There is no mobile ecosystem

This morning I saw a snippet of a new blog post by Benedict Evans (of a16z) float by on Twitter that made me scratch my head. It posited that while the iPad Pro and Surface Pro “look similar”, the former is part of an innovative new “mobile ecosystem” that is on the rise, while the latter is the dying breathe of a withering PC ecosystem. I called BS.

Why? Because there is no mobile ecosystem.

The PC ecosystem is real. For decades now it has had Microsoft and Intel at its center. Orbiting them were Windows software developers, IHVs, and PC OEMs – the latter fighting over customers and spewing out dollars, most of which were sucked in by Microsoft and Intel’s immense gravity. In the early 2000s this system was growing rampantly, and was stable enough to weather storms like a string of much-publicized security problems, and the drunken stumblings misteps of the Longhorn/Vista years.

However, as time went on, and complacency set in, things started to wobble a bit. As innovation in this system essentially stopped, demand for new PCs started to slow, and OEMs started driving down prices and building more and more “disposable” machines. Then a disastrous thing happened. Apple introduced the iPhone.

Almost over night, software developers (who had become bored and disillusioned by all the unrealized promises of Longhorn) turned their attention to this entirely new category of devices. Apple didn’t even have a public native app platform, so developers created the first mobile web apps and frameworks, and then began hacking and reverse-engineering Apple’s internal API set. Apple didn’t provide an app store, so developers built their own. Thus began the emergence of the Apple ecosystem.

Apple saw the fragility in the PC ecosystem and wanted no part of it. They already built the OS and the hardware, and soon decided they’d like to control even more. They even figured out how to exert unprecedented control over the cell carriers. However, that first iPhone showed that independent developers were not something they could ignore (as they did with the iPod, for example). So what did they do? They offered those developers a tight leash from the very beginning, with an institutionalized flow of cash to the mothership included for good measure. You would only sell apps and content through the App Store. You would only develop your apps using a Mac. You most definitely would not build anything that Apple felt threatened by, or use a technology they deemed inadequate. And you would give Apple a third of your income and thank them for the privilege.

The mobile smartphone market is clearly critical to the Apple ecosystem. It’s where they first got traction, because they basically invented it and had no (meaningful) competition early on. But even with the iPad included, mobile is the beginning, not the end, of the Apple ecosystem. They’ve leveraged it (in a very 90s Microsoft way) to get the vast majority of developers onto Macs. This is really important. It creates a very difficult-to-break feedback loop. There’s significant resistance to breaking developers away from the Apple ecosystem because the Apple ecosystem is where devs *live*, even if they try an Android phone, for example.

Apple is pushing to extend this thriving ecosystem into new markets like TVs/set-tops and wearables. The iPhone is the center of gravity in Apple’s ecosystem, and their other products revolve around it. The Apple Watch requires one. The Apple TV is meant to work with one. Macs are required to develop for one. But what Apple has never had any desire to participate in is a mobile ecosystem. Perish the thought.

Now that bit from Benedict’s post that caught my eye was about the iPad Pro. So where does that fit in? I don’t think the iPad Pro is a product Apple came up with to expand their ecosystem. I think it’s a bulwark against the Surface (and similar modern tablets/PCs) encroaching on their Mac (and to a lesser extent, iPad) business. It might get traction in some nice niches/verticals, but I think its primary purpose is defending their ecosystem, not expanding it.

Google isn’t cultivating a mobile ecosystem either. They’re cultivating a Google ecosystem. The mobile part of that ecosystem is also strong for two reasons:

A) Price. Or in more words: They played the Windows game of old and let OEMs race to the bottom, flooding the market with cheap Android handsets. Except they didn’t care about extracting profit because at that time phones were not an important part of Google’s ecosystem.

B) Carrier support in the early days from everyone who wasn’t AT&T and were desperate to be in the smartphone space. Apple got too comfortable and took too long to address this, which gave Android a larger in than they might’ve had if Apple had broken carrier exclusivity sooner.

