How to Check the Accuracy of Restaurant Sales Forecasts
Business Mastery

How to Check the Accuracy of Restaurant Sales Forecasts

Published On: 11 November 2026Written By: Briantama Afiq Ashari

Have you ever prepared ingredients for 200 servings but sold only 130? Or run out of ingredients during a busy service? These situations show why restaurant sales forecast accuracy matters for operational planning.

Sales forecasts do help restaurant owners plan operations. However, these forecasts need to be reviewed regularly to ensure business decisions aren't based solely on assumptions.

For example, a restaurant expects weekend sales to rise around payday, so management increases ingredient stock and plans for additional production capacity.

But fewer customers arrive than expected. When forecasts are too low, the opposite can happen: the kitchen struggles to meet demand, ingredients run out, and waiting times increase.

Therefore, simply making predictions isn't enough. Restaurants also need to compare them with actual results to determine whether the estimates align with actual conditions.

What Is Restaurant Sales Forecasting?

A restaurant sales forecast estimates transaction volume or sales value over a specific period, such as an hour, day, week, or month.

These predictions are usually made by considering previous sales data, customer visit patterns, demand trends, holiday calendars, promotions, and operational conditions.

For example, a restaurant that regularly experiences increased transactions every Friday night can use that pattern as a basis for estimating future sales.

However, predictions aren't exact figures. Weather conditions, events near the outlet, price changes, competitor activity, and even operational disruptions can impact sales results.

Sales forecast accuracy refers to how closely forecasts match actual results. It does not mean every prediction will be exact.

The forecast metric should match its purpose. Use projected portions or menu sales to plan ingredients, and projected revenue in rupiah to plan income.

Read also: How to Create a Restaurant Sales Forecast for Inventory and Production Scheduling

Why Do Forecasts Need to Be Compared to Actual Sales?

A forecast that seems reasonable may not reflect business conditions. Without comparing it to actual sales, management will have difficulty knowing whether the forecasting method used is working effectively.

For example, a restaurant forecasts IDR 15 million in Saturday revenue, but its POS report shows IDR 12 million in actual sales. Management needs to investigate the IDR 3 million difference.

Was the prediction too high because management used weekend data during the holiday season? Was there a decrease in visitors due to rain? Or was the planned promotion canceled?

These questions help management understand the reasons for the differences in numbers. Therefore, evaluation isn't simply about assessing whether predictions were right or wrong. This process also provides an opportunity to learn about changes in customer behavior and operational conditions.

Comparing forecasts with actual sales regularly makes it easier to see whether estimates are consistently too high, too low, or reasonably close to reality.

This information is also useful for decision-making. When forecasts are frequently too high, restaurants can review their ingredient purchasing plans. Conversely, if forecasts are too low, management needs to evaluate staffing needs, kitchen capacity, and inventory.

How to Check the Difference Between Predictions and Sales Results?

Source: iStockphoto

The simplest way to check the accuracy of a prediction is to compare the forecast figures with actual sales reports for the same period. Ensure the data being compared covers a similar scope, for example, the same daily turnover or number of transactions.

The first step is to collect both predicted and actual data. Take the figures set before the sales period begins and compare them with the results from the POS system or previously reviewed sales reports.

Calculate the variance using Difference = Actual Sales − Forecast Sales. For example, if projected revenue is IDR 10 million and actual revenue is IDR 12 million, the difference is IDR 12,000,000 − IDR 10,000,000 = IDR 2,000,000.

Actual revenue is IDR 2 million higher than forecast, which means the restaurant underforecast sales. Conversely, if it forecasts IDR 12 million but earns only IDR 10 million, the difference is negative IDR 2 million.

This is an overforecast because the estimate exceeds actual sales. To measure the size of the error regardless of direction, use Absolute Percentage Error = (|Forecast − Actual| ÷ Actual) × 100%, when actual sales are greater than zero.

For example, if forecast revenue is IDR 10 million and actual revenue is IDR 12 million, the absolute error is IDR 2 million. The absolute percentage error is (IDR 2 million ÷ IDR 12 million) × 100% = 16.67%.

From these results, management can investigate the causes of the increased sales. This could be due to group orders, a successful promotion, or increased dinnertime traffic.

Do not stop the evaluation after one day. Compare several weeks of results so a single unusual event does not distort the assessment of forecast accuracy.

When Should Restaurants Use an Error Metric Like MAPE?

