You might think that when a company the size of Facebook is gearing up to launch its first mobile ad product, it would have a large, organized team vetting each and every aspect before launch. But the origins of Facebook’s first mobile ad format, which promotes other apps and directs users to their install pages, were a little less coordinated.
In the beginning, the team consisted of just Deborah Liu, the product manager who helped launch Facebook Credits, and one engineer. By the time the feature first launched in October, it had added one more engineer.
Facebook’s app install ad would eventually become a success story for the social network. Between April and June of this year, more than 8,400 advertisers purchased the install ads, which together drove 46 million app downloads. But for a while, it looked like a flop. “We kept fixing things, and we were like, this isn’t working,” Liu told Fast Company. “The numbers were okay, but we just couldn’t get past it. We kept fixing things, saying, this is going to be the thing that takes us over the top, and it didn’t work.”
With only two engineers, the team had to prioritize potential solutions. A small bug that disappeared the ads after users scrolled passed them wasn’t high on the list. Unfortunately, it was also the problem.
The team discovered the bug in November and fixed it in December. “The minute we fixed that, everything changed,” Liu says. “Because it turns out that user behavior is very different than we expected.”
Instead of interacting with stories when they first see them, as the team expected, it turns out most people browse through their News Feeds before returning to the content they find interesting. With the bug, users who may have downloaded an advertised app couldn’t find it when they scrolled through their News Feeds a second time. “Once we fixed that, it was very clear that this was going to be a really big business and this was going to be good for our developers,” Liu says.
Facebook has a habit of constantly updating live features. When the company launched graph search, for instance, its users until after a design revamp. Most recently, it re-ranked the News Feed based on data from experiments. Trial and error and making fixes on the fly are all a part of solving problems. As Liu says, “You never know what it is that is going to take you over the top.”
[Image: Flickr user Scottydelmonte]