A restaurant manager in Chicago rewrote his entire menu in an hour based on what was selling well. A restaurant in Dallas knows which servers need help with wine knowledge. This all happened because of big data.
Big data, broadly defined, is pure information. It is the mass of messy bits and pieces of information that have become a flood through the operations of restaurants across the nation. It’s the details of every sale, the specifics of inventory, the money spent on overtime, the components of an online reservation, the very minutes a guest sits at a table. It’s overwhelming. It’s unwieldy. It’s also extremely powerful.
A 2012 study by market intelligence firm IDC illustrates just how bonkers the issue of data is becoming. The study looks ahead to 2020, when analysts expect the digital universe to encompass 40,000 exabytes (40 trillion gigabytes). That’s up from the study’s starting point of a mere 130 exabytes of data generated in 2005. This works out to 5,200 gigabytes for every person on Earth in 2020. All you really need to know is: that’s a lot of data.
Dan Simons, partner in the Founding Farmers farm-to-table restaurant in Washington D.C. has his own definition of big data. “Big data is all about tiny pieces. It’s what I can understand about the inch-by-inch, the second-by-second, the minute-by-minute behavior or results that our guests have at the table based on who is serving them, how we’re handling them and how we have a relationship with those guests.” It’s not about the mounds of data so much as it is what can be done with it once it’s analyzed. Here’s how restaurants are using big data to learn more about customers, customize menus, measure server productivity and keep a handle on labor costs.
Dan McGowan, president of the eight-unit, Chicago-based Big Bowl chain of Chinese and Thai restaurants, was contemplating the cost of fried rice in his restaurants. He thought $1.25 was a little high. He wanted to back it down to 99 cents, but that price reduction would need to be balanced out somewhere else. He found the answer in big data. By running the sales numbers in analytics program Avero, he found the perfect complement to the lowered rice price was a slight increase in the price of the restaurant’s custom-made premium ginger ale. “I can run a number that I can never run in my head. In five minutes, I can answer my question and I’m able to make a decision,” says McGowan.
The Big Bowl experience is an example of how much information can be gathered, sorted and used from a POS system. Your POS already knows what is selling well, how often items are ordered and what the prices are. “Guest check analytics is a burgeoning big data area. You can start to do some pretty good tuning about what your menu should look like. You can also look at what price points people are buying at and adjust your prices,” says Chuck Ellis, president of restaurant market research company Restaurant Sciences.
Since most restaurants are already running POS systems, it’s a natural way to get started with big data. Many of the big data software and service packages geared for restaurants rely on connecting to POS systems. Many POS providers now offer options to generate customized reports based on POS data, so a restaurant can handle tasks like evaluating historical sales data, judging the success of special promotions and comparing sales between different units. “There are so many great tools that are now built into a POS, as well as off-the-shelf tools for managing recipes and inventory. It’s very accessible to the independent operator,” says Ellis.
Every morning, McGowan gets up and spends a few minutes looking at sales data for all his stores, comparing them to the previous week and the previous year. If a new menu item like sushi is being tested at a particular restaurant, he will check the sales numbers for that product in particular. Having ready access to all this information, broken down into digestible bits, allowed him to tackle an entire menu rewrite based on what was selling well. “It was an exercise I could do out of my house in an hour on a Saturday afternoon drinking a cup of coffee,” says McGowan. Before big data programs came along, he would have had a team of accountants trying to amass and break down all that data. “Now I can make a decision in a day or two, rather than weeks,” he says.
Reducing labor costs
There’s a fine line between overstaffing and understaffing and it costs money when you wander too far in either direction. Restaurants amass a history of valuable labor data every day: labor costs compared to sales, the amount of overtime paid in a week, the breakdown in costs between FOH and BOH.
Sometimes big data services can be found tucked into existing labor management services. Service providers ShiftNote and the POS-integrated version of HotSchedules, for example, include overtime alerts and the ability to create schedules based on sales forecasts. It’s about being able to step beyond the present and use data to look ahead, schedule accordingly and keep costs under control.
Once upon a time, the Spillers Group’s three restaurants in Dallas relied on sharing PDF file spreadsheets with their managers in order to keep them in the loop on how much each store was spending on labor. These files were quickly out of date and couldn’t help the managers with really digging into the details of their labor spending. Knowing his managers were hooked on using their smartphones, Spillers Group co-founder Shane Spiller went in search of a mobile app that could help. He found Roambi.
