May 26, 2015 · 4 minutes

Bloomberg Business has published a fascinating feature about Snapchat chief executive Evan Spiegel's efforts to turn the startup into a functioning company. The entire thing is worth a read, but I want to call attention to one thread in the piece: Spiegel's efforts to build something that isn't controlled entirely by data.

There's a stark difference between analysis and intuition. The former tells you what people think they want; the second allows you to guess what they might really want. While there's a place for both -- it's foolish to completely ignore data and focus exclusively on gut feelings -- many companies favor the former.

Tech companies design their products based on information collected about how consumers use them. Marketers craft their pitches by using the increasing amounts of data they're able to collect. Advertising companies then show those pitches, banner images, and commercials based on their own data analysis.

Snapchat isn't doing that. According to Bloomberg Business' story, the company seems to be avoiding -- or willfully ignoring -- a data-driven approach to its product design, editorial efforts, and advertising platform. Just look at what the story's authors wrote about Snapchat's oft-criticized software design:

Some users also criticize Spiegel for developing a service that’s difficult to use and remains somewhat mystifying to anyone born before 1985. ('The user interface and design looks like the cross between a weird Japanese animation and a 1980s sitcom,' wrote the New York Times.) For example, there are no intuitive buttons, just cryptic icons and swipe gestures that trigger different functions. It’s nearly impossible to search for other users, unless you know their Snapchat names or cell phone numbers. 'I get that it looks different. It looks different because it’s something that is new,' Spiegel says. He says the company could simplify the service and develop such features as a user directory, but he’s more interested in innovating—70 percent of the company’s engineers are working on new products.
Then there's the section explaining how Snapchat handles its Stories, which were introduced earlier this year:
While Facebook and Google focus on technologies that advance material based on what’s popular or useful, Spiegel feels he has a responsibility to show Snapchat’s impressionable young audience things that are meaningful, not just popular. Instead of software decoding a user’s interests from search terms, clicks, and shares, he’s placed a bet on traditional media and old-fashioned editors. Earlier this year he signed up 11 media brands, including CNN, Comedy Central, ESPN, and People magazine, and invited them to contribute daily channels of videos and articles that disappear every night at midnight. 'There’s a sort of weird obsession with the idea that data can solve anything,' Spiegel says. 'I really haven’t seen data deliver the results that I’ve seen a great editor deliver.'
And then there's the company's advertising efforts:
Spiegel’s averse to most kinds of online advertising. He finds targeted advertising creepy, especially the experience of shopping for a certain product on one site, only to later see ads for it on another. 'It’s definitely weird when a vacuum follows you around the Internet,' he says. He’s also ruled out ads on Snapchat that accompany private one-to-one messages between users, judging it too invasive.

Instead, Snapchat started inserting full-screen video ads from such brands as Coca-Cola, McDonald’s, and Samsung into feeds in the media channels and in various 'stories' from cities and college campuses. The ads are about 10 seconds and resemble conventional TV spots, not some novel Internet format. I'm sure these excerpts represent a romanticized view of Snapchat's approach to designing its mobile app, creating an editorial product, and introducing more ad-focused tools as it tries to wean itself off venture capitalists' swollen teats. But it's nice to think, even for a moment, this is how things work at Snapchat.

Too many things are determined by data analysis. Google famously tested 41 different shades of blue for use on its search results page. Facebook shows its users whatever it thinks they want to see. Algorithms, at least when they're working the right way, present us with information that fits our world view.

That's boring. There's nothing worse than giving people exactly what they think they want. You know how that apocryphal Henry Ford quote goes? "If I had asked people what they wanted, they would have said faster horses?" Well, companies are now asking what people want, and we're saying faster horses.

Snapchat seems to be bucking that trend. Even if it doesn't work -- I almost gave up on figuring the app out even though I'm squarely in its target audience, and Bloomberg Business' story makes it clear that interest in the app's stories and advertising platform are waning -- it's nice to know it's at least trying.