By Dipock Das, VP of Technology at HotSchedules
The irony of “smart” technology is its stupidity under the wrong conditions. Smart devices are only as intelligent as the data they rely on, and no industry has had more difficulty capturing data than the restaurant industry. The “Holy Grail” of big data has evaded most restaurants because years ago, software vendors didn’t anticipate its importance. As big data became a big deal, restaurant software became antiquated overnight, yet vendors failed to adapt. They “chose poorly” to quote the knight from Indiana Jones and the Last Crusade. Restaurants are now hamstrung by their inability to collect and use data in a strategic way.
My goal here is to explain why big data has been such a struggle in the restaurant industry, and I will discuss how we can blaze a new trail to the grail. At this point in the quest, the key is that restaurants need to capture as much data as possible from a variety of sources. As restaurants start using artificial intelligence to make sense of data, they won’t just use it to find past mistakes – they will use it to prevent mistakes in the first place.
Restaurants have long relied on legacy software that is expensive, runs on Windows PCs and feels daunting to replace. So, restaurant management and IT staff are trapped — mentally and technologically — within the constraints of that software. Consequently, restaurants are in the dark about what they could do if they weren’t tethered to Jurassic technology.
For instance, most restaurant owners know that Yelp matters. As one Harvard Business professor found in 2011, a one-star increase on Yelp can lead to a 5-to-9 percent increase in revenue for a restaurant. However, there is lack awareness that Yelp data can be correlated with a restaurant’s operational data to identify what managerial choices lead to negative or positive reviews.
The limits of outdated technology shape the mindset of restaurant management. The first step is to stopping asking, “What can our tech do?” Instead, ask, “What do we want our tech to do?”
For the brave restaurants that pursue big data with legacy software, a world of frustration awaits. Integrating data from disparate, ’90s-era software programs is cumbersome at best and a fool’s errand at worst for several reasons. First, collecting data is much trickier that most people realize. It can actually take weeks to get data from old restaurant software to a data warehouse where it can be aggregated and analyzed.
Second, even if a restaurant gets that data under one roof, it still needs data scientists to make any sense of it. To use the same example, finding correlations between Yelp reviews and a restaurant’s operational choices requires a ton of analysis. Among staffing, inventory, menu selection and marketing alone, dozens of variables could influence a Yelp review to various degrees. Only the largest restaurant chains can afford to hire data scientists, and most expect Wall Street salaries.
After a restaurant spends tens of millions of dollars to collect and analyze the data, that’s not the end of the road. How do they push that data to someone who can act on it — immediately — when the only software at a restaurant’s disposal is stuck on a PC that sits in the back office? Most likely, the data will populate pretty PDF and Excel reports that arrive via email once a month, long after the event when the data could have changed business outcomes. At best, big data with legacy software can provide 20/20 vision looking backwards.
The New Way to the Grail
Adding data scientists and expensive integrations to legacy software is a bit like dropping a Ferrari on an air mattress and claiming you made a speed boat. You can put all the tech you want on top of primitive software, but that won’t change its nature. So today, the restaurant industry is shifting towards mobile technology that is designed to both collect and receive actionable data. The next wave of software will apply artificial intelligence (AI) to big data in order to solve restaurants’ most vexing operational challenges.
At present, three categories of data can feed smart devices and eventually AI: 1. in-store, 2. above store and 3. near-store data.
In-store data is every byte (pun intended) you can pull from the restaurant — data from your POS system, drive through, labor management and inventory system, to name a few sources. Above store data is cumulative in-store data viewed at different scales. It could compare data on a store-to-store, city, state, regional, franchisee, national or perhaps international scale. Last, near-store data, includes weather, traffic, demographics and of course Yelp. It’s about the environment in which restaurants operate.
The true Holy Grail is AI technology that can look at all this data, learn from it and make recommendations that influence a restaurant manager’s decisions in the moment, when there’s opportunity to change business outcomes. AI could learn that there is a correlation between the Wednesday evening shifts and low Yelp ratings. Perhaps it finds that the restaurant schedules two less servers than usual since guest traffic is lower at that time. The AI calculates that on a two-year scale, the cost of two additional staff members is ten times cheaper than the negative effect of these low star ratings, projecting that Wednesdays continue to generate them. Problem solved. Having learned from one restaurant’s experience (using in-store and above-store data) the robot now pushes this insight to other restaurants (above-store) before they suffer the same mistake.
Although it seems strange that an industry run on paper and PCs could soon rely on artificial intelligence to make decisions, this isn’t farfetched. The restaurant industry hasn’t progressed incrementally from “dumb” to smart tech, like many others industries. Restaurants will instead benefit from years of innovation in other industries to steal a few technological bases. If restaurant tech can deliver data-driven recommendations to a smartphone resting in the pocket of a person who takes action immediately, the technology is indeed “smart.”