Roambi isn’t specifically made for the restaurant and hospitality industry, but it fit the bill for solving the Spillers Group’s problem. Now, all that labor data is fed into Roambi, which translates it into graphs and charts and automatically shares the latest data with all the managers. “Anything that is measured can be improved,” says Spillers. “You can quickly identify an anomaly if you had an excess period of overtime. It’s visually jumping out at them from their mobile devices.”
The app has helped the restaurant group decide when to hire extra help and when to stick with overtime. It is also a key part of a bonus program for the managers. Reaching certain labor cost benchmarks results in tiered bonuses, encouraging smart management of the labor pool. The Spillers Group saw a 10 percent drop in labor costs within the first few weeks of using the program.
Evaluating the effectiveness and productivity of servers has historically been a fuzzy area of restaurant operations. Big data offers a new way to look at server performance. On a basic level, a POS system can be used to gather information on a server’s check average, but that’s only part of the story. Tip amounts can help measure customer satisfaction. The amount of time spent at a table can tell you if a server is rushing or taking too long. The challenge is in automating and accessing this data.
Boston startup Objective Logistics takes an unusual approach to this issue. It gathers data from the POS system and breaks it down into reports for each server. This gives managers a lot of information to go on when evaluating servers, but there’s a twist. It brings in the element of competition by allowing the high-scoring servers to choose their own shifts. Essentially, it uses data to turn serving into a game-like pursuit geared for motivating servers into better performance. It’s an intriguing enough idea to net the young company $5.3 million in startup funding this summer.
Kent Rathbun Concepts encompasses three different concepts and a catering company stemming from Dallas celebrity chef Kent Rathbun. Director of operations Matthew Scott uses Avero Slingshot to generate server reports. These reports can be likened to a baseball card, in that they contain statistics like the check average and how much of each kind of item is sold.
Every Friday, server reports are posted for everyone to see and discuss. “We use it as a tool to learn where we can get someone better, what it is they’re not doing as well as their peers. We’re able to mentor and teach employees individually,” says Rathbun. If a server isn’t selling much wine, that person can be identified and given extra training on understanding and promoting the restaurant’s wine offerings. Servers with good track records are enlisted to role-play and demonstrate how they handle these tasks successfully. In turn, the whole team is strengthened.
Scott’s biggest tip for using big data to evaluate servers is to keep the approach positive. It’s not about taking under-performing servers to task; it’s about discovering what areas they need extra help in and providing the appropriate training. “Keep it positive and keep it educational. You won’t effect change if you make it a negative. It has been very successful for us,” he says.
Let’s say you’re one of Founding Farmers’ top customers. You come in often. You bring friends. You buy wine and dessert. You’ve been busy and you haven’t stopped by the restaurant in a couple months. One day, an email arrives. It politely inquires how you’ve been and lets you know that your favorite Grey Goose cocktail is waiting for you on the rocks for when you come back. There’s a very personal touch to the message. Smiling, you click over to OpenTable and make a reservation for the weekend.
On the surface, this is a nice example of personalized customer service. Behind the scenes, big data tools did most of the heavy lifting to make it happen. One of the components is a service called Swipely, which replaces a restaurant’s traditional credit card processing provider, usually at no extra cost and with no new hardware requirements. It works with existing POS systems to gather information about the guest check. It identifies repeat customers, new customers, what they ordered, how long they were at the table and how much they tipped. All that minutia is fed into a system that gathers the information into a dashboard. From that dashboard, operators can monitor overall sales, track marketing campaigns and view individual customer profiles.
Founding Farmers gathers data together from Swipely, OpenTable and analytics service Avero Slingshot. “We look to build our profiles on each guest so we know what their favorites are, the booze they like to drink, the food they like to eat and how they like to be served,” says Simons. He then manually builds profiles of his top-100 guests within the reservation system. “You can have really targeted reach-outs. We look to communicate with them in a highly personalized way,” he says.
Some restaurants also use Swipely’s built-in loyalty program to gather even more detailed information on customers. This works best for restaurants that don’t have a high percentage of cash sales. “For the restaurants we work with, over 70 percent of customers pay with a debit or credit card. We’re trying to tap into the data behind all the transactions,” says Swipely founder Angus Davis. Founding Farmers is 90 percent credit cards and 90 percent reservations, so this system of combining information from Swipely and Open-Table works.
