Chicago startup Food Genius has been mining data from restaurant menus for the last two years hoping to plot out the changing appetites of the country. The idea was to help big food companies like Kraft identify taste trends for their product development teams, so they could get that new Navajo green chili and chicken frozen entrée to market at the peak of a Southwestern cuisine fad rather than at its tail end.
There was only one problem. The data they found was not what what they expected to find.
After aggregating menu data for more than a year, Food Genius wasn’t able to find any big nationwide food or eating trends. No matter what it looked at – the appearance of kale on the menu, the number of restaurants serving carnitas tacos – Food Genius was just seeing big flat lines across its graphs. The problem, according CEO and co-founder Justin Massa, was that, averaged out across the country, our eating patterns stayed static or rose and fell only incrementally quarter to quarter, year to year.
Justin Massa (Source: Food Genius)
For a company trying to make new product decisions – “Should we put jalapeno in our hot sauce or habanero?”– that data was pretty worthless, Massa told me in a recent interview.
“If you try to look at food as a big national trend, you’ll find nothing,” Massa said. “But what’s interesting is there’s dramatic changes happening in individual markets if you dig down into that data. … We originally thought of our service in terms of product development, but we have found our sweet spot in sales and marketing.”
On a state-by-state, market-by-market and cuisine-by-cuisine basis, Food Genius is able to pick out surprisingly detailed insights, Massa said. A company may need to develop a product for national consumption, but in the sale and positioning of that product it’s able to get fairly granular. So instead of losing Kraft as a customer, Kraft started using Food Genius as a market intelligence tool. As did 30 other companies, ranging from national grocers like Safeway to restaurant chains like Arby’s and DineEquity (which owns Applebee’s and iHop) and restaurant distributors like Reinhart Foodservice.
Food Genius’s new database interface
What are those companies doing with that data? It depends on their business, Massa explained. A food product maker might use it to target its advertising budget in cities where it’s witnessing certain patterns. For instance, a city like Los Angeles may be going through a big Neapolitan pizza craze, making it an ideal place to market a company’s thin crust “brick-oven” pie line. A distributor might see that hickory-smoked bacon may be showing up on a lot on burgers at a lot of tony gastropubs and then, armed with that data, try to sell the same kind of bacon to its fast-casual restaurant customers.
There doesn’t have to be direct correlation either, Massa said. If there are a lot of Laotian restaurants popping up in one city, that doesn’t mean a restaurant group should open their own Laotian restaurant in the same town. It’s probably already oversaturated with Laotian food. But that restaurant group could use Food Genius data to determine what other cities have similar demographic and eating patterns, allowing it to kick off the Laotian dining trend in new markets.
With that in mind, Food Genius has plans to bring in other data sources to its platform, from U.S. census data to social media sources. The idea is to be able to take the 1 billion “food concepts” – a kind of meta-dish generated from its data — and correlate them across socio-economic data, as well as to compare how food is really consumed against the “buzz” that food generates on the internet.
Despite the initial failed foray into food product development, Massa said the company is doing well in its new niche. Its customer base is growing, and it expects to turn an operating profit this year. But the whole experience seems to have made Massa retrospective.
He got into this business to identify the big trends in foods, but in the process he discovered there that there are not only no big trends when it comes to eating, but also many of the commonly accepted trends aren’t backed up by data. For instance, new food concepts don’t start on the east and west coast and then move into the country’s interior like everyone seems to assume. There are a few isolated instances of this happening, leading to the illusion of a trend. But for the most part, food gets popular at different times in different places almost randomly when you look at the data collectively, Massa said.
The culinary world is looking for “a theory of trend,” Massa said, but none appears to exist.
Related research and analysis from Gigaom Research:
Subscriber content. Sign up for a free trial.
What the fourth-quarter 2014 meant for tech buyers
Applying lean startup theory in large enterprises
Important notes for IT decision-makers from the fourth-quarter 2013