to the floor. You end up with a worse restaurant and the same percentage, which is the worst possible outcome.

How do you know which side of the labor equation is more important to address? If labor runs high on slow days but looks fine on busy ones, that's fixed labor, and you're carrying a staffing model built for volume you're not doing. The fix for that lives in your hours of operation, your shift structure, or your management coverage. It does not live in your server schedule, and cutting there will just cost you money in a different column. Finding the correct balance of salaried, fixed labor versus variable, hourly labor is a challenge and is often the biggest factor that affects labor percentage in restaurants. 

A common mistake I want to flag here is trying to force labor to meet your target percentage every hour of every day, every day part, all the time. Many large full service restaurants require a chef, a sous chef, a GM, and an AGM just to start, and this sort of management structure is expensive. If you don’t have the volume to meet your labor target during slow afternoons between lunch and dinner, or if you can’t get Tuesdays below 40%, trying to force it by cutting hourly staff is counterproductive, and will exacerbate revenue problems during slow times, not improve them. It is normal for labor to run slightly over on slow days, then run low on busy days, offsetting the overages. When you’re running significantly high on those slow days, say 50 - 60% or more, your problem is primarily low revenue, and secondarily too much fixed salary labor on the books.  

Where do the wasted labor hours actually hide?

Before making drastic scheduling or operations changes, look at the little things first. Often, a few points of labor overage can be saved with simple organization and better policies. Check here first:

The front end of shifts - Scheduling staff to come earlier than is actually necessary is a common and easy mistake to make, and it adds up to real money fast. Don’t schedule all five of your dinner servers or your three PM line cooks to start at an arbitrary, blanket time - schedule them to what demand actually requires. Otherwise, you end up paying staff to socialize and act busy. 

Just look at the above scenario - instead of those five servers and 3 cooks clocking in at 4:00 PM, we instead schedule to the demand required by both customer volume and prep / setup. Two servers and one cook come in at 4:00PM and are responsible for ensuring a smooth transition between shifts and verifying that all prep and set-up is done. Another cook and two more servers start at 5:00PM. Finally, the last server and cook clock in at 5:30. This saves 3.5 hours of server labor and 2.5 hours of cook labor. If just doing this on slower weekdays, say Monday - Wednesday, you’re saving 10.5 server hours and 7.5 cook hours weekly, which can easily add up to $1000 / month, or $12,000 / year. 

The dead tail - That last hour before close where you've got a full floor running for two tables and a bar guest nursing a beer. Create an efficient cleaning and closing procedure, and you should be able to progressively end shifts in accordance with demand, just like you did for the clock-ins. This can unlock at least another $1000 / month in labor savings.  

Overtime -  Occasional overages driven by real volume spikes are a good problem to have. Overtime that instead shows up every single week is just money you're handing over because the schedule was written carelessly, because you aren’t staffed properly, or because you’re allowing too many of your employees to request off for the same shifts. Look at who's hitting it, and be honest about whether it was predictable the day you wrote it. Be especially mindful when adjusting the schedule in real time. If you need to cover a shift for a sick employee, or you need an extra prep hand or dishwasher on Saturday afternoon, be careful not to inadvertently add overtime dollars to payroll. 

Early clock-ins. A few minutes per person per shift feels like nothing, and on any given day it is nothing. Spread it across a full staff over a year and it turns into a real number. This isn't about squeezing anybody, it's about the schedule meaning what it says. It is important to institute clear policies and expectations that clock-ins cannot occur more than 5 minutes - 10 at the most - prior to the published start time. 

Prep that ends up in the trash. Back of house hours spent making food that gets thrown out is labor you paid for twice, once to produce it and once to buy the product. Bad par levels show up as a food cost problem, but they started as a labor problem. Iron-clad SOPs for how prep occurs need to be designed and enforced to ensure this is avoided. 

Covering call-outs. If one person calling out forces overtime or drops a manager onto the line, that's a cross-training gap billing you as a labor cost.

Not one of those requires removing a position, and every one of them is real money, especially when combined. Investigating and fixing these areas alone is often enough to solve the problem. Always be sure to communicate clearly to your staff when making these kinds of changes - a simple conversation is the difference between your staff understanding the math and buying in, or feeling like they’re underappreciated and mailing it in. 

What actually reduces labor cost without hurting service?