Big data is just starting to filter into the hands of smaller operators in the form of services that take all data and shake and bake it down into manageable information. Simons, who currently puts his customer profiles together manually, is hoping to see greater convergence among providers. “It can be overwhelming for us because I’ve got three different software providers and they’re not all fully integrated,” he says. “For there to be widespread adoption, it has to become simpler and easier.”
Perhaps the best way to wrap your head around big data is to break it down into small data. It’s about controlling your food costs, fixing the flaws in your labor scheduling, or measuring the success of a new menu item. It’s like cutting a steak. You don’t try to down the whole rib eye in one mouthful. At this point, the challenge for small operators is to just get started. “Imagine if every night, Amazon deleted their database. This is what happens every day in restaurants,” says Davis. “If you’re not using this data, you’re not competing.”
Get started with big data
You don’t have to wrap your head around the entire concept of big data in order to make use of it. The best way to get started is to cut it down into smaller pieces. “Today, the problem is too much data and not enough time. The number of data sources that are being created is just breathtaking. My customers are feeling completely overwhelmed,” says Damian Mogavero, founder and CEO of restaurant data analytics company Avero. “Big data in and of itself is not a good thing. It only works if you transform it.”
Start small. Right now, there is no single big data solution that can swoop in, analyze everything from your POS to your social media hits to your cost of goods and then tell you what to do. What is available are services that address more specific needs. “A restaurant operator shouldn’t just jump into big data unless they have a problem they are trying to solve,” says Benjamin Stanley, co-founder of Food Genius, a big data restaurant menu trends company. For example, Food Genius tracks menu trends and mines the data for insights into what is popular and what is falling out of favor. This sort of service makes sense for larger chains that need to stay a step ahead of the competition when it comes to designing new dishes or adjusting existing menus.
The advent of big data services geared for small businesses is in its infancy, but some powerful tools are out there. Chuck Ellis, president of restaurant market research company Restaurant Sciences, suggests approaching big data services the same way you would shopping for any other service. Talking to colleagues in the industry about what works for them can be a good start. “If you spend most of your time and energy in finding the right few partners for these different services, you’ll be okay,” he says.
Big data on a bigger scale
Zaxby’s, a chain of nearly 600 chicken restaurants, is big on LTOs. A recent special, a banana pudding milkshake, was scheduled to disappear, but a look at sales data convinced the company to put it on the menu full-time. “Analyzing the sales results helps us make good business decisions,” says Frank Knight, senior director of information technology. When chicken wing prices went through the roof last year, Zaxby’s was able to respond by promoting other products. Once again, the prompt for this move came from big data.
Zaxby’s uses two different POS systems. Most locations use Micros and a few use POSitouch. Data from those systems is fed into Cognos, a business intelligence package from IBM. Cognos can analyze data and generate detailed sales, food cost, labor cost and financial reports. Zaxby’s employs a financial analyst who works with the program and shares performance reports with top-level management. It’s that big data information that leads to decisions like the one to place the hot-selling banana pudding milkshake on the permanent menu.
At the regional and store level, Zaxby’s managers have access to mymicros.net, an analytics component available with Micros POS. This online business intelligence tool consolidates financial information. “District managers have the ability to set up their own reports. They may be focused on labor one month and food costs the next month. They can use reporting as their needs change,” says Knight.
Though mymicros.net can make data available within 15 minutes of it being generated, Knight sees managers using it more to analyze larger time periods, such as weekly sales. “We have district managers who get up first thing in the morning and see how stores did the prior day from a cost, sales and labor standpoint,” he says. It helps them quickly spot and fix problems on the store level. For example, if sales are slow at a certain time of day, but labor costs are high, it could lead a district manager to work with a store manager on refining schedules in order to find a better balance.
Smaller operators may not need a package as powerful as Cognos, or be able to employ a full-time financial analyst, but that doesn’t mean big data is out of reach. “Know your size. Know who you are,” says Chuck Ellis, president of restaurant market research company Restaurant Sciences. “If you’re a mom-and-pop, you’re not going to go out there and try to find an enterprise-class solution. Do what makes you comfortable.”