McDonald's: Big Mac + Big Data

McDonald's Acquires Dynamic Yield

McDonald's has acquired AI startup Dynamic Yield for $300 million. I'm particularly excited about this deal because it signifies that finally, we are inching ever closer into a world where firms can efficiently and effectively squeeze out more of a customer's "willingness to pay" (this pleases the economics major within me) based on data-driven decision making (this pleases the statistics major within me).

For starters, let's look at what Dynamic Yield does (I'm betting all of us already know what McDonald's does, so I won't go over that — let's ignore that McDonald's isn't really a fast food chain, but actually a real-estate company. But that's a story for another day). From Dynamic Yield's website, they "help [firms] tailor personalized experiences based on any data [they] own." Experiences can be tailored according to weather, geography, what platform a consumer is using, and much more. Wired Magazine gives an example of how this would look like in real life:

"Here’s what that looks like in practice: When you drive up to place your order at a McDonald’s today, a digital display greets you with a handful of banner items or promotions. As you inch up toward the ordering area, you eventually get to the full menu. Both of these, as currently implemented, are largely static, aside from the obvious changes like rotating in new offers, or switching over from breakfast to lunch.

But in a pilot program at a McDonald’s restaurant in Miami, powered by Dynamic Yield, those displays have taken on new dexterity. Algorithms crunch data as diverse as the weather, time of day, local traffic, nearby events, and of course historical sales data, both at that specific franchise and around the world. In the new McDonald’s machine-learning paradigm, significant display real estate goes toward showing customers what other items have been popular at that location, and prompting them with potential upsells. Thanks for your Happy Meal order; maybe you’d like a Sprite to go with it."

The core idea is that everyone has a different level of desire for different McDonald's products, and that these desire-levels are dependent on many factors. For example, if it's a very hot day, we may have an increased desire for Sprite or McDonald's ice cream. At noon, there is probably more demand for burgers than at 8am (that's why McDonald's already does alter some of its offerings to account for this, such as by having separate breakfast and lunch menus). What this means is that offering one flat, static menu is inefficient. Instead, McDonald's would be able to optimize their profit-making capabilities by catering to those factors that determine desire, or, to put the word more technically, demand.

I usually wouldn't buy the frappe, unless it's a hot day. Missing out on my increased demand on hot days means that McDonalds's is missing out on revenue opportunities.

Theoretically, a firm is able to make as much money as a consumer is willing to pay. We've seen that this willingness to pay can fluctuate. Thus, it would be suboptimal not to capture the value presented in times where consumers have an increased willingness to pay. What Dynamic Yield does is allow McDonald's (and other firms) to better estimate what that level of willingness to pay is, and then change their product offering to that customer accordingly. In a simple application, if Dynamic Yield sees that the temperature is 90 degrees Fahrenheit, then it could infer (quite accurately) that my willingness to pay for ice cream and smoothies has increased, so that on those hot days, when I go into a McDonald's, I'll have a ton of ice cream and smoothie products pushed upon me. Obviously, the applications get much more complex than this (McDonald's doesn't need an AI to tell them hot weather means people want more ice cream, but there probably exist a ton of other strong correlations between external factors and demand that aren't so obvious to the naked eye and require tools like Dynamic Yield to detect). It's a big data-driven approach that will allow McDonald's to sell more Big Macs.

This is part of a larger trend. For almost the entirety of history, large firms have only been able to give fixed offerings to a vast population of different consumers. Now, firms are becoming ever more capable of providing offerings that cater individually to each of those different needs. This allows for better, more personalized service for those individuals, and for the firms to squeeze out every last bit of an individual's "willingness to pay." The personalization revolution is well on its way.