McDonald’s ice cream machines have a bad habit of being unavailable when you need them most, and now a programmer has reverse-engineered the McDonald’s app to track inoperable ice cream machines across all 14,000 locations in the U.S.
“I’m sorry McDonald’s data analyst I’m afraid I’m ruining your entire mobile conversion metrics for my own personal amusement,” Rashiq Zahid, the programmer, tweeted.
But to be clear, Zahid isn’t actually ordering tens of thousands of dollars worth of McDonald’s ice cream, but rather essentially querying the McDonald’s API for locations where adding ice cream to your order is marked as unavailable. “No restaurant has to actually prepare any ice cream,” Zahid told Motherboard in a direct message.
The site, called McBroken, is pretty straightforward in its breakdown. Of all U.S. McDonald’s locations, 9.3 percent had ice cream machines that weren’t working at the time of the query, with the worst numbers in Seattle (20 percent), New York City (19.57 percent), Washington DC (13.89 percent), Phoenix (13.64 percent), Houston (13.25 percent), and Philadelphia (12.9 percent).
Importantly, most of these machines probably aren’t “broken” per se, but rather temporarily turned off because they need to be cleaned.
“All the machines I list as broken are only unusable for a short while, while they’re getting cleaned,” he explained, adding that he updates the McBroken site when they become available again. “None of the machines are actually broken, but from the customer’s perspective they are because they can’t get any ice cream.”
“Sometimes they’re done cleaning them within an hour or so, sometimes it takes a day,” he said.
The McBroken experiment inspired others to analyze the data themselves and find some interesting things.
One person on Twitter ran the McBroken data for Houston through the Urban Institute’s Spatial Equity Data tool—created to help assess how unequally resources were distributed across cities and locales—and found McDonald’s locations were overrepresented in white areas while locations with broken ice cream machines skewed Black and low-income.
This is interesting considering it has long been documented that poorer areas tend to have limited access to healthy food, not only because of efforts by fast food chains but the federal government’s own subsidies.
Another longer analysis by Ajjit Narayanan, a data analyst at the Urban Institute, covered even more cities and found more or less the same patterns. In Chicago, McDonald’s locations are underrepresented in low-income, Black, and Latinx areas, but broken machines are overrepresented in Black and renter-heavy neighborhoods. In New York City, McDonald’s restaurants are overrepresented in Latinx and renter communities, while broken machines are overrepresented in Black, high disability rate, and low-income communities.
McDonald’s did not respond to Motherboard’s request for comment.
Although focusing on McDonald’s ice cream machines is a bit silly, the data paints a grim reality for Black, Latinx, and low-income neighborhoods.