When the labor line looks heavy, every instinct — and every scheduling tool — says the same thing: cut hours. And the easiest hours to justify cutting are the expensive ones, which are the busy ones. That's the trap. Cut a quiet Tuesday and you save a little. Cut the Friday peak and the ratio improves while the till quietly takes the hit.
The arithmetic of a peak-hour cut
Drop one server from the three-hour Friday rush at 25 an hour fully loaded: 75 saved, guaranteed. Now the floor runs slower. Suppose the thinner team turns four fewer tables across the rush — two covers each at a 45 average, that's 360 in revenue that never rings. At a 70% gross margin, you gave up about 250 of margin to save 75 of wages. The worst part: your labor percentage *improved*, because the denominator shrank along with the cost.
That's the mechanism researchers call denominator endogeneity: understaffing the peak suppresses the very sales the ratio divides by, so the cut flatters the metric while destroying contribution. The clean evidence comes from retail, where the data is easier to isolate — one study of 41 stores found every single one understaffed its three busiest hours, losing around 8.5% of peak sales, and traced the cause to budgeting labor as a percentage of sales. Companion work found a dollar of well-placed payroll returned several dollars of sales, and that simply moving existing hours to the right times lifted revenue 2–3%. Different industry, identical mechanism — a restaurant's peak is where the same logic bites hardest, because a table not turned on Friday night doesn't come back on Tuesday.
Where the real savings live
- Dead dayparts — hours where even the minimum crew can't earn its own wage bill; shorten or close them (the margin-per-hour test finds them)
- Opening-hours surgery — a small venue commits half to two-thirds of its labor just to being open; the rota follows the opening hours, so that's the lever that moves it
- Cross-training — a team where the bartender can plate and the runner can pour lowers the minimum crew a service needs, which is the floor under every quiet shift
- The mix, not the rota — if a busy shift earns badly, the problem is usually what it sells, not who's working it
Peaks pay for everything
At Chat Noir, my club in Geneva, a few hundred covers cross the register on a normal night and the peaks go well beyond. Those two or three loud hours pay for every quiet one — the last place I'd economize is the crew that keeps them moving. When the labor percentage looks heavy, the honest question isn't "who can I send home Friday" — it's "which hours shouldn't exist at all."
The full labor cost guide covers the ratios; the short version is that a percentage is a rear-view mirror. Steer with margin, cut the hours that earn nothing, and protect the ones that pay for the rest.
