Sales forecasting for next week, in twenty minutes

Sales forecasting is the habit of estimating, before the week begins, how much the venue will sell and how many guests it will serve. It sounds like a job for an analyst with a spreadsheet full of formulas. For a single restaurant, it can be done in about twenty minutes on a Thursday or Friday, with four inputs most venues already have: what happened last year, what happened in the last four weeks, what is already in the diary, and what is expected that week, such as the weather and local events.

A forecast matters because almost every decision for next week depends on it. How many staff to schedule for each shift. How much to order from each supplier. How much to prepare each morning. Whether to promote a quiet night. Without a forecast, those decisions are made from memory or habit, which usually means overstaffing quiet shifts, understaffing busy ones, and either running out or throwing food away.

This guide sets out a simple, repeatable method for sales forecasting for next week. It covers the four inputs, how to combine them into a covers forecast and a sales figure by day and service, how to account for seasonality and special events, how to measure forecast accuracy and improve it, and how to use the forecast for staffing, ordering and preparation. It is practical demand forecasting for a venue that does not have time for anything complicated.

A forecast turns next week from a guess into a plan. With a figure for each day and service, the manager can build a rota against it, the chef can order and prepare against it, and the owner can see whether the week is likely to be strong or weak before it happens.

It does not need to be perfect to be useful. A forecast that is usually within ten or fifteen per cent of what actually happens is far better than no forecast at all. The aim is to be roughly right, consistently, and to get a little better each week.

The general practice of demand forecasting covers sophisticated statistical methods. A restaurant can get most of the benefit with a simple method done every week.

Forecast two things: covers and sales. Covers, the number of guests, drive staffing, preparation and the shape of service. Sales, the money taken, drive ordering, cash planning and the week's financial picture. Forecasting both, and checking that they make sense together, catches mistakes.

Forecast by day and by service. A single figure for the week hides the difference between a quiet Tuesday lunch and a busy Saturday dinner, which is exactly the difference the rota needs. For most venues, a figure for each lunch and dinner service, or each meal period, is the right level.

Keep it in one simple sheet, with a row for each service of next week and columns for each input.

The first input is recent trading: what happened on the same day and service over the last four weeks. Take the covers for each of the last four Tuesdays' dinners, for example, and average them. That average is a baseline: what a typical recent Tuesday dinner looks like.

Using four weeks smooths out a single unusual night, such as a large party or a rainy evening, while staying recent enough to reflect current trading. If one of the four weeks was clearly unusual, such as a holiday, leave it out and average the other three.

The baseline is the backbone of the forecast. The other inputs adjust it.

The second input is last year: what happened in the same week last year, compared with the weeks around it. This captures seasonality, the predictable rise and fall of trade through the year. The run-up to the festive season, the summer slowdown, the school holidays, Ramadan in the Gulf: each has a pattern that recent weeks alone will not show.

A simple way to use it is to compare last year's figure for next week's equivalent with last year's figure for the four weeks before it. If next week last year was about 10 per cent busier than the four weeks before, adjust this year's baseline up by a similar amount.

Be careful with dates that move. Holidays, school breaks and Ramadan fall on different dates each year, so compare the right weeks, not simply the same calendar dates.

The third input is the diary: the bookings already in for next week. For venues that take bookings, this is often the most powerful input, because it is a direct signal of demand. If Saturday already has more bookings than a typical Saturday at this point in the week, Saturday is likely to be busy.

To use the diary well, you need to know how bookings usually build up. If a typical Saturday has 70 per cent of its final covers booked by the Thursday before, and next Saturday already has 60 covers booked, you might expect around 85 covers in the end. The pattern of how much notice guests usually give is what makes the diary useful.

Add an allowance for walk-ins, based on how many walk-ins each service usually sees, and for no-shows and cancellations. Our guide to a covers forecast your kitchen will believe covers this in detail.

The fourth input is what is coming next week that the numbers do not yet show. The weather, especially for venues with terraces or those that depend on passing trade. Local events: a concert, a match, a conference, a market. Roadworks or closures nearby. A promotion or event at the venue itself. A competitor's opening.

Adjust the forecast for each, using judgement and past experience. A sunny weekend might lift a terrace venue's covers noticeably; a major match might fill a sports bar and empty a fine dining room. Note each adjustment and its reason, so you can check later whether it was right.

Combine the four inputs into one figure for each service. A simple approach is to start with the four-week baseline, adjust it for seasonality using last year, then compare the result with the diary and adjust towards whichever is more reliable for that service, and finally apply any adjustments for what is coming.

For example, for next Saturday dinner:

The numbers are illustrative. The point is a clear, written chain from inputs to forecast.

Once you have a covers forecast for each service, multiply by the expected spend per cover for that service to get sales. Use the recent average spend per cover for each service, from the last four weeks, adjusted if something is different, such as a set menu or a price change.

Spend per cover often differs between services: lunch guests may spend less than dinner guests, and weekend guests more than weekday guests. Using a separate figure for each service makes the sales forecast more accurate.

Check that the result makes sense. If the forecast for a Tuesday lunch is higher than the busiest Tuesday lunch you have ever had, look again.

Seasonality is the biggest reason a forecast based only on recent weeks goes wrong. A venue coming out of a quiet summer into a busy autumn will under-forecast if it looks only at the last four weeks; one heading into a quiet January after a busy December will over-forecast.

Keep a simple record of how each week of the year compares with the weeks around it. After a year or two, that record shows the venue's own seasonal pattern, which is far more useful than any general rule.

In the Gulf, seasonality often runs differently: busier cooler months, quieter summer months and the shift in trading hours during Ramadan. Our guide to Ramadan restaurant hours and the service that fits covers how the month changes the day, and our guide to mapping your own restaurant year covers how to build a seasonal picture from your own figures.

