How does Apple fix a broken App Store?

The App Store isn’t broken. At Apple’s Worldwide Developer Conference (WWDC) earlier this month, it announced that there are now 1.2-million apps in the store and that users visit it 300-million times every week. To date (since July 2008), 75 billion apps have been downloaded by iOS users. That makes it the most successful (by far) software market in history.

Except, the App Store is broken. Benedict Evans, now partner at venture capital outfit A16Z, repeats this mantra often: mobile discovery is in (what he terms) the ‘pre-pagerank’ phase. I argued back in October on this site that app discovery on both iTunes and Google Play is dead.

Apple’s App Store-related announcements at WWDC this year were largely expected. It wasn’t going to do nothing. And some of the additional features have been trialed already (like video previews). It’s spent some time working on integrating app discovery in Spotlight search (ie. outside of the actual App Store), there’s additional focus on surfacing popular (trending) apps, editor’s choice adds another much-needed level of curation and app bundles solve a pain for developers in a rather eloquent way (and get customers to buy more apps!).

But, all of these changes are iterative. Human-powered curation at the scale of a 1.2 million-strong app catalogue is ambitious at best. How much curation can Apple realistically do?

None of this human-intervention solves the top charts/what’s hot/what’s new dilemma. Search in the App Store is a mess (and increasing the number of categories – as has been argued – does not improve the experience substantially).

In April, developer (and podcaster) David Smith made a number of detailed, well thought-through suggestions to improve the App Store. But (despite being thought through), some of these are very limited in their vision. Again, iterative. Not a fundamental change of what an App Store is or means.

The two that stand out in Smith’s piece are repeat approvals of old apps and refunds. The former would be tremendously useful in weeding out abandoned/old/deprecated/barely working apps that clutter practically every search result. Apple’s two options to fix search have been to improve the algorithm, or to actually focus on the quality of the catalogue. Given the difficulty of the latter, it’s no surprise that Apple’s chosen to work on algorithm tweaks. And its 1.2 million headline number would be affected (perhaps significantly) if it forces apps to be re-approved annually, for example.

Refunds would make a big difference. But that’s still not thinking big enough. In fact, refunds would make a bigger difference to developers than to customers. (Apple announced a few new developer-focused App Store features at WWDC. The integration of TestFlight is a big deal if you build apps, but customers don’t care. Analytics in the new version of iTunes Connect which will help developers track downloads are – again – invisible to customers.

A social solution?

Blogger (and XOXO Conference founder) Andy Baio’s post on Medium suggests adding a social layer to the App Store (hardly a new idea in itself). Apple’s (valid) obsession with privacy has made any automatic layering of a social graph from Facebook/Twitter/the user’s address book a non-starter, and for very good reason. Would you want the fact that you’ve downloaded Tinder (for example) to be broadcast to your friends?

“It’s not just about friends, but about following people, developers, or organizations that are finding and curating stuff you like.” The point about this ‘following’ process not being like Facebook but more like Twitter is an important one. In fact, with apps, common interest (Twitter) is far more important to discovery than common history (Facebook).

But it’s not enough to simply layer a social graph on top of the App Store. The actual mechanics of ‘following’ and ‘sharing’ need lots of focus. This is where Baio falls into the trap of summarizing how this could work in little more than two sentences. Following is not as simple as leveraging a user’s entire social graph from Twitter. What if they don’t have Twitter? What if they’re only interested in a specific sub-set of the people they follow on Twitter? How many people/connections does it need to be useful? How does that social graph manifest itself?

Perhaps (as Baio suggests) the whole notion of reviews needs to be reimagined as a recommendation layer. Importantly, the process of favouriting or sharing an app needs to be as close to frictionless as possible. Whether this is done in-app/per-app or within the App Store every so often is a consideration. Using push notifications to generate engagement would be useful (but obvious).

This is a very, very difficult problem to solve.

Baio sums it up brilliantly: “Apple is using discovery methods from the age of brick-and-mortar bookstores and videogame shops—shelves of staff picks and bestseller lists are useful, but they’ll never be able to expose more than the very surface of what’s in the App Store.”

Apple hasn’t moved aggressively to address the app discovery problem this year.

Perhaps Google will (at I/O this week?).

Image: Mike Lau via Flickr.