It’s one of the strangest contrasts in today’s tech world. Mobile hardware keeps advancing screamingly fast, with each new smartphone model made obsolete by a faster, cooler successor within months. And the number of mobile apps keeps multiplying—it’s surpassed 1 million in the Google Play store and is approaching that number in Apple’s iTunes App Store. But the mechanisms for finding new apps seem frozen in 2007.
Or maybe a better comparison would be 1997— before the rise of Google, when there were already a couple million sites on the Web, but the only way to find the one you needed was to use a directory site or a primitive, keyword-based search engine. Today, discovering cool new apps for your smartphone or tablet is still disturbingly similar to plowing through endless lists at Yahoo or Altavista. Almost no one bothers to do it, which means apps that don’t get picked for the human-curated “Best New Apps” section of iTunes or the “Play Picks” section of Google Play stand almost zero chance of hitting the big time.
That’s the problem Cole Ratias hopes to fix. He’s the founder and CEO of Loytr, a San Francisco startup that makes a social-media app for iOS devices called MyPad. The app, which I first wrote about a couple of years back, has millions of users, who like it mainly because it lets them browse and post on Facebook, Twitter, and Instagram without having to switch between apps. But an updated version of MyPad, released by Loytr over the weekend, includes an important new feature: it gives users the ability to browse iOS apps based on information about the apps their Facebook friends are using. It’s an interesting example of “social app discovery,” and it could become the true killer feature inside MyPad—while at the same time showing other mobile developers a way out of the app-discovery mess.
“The goal of MyPad has always been to provide the best social context on these devices,” Ratias says. “Initially, that meant Facebook and then Twitter and then Instagram. But now, since none of those do a good job on mobile of helping you find out what apps your friends use, we have taken the next step.”
I’ve written about the whole app-discovery crisis before, in a November 2011 story about Chomp. The San Francisco startup had come up with a better way to index and search apps than the mechanisms built into Apple and Google’s own app stores. It also invented a card-based interface for browsing app search results that was way more useful than what iTunes and the Android Market offered at the time. When Apple bought Chomp the following spring, I hoped it might be a harbinger of big changes in iTunes, and indeed, the new version of iTunes that came out around the same time as the iPhone 5 incorporated a lot of Chomp’s technology, including the sliding app cards.
But what still hasn’t improved much is the quality of search results within iTunes, which depend greatly on whether you can come up with the right search terms. “Generic search results for apps are great, if you can describe what you are looking for, but in most cases, people can’t,” observed Hooked Media CEO Prita Uppal in a September 2012 commentary in these pages. Nor was Uppal impressed by the “Genius” feature in iTunes, which tried to find apps similar to the ones users already had installed. “By pulling recommendations based on all the apps that are on our devices, even the ones that we never touch, Genius often just piles on bad recommendation after bad recommendation,” Uppal wrote. (Apple has since scrapped the Genius feature, replacing it with a location-driven “Popular Near Me” option that is also of dubious utility.)
Why hasn’t anyone solved the app discovery problem before now? Because it’s devilishly difficult—more difficult, in some ways, than the Web search problem that Google co-founder Larry Page set out to solve with his PageRank algorithm in 1998. You can evaluate the importance of a Web page simply by counting how many other pages link to it, then counting how many pages link to those pages, etc. But apps live in their own universe, largely separate from the signals that might be used to rank them, such as Web links, Likes on Facebook, or +1’s on Google Plus.
In the new version of MyPad, Ratias’s team is starting to bridge that gap. The only reason they have any hope of doing so, of course, is that MyPad has a lot of users—it’s been downloaded 60 million times and has 12 million registered users, with roughly 3 million of them coming back at least once a month. It also knows who those users are connected with on Facebook. With a sample that large, it turns out that Loytr can gather useful data about which apps are hot.
It works like this. In the new version of the app, clicking the “apps” section in the toolbar brings up an app window with four tabs: Recent Apps, My Apps, Friends’ Apps, and Popular Apps. On the My Apps page, which shows all the apps installed on your device, you can tell MyPad which apps you like best. At the same time, Loytr’s algorithms analyze your app lineup and gather other social and behavioral signals. If you opt into participating, Loytr sucks up that data, mashes it together with data from your friends and other MyPad users, and publishes anonymized app rankings in MyPad to people you know from Facebook, Twitter, and Instagram. It shows up on