As data volumes grow, comparing individual differences can become cumbersome. Restaurants with multiple locations also require more concise evaluation measures to ensure consistent comparison of predictions.

One useful metric is MAPE, or Mean Absolute Percentage Error. It measures the average absolute forecast error as a percentage of actual sales.

The formula is MAPE = Average [(|Actual − Forecast| ÷ Actual) × 100%]. Calculate the absolute percentage error for each period, then take the average. For example, suppose the errors over three days are 10%, 15%, and 5%. Only include periods with actual sales greater than zero.

MAPE = (10% + 15% + 5%) ÷ 3 = 10%. This means the forecast's average absolute percentage error across the three days is 10%. Restaurants can compare MAPE across periods or outlets with positive actual sales.

For example, outlet A earns IDR 20 million a day while outlet B earns IDR 5 million. Percentage errors help compare their forecast performance without focusing only on nominal differences.

However, MAPE has limitations. This calculation is problematic when the actual value is zero, as division by zero cannot be performed. MAPE can also provide an unbalanced picture when actual sales are low.

In these cases, restaurants can consider MAE, or Mean Absolute Error, which expresses the average absolute difference in the original unit, such as rupiah or portions.

Therefore, the choice of evaluation measure needs to be tailored to the characteristics of the data. MAPE is suitable for viewing the average error as a percentage, while MAE is easy to read when management wants to determine the magnitude of the nominal difference.

Read also: Why Sales Figures Differ Across Restaurant Dashboards, POS Systems, and Banks

How to Improve Forecasts That Frequently Miss the Mark

Source: iStockphoto

After identifying the discrepancy, the next step is to refine the forecasting process. Don't immediately change your forecast just because one day's sales results differ.

First, investigate the causes rather than changing forecast figures immediately. Check historical data quality and make sure sales reports include all transactions for consistent periods.

Missing transactions, cancellations, and entry errors can distort the data used for forecasting. Next, separate sales patterns by day and time.

Monday morning sales aren't necessarily directly comparable to Saturday night sales. Restaurants also need to consider special periods, such as national holidays, Ramadan, vacations, or events that impact customer traffic.

Third, note the factors that cause changes in sales. For example, promotions, new menu items, changes in operating hours, extreme weather, or delivery service disruptions.

These notes help distinguish normal patterns from unusual events. Next, review forecasts by outlet and menu category. Total revenue may match the forecast even when beverage sales rise sharply and main course sales decline.

Looking only at total revenue can hide these differences. Establish a consistent review schedule, with daily checks for operational decisions and weekly or monthly reviews for longer-term patterns.

These steps help restaurants improve forecast accuracy over time. Adjustments should be based on observed data, not guesses about the safest numbers.

Additionally, restaurants need to determine who is responsible for updating data and reviewing prediction results. When the evaluation process has clear responsibility, recommendations for improvement are easier to follow up on.

Restaurant Sales Forecast FAQs

1. How often do sales forecasts need to be evaluated?

The frequency depends on operational needs. Restaurants can review daily results for inventory and staffing needs, then conduct weekly or monthly evaluations to identify error patterns.

2. Do sales predictions always have to be accurate?

No. Predictions are estimates based on certain data and assumptions. What's more important is understanding the error rate and correcting the method when recurring patterns are found.

3. What is the difference between overforecast and underforecast?

An overforecast occurs when the prediction is higher than actual sales. An underforecast occurs when the prediction is lower than actual results.

4. Is MAPE suitable for all restaurants?

MAPE is useful for comparing prediction errors as a percentage. However, this measure is less appropriate when the actual value is zero or very small. Restaurants can use MAE as an alternative, depending on their analysis needs.

Conclusion

Making sales forecasts is an important step in planning restaurant operations.

However, these estimates need to be compared with actual results so that management knows whether the method used is appropriate to business conditions.

Through evaluating variances, calculating percentage errors, using MAPE, and examining the factors causing sales changes, restaurants can gradually improve their forecasting process.

To support this process, restaurants need sales data that is neat and easy to analyze.

OLIN by ESB supports sales forecasting and business analysis using data from F&B operations.

Forecasts should still be compared against actual sales regularly. Contact the ESB team to discuss how OLIN can support your restaurant's sales analysis.

ESB

Ruko Paramount Center II Blok B/8,
Jl. CBD Gading Serpong, Tangerang 15810

150358

(08.30 - 18.00) GMT+7

InstagramFacebookTikTokLinkedInEmail
Copyright 2026 © PT. Esensi Solusi Buana