Why Android First is a Myth
In mobile, particularly in consumer markets, there has been an ongoing debate about when or if Android will become the first platform that sophisticated startups develop for. Often times, this argument is wrongly centered around vanity metrics for Android such as worldwide shipments or installed base of each platform.
The reality is that platform constraints at the engineering and financing levels tell a much different story. “Android-first” faces structural and financial barriers which are unlikely to be overcome. iOS will remain the primary platform that startups develop for regardless of how much more quickly Android grows share. Here are the reasons why:
- In the US, iOS market share is still extremely strong (even pre-iPhone 5s launch data showed Android having peaked, so Q4 data will be interesting with Apple’s refresh). Since the vast majority of innovative mobile startups come out of the US, Apple’s stronghold domestically has an absolutely massive impact on developer mindshare—e.g. even if China gains another 400M subscribers this year, this fundamental fact won’t change.
- All of my conversations over the past year with Android developers, 3rd party dev shops, more mature startups developing on both platforms and investors confirm a simple hard reality: building and releasing on Android costs 2-3x more than iOS. This is due to a multitude of reasons: less sophisticated tools, generally more cumbersome APIs, fewer exposed advanced features, enormous QA issues brought on by fragmentation, etc. The rough rule of thumb is for every iOS engineer you actually need two Android engineers—or twice the development time.
- The effort required to build and release an app is severely gated by capital-raising. Today’s startup seed rounds typically range between $800K to $1.2M. With that amount of capital, startups are expected to not only release a polished app, but also show demonstrable traction before raising capital again (generally 5-10x the user traction versus what was required a few years ago).
- These structural limitations around capital raising for venture-backed companies force startups to take a non-linear path to development which is gated by fundraising—the types of milestones that a company must hit to raise a seed round (great founding team, big market, good idea) are radically different than at the Series A round (significant traction, repeatable user acquisition strategy, early ideas toward monetization, etc).
- To build a mobile app with $1M in capital, a startup can roughly afford to hire one designer, one client developer (iOS or Android) and one back end engineer. Often the technical co-founder is a hybrid back-end engineer and the business founder plays a hybrid product role. This will allow the startup around 18 months with which to release a mobile app and demonstrate product-market fit.
- Almost zero startups are going Android-first under these constraints. ((There are always exceptions. One is Grand St, which has a founding team with extremely strong Android DNA)). Why? Because founders know they have an extremely high bar to prove traction on the primary platform, before they can raise additional financing and accelerate into two platforms. The second platform basically looks like a step function at Series A. (e.g. go iOS with the seed round, show traction, raise an A round, then build for Android). ((Third platforms like Windows Phone are ignored even after the company is much bigger. Major companies like Instagram and Pinterest still don’t have Windows Phone apps)).
- So it’s well known in tech circles today that seed round sizes constrain app development to a single primary platform. And startups are choosing to go iOS first not only because development is cheaper and easier, but also because money for in-app purchases and advertising is overwhelmingly skewed toward iOS . In fact, a recent study of Facebook ads shows ads were 1,790% more profitable on iOS. This is extremely incriminating for Android and is the worst kind of news for Google. Money and time spent in Android is simply not catching up.
- Since iOS better supports startups’ ability to prove metrics requisite for raising Series A rounds from institutional investors, the earliest most innovative services are almost always available first on iOS. And since the lag between seed and A rounds is often 12-18 months, Android tends to lag this far too. Of course many of these innovative startups produce beautiful apps but don’t get traction, and fail to fundraise again.
- For every seed-funded startup which successfully raises a series A, about 80-90% die trying. One by-product of these capital constraints proves particularly interesting: often these startups become acquihires for the top mobile acquirers (FB, Google, Apple, Yahoo, Dropbox, etc). Because they are almost always shutdown at acquihire, big companies often have some of the most talented iOS engineers and product people in residence for a 1 to 2 year earn-out period. Without a doubt, these employees skew toward iOS when they join internal projects or think up new ideas. And when they eventually leave, there is a good chance they’ll stick with iOS again. There is no doubt this forms a sort of virtuous circle of iOS-first talent in the startup community.
While in theory Android provides a very modern platform for mobile development, the realities around Android-first are quite different. Startups simply cannot afford to bypass iOS and go Android out of the gate. One could even argue the gap is widening.
Uninformed people who reference innovation frameworks often like to declare that the mobile market has matured to the point of being “good enough”. I disagree. The reality is that software innovation at the app layer is accelerating, and converged hardware / software development costs a lot of money.
We all know that creating valuable mobile services is extremely difficult—but it’s interesting that today iOS is still far ahead when we take into account the relationship between development costs and available financing. And because Android lags so much here, to me this is further evidence that platform parity and “good enough” in mobile may be a long ways off.