How to Forecast Sales to Schedule Staff

To forecast sales for scheduling, use your historical sales by day of week and daypart as the baseline, then adjust for known factors — weather, events, holidays, and reservations. That forecast becomes the demand curve you staff against, so you schedule to how busy you'll actually be instead of guessing. A good forecast is the single most valuable input to controlling labor cost.

Why forecasting comes first

You can't build a cost-effective schedule without knowing how busy you'll be. Staff to a guess and you're either overstaffed (wasted labor) or understaffed (slow service and lost sales). The forecast turns scheduling from a favor-juggling exercise into a demand-matching one.

Build the baseline from history

1. Pull sales by day of week. Mondays behave like Mondays; Saturdays like Saturdays. Average several weeks per day.
2. Break it into dayparts. Lunch, mid-afternoon, dinner, and late each have their own curve. Staff the curve, not the day total.
3. Look at guest counts and check averages, not just revenue — labor tracks covers more than dollars.

Adjust for what history can't see

The baseline is the starting point. Then adjust for:

• Reservations and large parties on the books.
• Local events, holidays, and paydays that push demand up or down.
• Weather — a patio-dependent restaurant lives and dies by it.
• Trends — is this month running above or below last year?

Turn the forecast into a staffing plan

Once you have forecasted sales by daypart, apply your labor standards: how many servers, cooks, and support staff each level of volume needs. That converts the forecast directly into a headcount and shift plan you can schedule to.

The forecast is where labor control is won or lost. Everything downstream — the schedule, the labor percentage, the service level — flows from how well you predicted the day.

Let the forecast build automatically

Manual forecasting from spreadsheets is slow and gets skipped in a busy week. Cobblestone POS has your complete sales history by day and daypart, so it forecasts demand automatically — and its AI assistant Daisy can turn that forecast straight into a draft schedule with a projected labor percentage. It's a free, all-in-one platform for independents, a reason owners leave pricier systems like Toast ($470+/month) that charge extra for forecasting and scheduling. Accurate forecasts also feed your prep list and ordering.

The bottom line

Scheduling without a forecast is guessing with your payroll. Build a baseline from your own history by day and daypart, adjust for events, weather, and reservations, and translate the result into a headcount plan. Do that consistently and you'll staff to real demand — protecting both your labor budget and your service on the nights that matter most.

Frequently asked questions

How many weeks of history should I use? Enough to smooth out one-off spikes — often four to eight weeks per day of week, weighted toward recent trends and adjusted for seasonality.

Should I forecast dollars or guest counts? Both. Guest counts drive labor needs more directly than revenue, so forecast covers by daypart and cross-check against sales.

What's the biggest forecasting mistake? Using a single daily total instead of dayparts. You staff the shape of the day, not one number, so a lunch-heavy and a dinner-heavy $5,000 day need very different schedules.