Sponsored by HotSchedules.
Algorithms play an increasingly prominent role in our everyday lives, including restaurant experiences. Artificial intelligence can now be used to process orders via kiosk systems, suggest upsells to frequent guests, respond to customer service inquiries, and even to prepare and plate food in the BOH.
In terms of restaurant sales forecasting, machine learning is changing the game, and like many aspects of AI, the most successful and accurate outcomes are the results of a carefully crafted partnership between manager and machine.
While the restaurant industry has long used “big data” for marketing and research purposes, data alone doesn’t make for the most accurate sales forecasts. Machine learning applications still require local input and constant fine tuning in order to maximize their predictive analytics capabilities. While the machine possesses raw, unemotional, number-crunching power, the manager brings contextual awareness and human intuition that can only be learned on the job. Both bring value to the forecasting process, and both need to operate efficiently in order to most effectively predict future outcomes.
As useful as algorithms are, one of the biggest problems that operators face when it comes to implementing machine learning applications is gaining the trust of the manager. In fact, a study of supply chain companies found that local managers overruled computer-generated forecasts a whopping 90 percent of the time. However, research also showed that when given the opportunity to adjust the forecast instead of simply accepting or rejecting it, managers are less likely to experience “algorithm aversion” and adoption rates tend to be higher.
Clarifi, a cloud-based software platform by HotSchedules, allows managers the flexibility to provide contextual awareness to forecast results, preserving institutional knowledge while highlighting each store’s unique characteristics. Events account for exceptions or atypical occurrences—weather, large catering orders, or construction in front of the store, for example—that can throw off a forecast and cannot be accounted for at the corporate level.
Documented events, along with the raw forecast data, are presented in an easy-to-understand summary so that managers can learn from the forecast as it evolves. Over time, Clarifi learns the seasonal rhythm of the store and spots changes in trends by analyzing new data. As a result, managers are able to better understand how daily decisions impact the top and bottom lines, leading to a greater feeling of success and job satisfaction. This structured method of capturing a manager’s local market expertise and years of experience is invaluable to a restaurant operator’s forecasting capabilities.
When a unified forecast is used to facilitate coordinated labor and product decisions, operators see the accuracy of ordering, production, and waste management improve over time. For example, suggested ordering with clearly stated forecast rationale prevents managers from making assumptions, lowering the potential for waste or under-ordering. Clarifi also helps to more accurately predict scheduling needs, especially important in light of recent legislation in states and locales like Oregon, San Francisco, Seattle, and New York City that requires employers to post schedules 14 days in advance or face potential fines.
As quick serves continue to develop more online, specialty, and marketing options, capturing a location’s historical data becomes increasingly important. Savvy restaurant operators can use systems like Clarifi to make the most of those opportunities.