Forecast accuracy is how close the forecast was to what actually happened. Measure it every week: for each service, compare the forecast with the actual covers and sales, and note the difference as a percentage. Over several weeks, the average difference shows how accurate the method is.

Look for patterns. If Tuesdays are always over-forecast, the Tuesday baseline or adjustment is wrong. If weekends are always under-forecast when the weather is good, the weather adjustment is too small. Each pattern is a chance to improve.

Do not expect perfection. Restaurants are affected by many things no forecast can see. A forecast that is usually within ten or fifteen per cent is doing its job.

The most immediate use of the forecast is the rota. With covers forecast for each service, the manager can decide how many servers, cooks and other staff each shift needs, rather than copying last week's rota. Quiet services get fewer staff; busy ones get more.

Our guide to scheduling from forecast, not from memory covers how to turn the forecast into a rota, and our guide to labour cost covers how to keep the cost of the rota in proportion to sales.

The forecast also guides ordering and preparation. A chef who knows that Friday and Saturday will be busy, and Monday quiet, can order produce and plan the prep list accordingly. Ordering against a forecast reduces both running out and waste.

Our guide to the prep list that ends the morning guesswork covers how to plan preparation from expected covers.

Holidays and one-off days break the usual pattern, and they are where forecasts most often go wrong. A public holiday can make a Monday behave like a Saturday, or empty a business district completely. A festival, a national day or the first night of a new season can do the same. The four-week baseline knows nothing about them.

For each holiday or special day, look at what happened on the same occasion last year, rather than the same calendar date, and compare it with the weeks around it. Note whether it fell on a weekday or a weekend, because that changes its effect. Build a small list of the venue's special days and what each usually does to trade, and add to it every year.

When in doubt on a special day, forecast a range rather than a single figure, and plan staff and preparation so the team can flex up or down.

A single figure is convenient, but a range is often more honest. "Saturday dinner: about 95 covers, likely between 85 and 105" tells the manager and the chef more than "95". The range reflects how much the service usually varies, and how uncertain this week is.

Use the range for planning. Schedule staff for the middle of the range, with a plan to call someone in or let someone go early if trade lands at either end. Prepare for the middle, with a pad for the most popular dishes. Order with the range in mind, especially for ingredients that spoil quickly.

Over time, forecast accuracy shows how wide the range needs to be for each service.

Some decisions need more detail than others. The rota needs covers by service, and sometimes by hour, because staff start and finish at different times. Ordering needs covers by day, and the expected mix of dishes. Cash planning needs sales by week.

Build the forecast at the most detailed level you need, usually by service, and add it up for the others. For staffing by hour, use the shape of a typical service from recent weeks: how many guests usually arrive in each half hour. Our guide to sales by hour covers how to read that shape.

Keep every forecast and every actual result in one place, week after week. That log becomes the most useful document the venue has for forecasting. It shows how accurate the method is, which services are hardest to predict, how holidays and weather affected trade, and how the venue's seasonal pattern really looks.

Note the reason for every adjustment beside it, and whether it turned out to be right. After a few months, the notes become a guide to which adjustments to trust.

A forecast that lives only in the manager's sheet helps nobody else. Share it with the people who act on it: the chef for ordering and preparation, whoever writes the rota, and the owner for the financial picture. A short message on Friday, with covers by service for next week and a note of anything unusual, is enough.

Share the results too. When the team sees that the forecast was close last week, they rely on it; when it was wrong, the reason helps everyone plan better. A forecast that the team trusts changes how the whole week is run.

A forecast made on Thursday for the following week will be out of date by the following Wednesday. Check the diary each day and adjust the forecast for the next few days if bookings are running well above or below expectations. A sudden rush of bookings for Friday is worth acting on before Friday arrives.

Keep the original forecast as well as the updated one, so you can see how much each changed and learn from it.

The whole process can fit into about twenty minutes a week:

Here is an illustrative forecast for a venue's week, to show how the pieces fit. The figures are examples.

servearound does not produce the forecast for you, but it holds the numbers the forecast is built from. Reservation analytics shows covers day by day and by meal period, covers by day and hour across the week, bookings by source and party size, and how much notice guests give, which is the booking pace a diary forecast depends on. The book itself shows what is already booked for next week, and the waitlist shows walk-in demand on busy nights.

Payment analytics shows revenue by hour and by day, and the dashboard shows spend per cover, so covers can be turned into sales. Everything is read on the venue's own clock. See Analytics and the reservations module.

Within ten or fifteen per cent on most services is a good working standard for a single venue.

No. A simple sheet and the venue's own figures are enough. Software helps by keeping the figures in one place.

Next week in detail, updated during the week. A rougher monthly view helps with bigger decisions.

Rely more on the four-week baseline, last year and what is coming, and track how weather and events affect trade.

Start with a cautious estimate, measure every service, and build the baseline as weeks of real data arrive.

One person, usually the manager, sharing it with the chef and whoever writes the rota.

Covers first, by service, then multiply by spend per cover. Covers drive the rota and the kitchen; sales follow from them, and checking one against the other catches mistakes in both.

Sales forecasting for next week can take twenty minutes. Combine four inputs, the last four weeks, last year for seasonality, the diary and what is coming, into a covers forecast for each service, then multiply by spend per cover to get sales. Write each step down, so the chain from inputs to forecast is clear.

Measure forecast accuracy every week and improve the method where it is consistently wrong. Use the forecast to build the rota, order and plan preparation, update it during the week as bookings change, and treat demand forecasting as a weekly habit rather than a project.