How to Make Demand Forecasting Your Competitive Advantage

• This article was previously published on RestaurantDive.

In the dynamic and ever-evolving world of the restaurant industry, one key challenge persists decade after decade: accurately forecasting demand. Stuck using traditional forecasting models that primarily predict demand based only off historical sales, operators are all too often left to grapple with frequent scheduling errors and costly inefficiencies. But why?

Seeing into the future is no easy task, even for the most experienced operator. Not to mention, forecasting demand and creating schedules is only one of the numerous administrative tasks GM’s take on in addition to their primary responsibilities. But after years of minimal advancements in predicting demand, AI-powered forecasting is ready to transform this challenge into your competitive advantage.

What is AI-powered forecasting?

The standard practice of forecasting relies on POS integrations to bring in historical sales data for the same time period. Then, managers make adjustments based on known internal and external factors like an average of a 5% uptick in sales from last year or an unexpected summer storm making the patio unusable. Unfortunately, this leaves the forecast susceptible to guesswork and creates a terrible trickle effect of ineffective weekly schedules, unstable labor budgets and dissatisfied employees and guests. But by leveraging the computing power of AI, every manager can start their day with an accurate forecast, rather than calculating one.

While there are various models AI can use to forecast demand, they all begin with proper data collection. Fourth, for example, has compiled the largest repository of restaurant data from serving more than 100,000 locations and 14 of the top 20 US chains. With this data-rich depository at its disposal, the AI forecasting algorithm can perform an in-depth analysis of intrinsic and external data points including historical sales data, guest counts, in-store promotions, holidays and more. Then, the algorithm begins data preprocessing, or the process of cleaning data by detecting and removing event outliers from the formula. From here, the AI algorithm works behind-the-scenes to train itself by forecasting demand and analyzing its accuracy based off actual results. With each new data input it receives, the algorithm continues to refine the formula, making small but mighty daily improvements.

Improved accuracy, however, should be an assumed benefit when switching to an AI-powered forecasting model. Let’s explore how AI forecasting can improve your performance, profits and people when used specifically for labor modelling.

Optimize performance with demand-driven scheduling

AI-driven forecasting should deliver a unique demand forecast for every location, equipping all frontline managers with predictive, actionable insights irrespective of their tenure or overall experience. This creates a huge competitive advantage for growing chains and established enterprise brands alike, ensuring all locations and operators begin schedule creation on a consistently solid foundation. Going one step further, AI forecasts should directly feed into your scheduling solution, automatically creating house shifts that align with the forecasted demand and your restaurant’s labor rules. Now your forecast not only predicts demand with improved accuracy, but automatically optimizes your schedules to match demand so guests are happy and labor budgets are stable.

Using AI forecasting to directly influence scheduling can also help restaurants better overcome labor shortages without hiring more employees. How? By ensuring you deploy the right people at the most optimal times. For example, let’s say your sales forecast varies drastically within 15-minute increments between your lunch and dinner rush. By feeding this data into a smart scheduling solution, your managers should receive the best recommendation for scheduling employees to meet varying demand, lowering your risk of lost sales or overtime pay. “Best” recommendation is key, here. Managers may be able to find an “accurate” answer on their own, guessing when to cut staff. But AI coupled with advanced scheduling solutions will automatically provide the best business outcome.

Eliminate under and overstaffing to take control of your labor budget

Ranging from 25-35% of sales, labor is not only one of the largest expenses a restaurant faces, but often the most unstable due to staffing issues. Understaffed shifts mean hefty overtime fees, burnt out employees and missed sales opportunities. In fact, according to Fourth’s internal data, restaurants lose an average of 10-25% of gross revenue per location when not staffed properly to meet demand. On the flipside, overstaffed shifts mean wasted labor dollars and unhappy servers.

The best attribute of AI forecasting is its ability to continuously learn and improve its algorithm, creating a highly personalized forecasting model executed at scale. Overtime, this model can systematically eliminate inefficiencies like under and overstaffing. With the ability to create unique algorithms for every location, GM’s will have all the necessary insights to build schedules based off demand, consistently deploying the right people at the right time, every shift. And when you staff to meet demand, you not only regain control of your labor budget location by location but will also see improvements in overall sales and customer satisfaction rates. Best of all, this can all be realized without requiring additional manpower.

Give managers the gift of time

In an industry agnostic transformation, organizations across the United States have begun calling their HR teams “People Operations.” Signifying a prioritized effort to better meet the rapidly evolving expectations of employees, People Ops is about ensuring your employees have the necessary tools, resources and support to develop their wellbeing and competencies. And while you may initially relate People Ops initiatives to providing in-demand employee benefits like earned wage access or tuition reimbursement, it should also extend into your tech stack. The best restaurant technology should not only contribute to your revenue and performance goals, but your retention goals as well.

1 in 2 workers in the food service and hospitality industry quit because of burnout, while 48% of food service managers’ report feeling burnt out on a daily basis. AI forecasting can greatly reduce the number of administrative burdens falling on frontline managers. When used in tandem with automated employee scheduling tools and labor optimization, managers can create fully optimized and compliant schedules in minutes rather than hours. This means less time in the back office and more time pursuing the work they love like creating memorable guest experiences or strategic promotions.

This drastically improves the hourly employee experience, as well, effectively eliminating burnout from understaffing or “bore-out” from overstaffing. And as all restaurant operators know, employee engagement and happiness directly influence retention rates and guest experiences.

According to the National Restaurant Association’s 2023 State of the Restaurant Industry report, more than 40% of operators are planning to invest in technology this year to help combat rising costs. While there’s a plethora of technology, in and out of the world of artificial intelligence, I highly encourage operators to take control of their labor budgets, empower frontline managers and conquer every shift with AI-powered forecasting.