How to reduce restaurant labor cost without hurting service

A daily labor percentage can look fine while a few hours a day cost too much. The fix is rarely a whole shift. It's usually the hours after the rush, carried by the rush's crew.

Labor is the cost a store can change fastest. A schedule is rebuilt every week, so a better one shows up in the next week's numbers. It's also the cost that can do the most damage when cut badly, because a short crew loses sales that no report records. This guide covers how to build a schedule that fits the work, how to find the hours where it doesn't, and how to make changes without paying for them in service.

The numbers to use

Schedule by the labor the GM can move. That's hourly labor from the POS, not the fully loaded figure on the P&L, which includes salaries, payroll taxes and benefits. Hourly vs P&L labor explained.

• Hourly labor % = hourly wages ÷ net sales × 100.
• Sales per labor hour (SPLH) = net sales ÷ labor hours worked in the same period.
• Labor hours needed in an hour = forecast sales for that hour ÷ SPLH target.

Sales per labor hour is the better scheduling tool, because it doesn't move when wage rates change. Hourly labor percentage tells you what the week cost; SPLH tells you whether the schedule fit the work. Never divide labor cost by labor hours to get an average wage: salaried hours are often in the hours but not the cost, so the answer comes out too low.

Build the schedule from a forecast

1. Forecast sales by hour from the same weekday over the last four weeks, adjusted for anything known: a holiday, a local event, a promotion.
2. Set a sales-per-labor-hour target from your own best weeks. Menus and formats differ too much for a published number to fit.
3. Divide forecast sales by the target to get labor hours for each hour.
4. Set a floor: the minimum people needed to open safely and run each station, whatever the forecast says.
5. Stagger start and end times so people arrive for the rush and leave as it ends.

The forecast says the store needs nine people for one hour and three by mid-afternoon. A schedule built in whole shifts, with everyone from 10 to 4, would carry nine people through all five hours: 45 labor hours for work that needs 28. Staggered starts and shorter shifts close most of that gap. Some of it can't close, because shifts have minimum lengths and people need breaks, but the forecast shows where the hours are going.

Find the overstaffed hours

Compare people on the clock with orders, per hour, over four weeks, split into weekdays and weekends. A run of hours carrying the peak hour's crew for under half its orders is the gap. In one illustration: about three people from 3 to 7 PM for nine orders an hour, against 2.9 people at noon for 27.

Leave out catering and scheduled orders placed on another day. They don't need someone at the counter that hour, and counting them hides the gap.

Look at weekdays and weekends separately, and at each daypart on its own. A Saturday afternoon can be busy while Tuesday's is empty, and averaging them hides both. Four weeks is usually enough to see a pattern without one odd day deciding it. If a gap appears in only one week, it's probably a one-off, such as a call-out covered by a double, rather than a schedule problem.

Provo's daily labor percentage looked normal, because a busy lunch covered for the afternoon. The cost only shows up when the day is split by hour. The fix is one shift ending at 3 instead of 7, four days a week, not a cut to the lunch crew.

Other places labor goes

• Overtime: paid at a higher rate, so a few hours of it cost more than the same hours spread across the crew. Look for people near 40 hours by Thursday.
• Clock-in and clock-out drift: early punches and punch-outs well after a shift ends add minutes every shift. Set a rule for how early people can clock in.
• Training hours: necessary, but schedule them in slow hours and track them separately so they don't hide in the daily number.
• Turnover: every new hire carries training hours and slower shifts. Keeping a good crew is a labor saving. How to retain restaurant managers.
• Prep timing: prep done during the rush pulls people off the line, so more get scheduled. Move prep to the quiet hours.

Check the rules first

Some states and cities have predictive-scheduling or fair-workweek laws that can require schedules to be posted in advance and extra pay for changes made at short notice. Rest-break, minor-labor and overtime rules also vary by location. Check what applies to your stores, and build changes into next week's schedule rather than today's.

Don't count understaffing as savings

Cutting too far costs sales that appear in no report — the car that left the drive-thru line. Try a change for two weeks and watch orders and ticket times in those hours before calling it a saving. If orders fall or wait times climb in the hours you cut, put the hours back. A schedule change that saves $240 a week and loses $400 in sales is not a saving.

A weekly labor checklist

1. Pull last week's hourly labor and sales per labor hour by hour, for each day.
2. Mark hours where crew outnumbered the work, using people on the clock against orders.
3. Build next week's forecast by hour from the last four of the same weekday.
4. Apply your SPLH target and the store's floor to set hours for each hour.
5. Stagger starts and ends around the rush, and check overtime exposure before posting.
6. Check the schedule against local scheduling, break and minor-labor rules.
7. Post it on time, then compare actual hours with scheduled hours each day.
8. Review orders and ticket times in any hours you cut after two weeks.

Common mistakes

• Judging labor by the daily percentage. It hides the hours that cost too much.
• Cutting the rush to fix the afternoon. The rush is where the sales are.
• Using a published SPLH target instead of the store's own best weeks.
• Leaving catering and scheduled orders in the hourly order count.
• Changing schedules at short notice where local law requires notice or extra pay.
• Calling a cut a saving before checking orders and ticket times.
• Dividing labor cost by labor hours to get an average wage. It comes out too low.

Where Franlever fits

Franlever compares crew on the clock with orders by hour at each store and flags runs of hours where the gap is large. It measures each store against its own normal range and the brand's median, and when a gap is worth fixing, AI COO sends it as that store's one lever for the day, to the GM by email or text. It never prices a staffing change as a saving, because understaffing costs sales no report shows. How AI COO works.

General information for restaurant and franchise operators, not legal, tax, accounting or financial advice. Figures in examples are illustrations, not predictions or promises of results. Laws and platform policies change and differ by location, so check the current rules and talk to a qualified advisor about your situation.