As discussed in a previous post, when using big data to formulate conclusions about the world at large, one must always consider what and who is being left out. I had a dig around GrubHub's site and found a number of potential caveats to its position as an authoritative voice on the "gender wars" of America's eating preferences:
- GrubHub acts as an intermediary between consumers and restaurants offering takeout ONLY. It is not involved in transactions between consumers and grocery stores, farmers markets, and non-takeout restaurants. To be fair, GrubHub doesn't claim that its data represents anything other than takeout dining.
- GrubHub has 4.57million “Active Diners”, so less than 2% of Americans. To engage with GrubHub, customers must have Internet access so its user base will skew white, young, educated, wealthy, and urban.
- GrubHub’s analysis is based on a year’s worth of GrubHub orders. At no point in the current ordering process is the customer asked to confirm their own gender (possibly assumed from the customer's first name) or the gender and number of people that will actually be eating the food that is being ordered.
- GrubHub offers its customers a choice of 30,000 takeout restaurants but its partners do not seem to include the top ten takeout chains (McDonald's, Subway, Starbucks, Burger King, Wendy’s, Taco Bell, Dunkin' Donuts, Pizza Hut, KFC, or Applebee’s).
Image from Wikipedia