Last week, when this movie theater candy map suggested that Wisconsinites prefer Kraft Singles above all else, lots of people laughed. Some Wisconsinites defended themselves, arguing that they had never seen a cheese slice at the movies before, while others suggested they were just at the wrong theaters.
A few raised eyebrows at North Dakota’s preference for Baked Beans, while a handful of others argued that surely the mapmakers meant Boston Baked Beans the peanut-based candy. No one pointed out that neither Wal-Mart nor Target have released Halloween candy data this year. In fact, most commenters on Lights, Camera, Pod’s parody map just debated their own state’s preferred candy.
This very silly map highlights a very serious problem: we rely on maps to give us the truth, but maps can be as unreliable as any other data source. Just ask Africa.
We can’t resolve centuries of imperialism today, so let’s stick with candy to address two problems with how viral maps get made.
Search frequency ≠ popularity
Let’s start with this map of each state’s favorite Halloween candy.
From a usability perspective, there’s a lot to like about this map. For example, the use of simple colors and branding makes it easy to see that the most beloved Halloween candy is a Reese’s Peanut Butter Cup, and that I’m never trick-or-treating in North Dakota, Nebraska, or Oklahoma.
I find my home state’s preference for Mars bars inexplicable. This makes me want to tell other people that Ohioans are insane and shout to the rooftops that clearly the best Halloween candy is a thin box of Good & Plentys.
Although you may have been too busy adding Swedish Fish to your shopping list to notice the attribution, this map comes from a company called Bid-on-Equipment. While it’s not unheard of for used machinery trading sites to run best-of candy lists, it’s certainly unusual, and worth reflecting on for a moment. So instead of asking why Bid-on-Equipment got your state’s favorite Halloween candy wrong, it might be better to ask why the company created the map in the first place.
The genius of this map is that it inspires both love and hate, and both love and hate inspire social media sharing. I’m guessing Bid-on-Equipment has seen a nice traffic bump this month. Some of those candy lovers and haters might even have used machinery buying and selling needs.
The downside of the map is the slim research supporting it. Here’s how BoE summarizes its research methods: “Using the Google AdWords platform, we analyzed search volume trends for more than 100 different types of candy, over the period of September 2018 to October 2018 in all 50 states and the 20 largest cities in the country.”
To summarize the summary, BoE checked how often each state googled “Peanut Butter Cups” (and 99 other types of candy) and used the results to determine how popular each candy was.
People running searches for particular types of candy could be doing so for lots of reasons. Perhaps they’re curious, as I was, about butyric acid in Hershey’s chocolate. Perhaps they’re searching for information about a candy recall. Perhaps they want to know the correct pluralization of Good & Plenty. The fact that someone has searched for a candy name doesn’t mean they like that candy.
Bulk purchases ≠ household purchases
If you don’t like the map you’re using, get a new one.
This map has a few things going for it. It’s produced by CandyStore.com, a company we can assume has some candy-industry expertise. It’s interactive, so you can mouse over your state to see the runner-ups. That feature let me see that my state’s preference for M&Ms, Starburst, and Blow Pops means my trick-or-treaters are going to be really disappointed with the candy offerings at my house.
Instead of using search data, Candystore.com tracked candy sales from 2007-2018. This approach seems more sensible because while searches of candy could be about anything, sales seem pretty unambiguous.
Unless you hate the neighbor kids. Or maybe you don’t even hate them, but you don’t want to shell out for the top-shelf candy. As a wise commenter on CandyStore.com’s website pointed out, candy that you buy isn’t necessarily candy that you like — you may be buying the candy that you can afford in order to feed a whole neighborhood, or you may be buying the kind you’re less likely to wander away from your desk and eat.
But even assuming that people are buying their favorite types of candy, CandyStore.com may not be the best indicator of what Halloween candy individual households are buying. The site specializes in bulk candy sales, which suggests that the clients buying from Candystore.com are not the same clients buying candy for trick-or-treaters. Or maybe they are? Are other houses on the street buying a lot more candy than me? Should I be buying more candy? CandyStore.com doesn’t sell Mars Bars, so maybe a 20-pound case of fun-size M&Ms? A 500-count mix of Pop Rocks? A black-and-orange candy buffet?
Candy buyers ≠ candy eaters
The maps are there to make us talking, and they work. But that highlights one more small problem. The people googling and buying candy aren’t necessarily the people eating the candy. If we wanted the most accurate Halloween candy map, we’d need to start polling kids.
Maybe 538 should get on that. They don’t have anything else going on this month, do they?