The most important step you can take is to build the schedule using data, designed to meet the demand curve. Stop writing shifts in blocks. Pull your sales by hour, look at what is needed to prepare for those sales and when the work actually needs to start, and start people then. Half the crew at four, the rest at five thirty. Then cut in the same order on the way out, one at a time, as the room empties. For most restaurants this is the fastest money on the table, and the guest never notices, because you're removing coverage from hours that didn't have any guests in them to begin with. Here are some other areas that help: 

Go after turnover. This is the most important point here. Training is a labor cost that never once shows up on your P&L labeled as training. Every new hire requires several technically unneeded shifts while they train. Their work quality then runs slow for weeks and pulls a productive person off their own job to teach them. A restaurant that keeps its people is running a lower labor cost than one that doesn't, even paying the exact same wage.

Write the cut order down. If the closing manager is making a judgment call at nine about who goes home, or the choice is left to the staff, somebody's staying who shouldn't be. A written cut order takes the awkwardness and potential drama out of it, and it's the difference between a floor that clears out on schedule and one that's fully staffed at ten thirty for no reason.

Cross-train. One person who can cover two stations means you need fewer bodies for coverage, which is a very different thing from needing fewer bodies during a rush. It's what lets you absorb a call-out without overtime and run a quiet Tuesday with three people instead of five. This is the highest-leverage item on the list and also the slowest to pay off, which is why almost nobody gets around to it.

Create written, enforced SOPs. A task done to a written standard takes less time than the same task figured out fresh by whoever's on that day. Prep sheets, par levels, opening and closing checklists. They shave minutes off the same job and, more usefully, they make the time predictable enough that you can actually schedule around it.

Look at the menu. Forty items takes more prep hours, more skill on the line, and more waste management than twenty five does. Menu complexity is a labor driver that hardly ever gets counted as one. If you've been meaning to cut the menu down anyway, the labor line is another argument for doing it. Beyond cutting the menu down, look into products your food purveyors have available that can cut down on prep without cutting food quality. 

Which labor cuts always backfire?

Some cuts look good on paper and cost more than they save, and they all fail the same way: they pull capacity out of the hours that actually generate money.

Cutting the host / busser / runner. Support staff are often the most expensive hourly staff in the FOH, and cutting them on slow nights seems like a no brainer. What it actually does is slow your seating, slow your table turns, and create added stress for your servers. It also creates bad first impressions and negatively impacts the all important immersion factor. To reduce labor % on slower nights, you almost always need to raise revenue. Cutting the support makes your customer experience worse and reduces the likelihood that they’ll come back again, which just makes the problem worse.

Widening sections during a rush. One less server sounds efficient right up until every table's waiting longer, the upsells stop because nobody has thirty seconds to spare, and your check average slides. Labor improves on the report and sales fall further, so the percentage lands right back where it started.

Pushing prep onto the line during service. Ticket times go up immediately, food quality suffers - this is the one guests feel faster than anything else on the list. It is also a fast track to killing morale in the kitchen, which is essentially trying to solve one problem by creating a bigger one. 

Cutting ten percent across the board. Trimming everywhere hits your productive hours and your dead hours in equal measure, but the dead hours were the entire opportunity and you only took a tenth of them.

The rule underneath all of it is that understaffing a peak never saves money. Peak is when you make everything you make. Protect it, protect your guest experiences, protect your staff’s sanity, and take the hours out of the parts of the day that aren't producing.

Focus on revenue too 

Everything above works on the top of the fraction. The bottom deserves the same attention and it doesn't cost you anything with the guest.

The same server hour is worth more when the check average is higher, so training on beverage, dessert, and add-ons raises sales per labor hour without adding a single minute to the clock. The same goes for anything that improves turns during a rush, because more covers in the same hours is what better labor productivity actually looks like.

Nobody talks about this version because cutting is faster and easier to point at in a meeting. But an hour that produces more is worth more than an hour you deleted, and it costs you nothing at the table.

The part that makes it stick

Labor control isn't a decision you make once, it's a habit you run weekly.

Write the schedule against the demand curve. Check what actually happened against what you planned. Find where you were over, find where you got caught short, and correct it on the next schedule. Take notes and keep a file that documents your findings and changes over time. The restaurants holding a good labor number aren't the ones that made a dramatic cut last spring, they're the ones looking at it every week and adjusting by small amounts, which is far less painful and works considerably better.

---

Labor % only tells you part of the story, and it’s only useful when you know that percentage accurately and in real time. We have three products available to help streamline your labor. The SHS Operating System lets you track labor, COGS, your custom budget categories and 30 other metrics live. Our Labor Matrix tool let’s you import your roster and create 3 schedules simultaneously with costs broken down in detail, making it easy to game the best schedule for your operation. Finally, let us do some digging with our Health Check analytic and we’ll send you a 12 page report with detailed findings on which areas of your restaurant can improve and how.