Originally, the center of Google’s ecosystem was search. They had content publishers and advertisers orbiting their search advertising behemoth, and projects like Android were just meant to feed and protect that lucrative system. Times change, though, and so have Google’s priorities. Where they once saw devices as Amazon does (places to offer their goods and services), they seem to be increasingly seeing devices (and Android in particular) as the center the Google ecosystem, with their services orbiting Android the way Apple’s watch is tethered to the iPhone. This is probably less a matter of their desire, and more a response to competitors. Apple, Microsoft, and Amazon are all reorganizing their businesses around a model that puts them at the center of their own ecosystem. Ecosystems which are much more self-sustaining than the PC ecosystem of old.

That brings us to Microsoft. Initially they were either oblivious or in denial as to ecosystem amassing around Apple. Later, they saw Apple raking in cash with their model, and Google “succeeding” with Android using the old Windows/PC model (sans the profit), and said “let’s do both of those things!”. In phone land, the result was a handful of third-party Windows Phone devices that all had identical components. With no ability to make competitive deals with component vendors, and hamstrung waiting for Microsoft to release new device class specifications, none of these OEMs stood a change of competing with Apple or Google’s partners who could do whatever they wanted. They had no userbase to offer to developers, and a severely limited platform built on, no joke, Silverlight. They even wanted to imitate Apple so badly they bragged about leaving out copy-and-paste from the first release of Windows Phone 7. That’s how seriously these people took mobile.

On the PC side, they took a slightly different approach. Being like Apple meant spinning up an in-house hardware unit to build the Surface. At the same time, they also tried the Windows Phone model of building detailed ARM device specifications and “permitting” a few vendors to build virtually identical devices. Oh, and they also tried letting other OEMs roam free on Intel hardware. Cynically I call this the “have your cake, eat it, and sell it back to them too” approach. You could also consider it the “try everything and see what works” approach. Neither conveys a high degree of focus.

To their OEM partners, this was all an experiment, an attempt to establish a “premium” tier and to motivate and inspire those partners to innovate. In reality, whether the original intention or not, this was the beginning of realigning the PC solar system. If this was the plan, it was sort of brilliant. Microsoft saw the PC ecosystem slowly collapsing, and began preparing itself for a new universe where it would need to take on a different role, with Windows at the center of an ecosystem surrounded by hardware of all kinds (including servers in an Azure datacenter). All the while keeping PC OEMs feeding them while Microsoft prepared for their inevitable demise.

Whatever the plan was back in that Windows 8/RT timeframe, the execution, obviously, did not work out so well. The real sin wasn’t the failure to burst into the mobile market, but that in doing so they destabilized the strongest part of their ecosystem at that critical moment. It’s taken years to even begin undoing that damage. In the long run maybe this will turn out to be a growing pain. Or it may turn out to have been a crucial misstep from which they’ll never recover.

In fact, I’d attribute the Windows 8/RT disaster to exactly the kind of thinking that Mr. Evans shared in his post. That’s how you get to the idea that those 1.5bn PCs in use, and the ~0.4 billion a year sold, don’t matter, and that it was “okay” if they all stayed on Windows 7. Because all that matters is some non-existent “mobile ecosystem” that everything needs to immediately be a part of. So why not bet the company on it?

It turns out, however, that mobile didn’t eat the world. PCs (at least in the traditional sense) may not be a growth market, but they are a crucial part of Microsoft’s ecosystem, and will be for some time. Mobile is a growth opportunity for them. A huge one. But succeeding there these past few years was made so much harder by the disarray of their existing ecosystem.

That brings us to today. Microsoft’s ecosystem is showing signs of stabilizing, with the Surface Pro 3 and now Windows 10 providing some much-needed energy and positivity in their PC business. The cost of this is clear: Microsoft has all but entirely exited the mobile market over the last year or two. Whatever little progress they’d made with Nokia and WP 8.1 was left to evaporate while the company fixed PCs. Sure, they have promised two new flagship Lumias before the year comes to a close. But everyone paying attention can see that these are (far too late) stopgaps while they “retrench”. The excitement their new head of hardware had for these at the announcement event was… was… what’s the opposite of “palpable”?

That’s not to say they’re giving up on mobile. Not at all. But it’s going to take more than competitive devices and a competitive platform to make a dent in that market. It’s going to take a Microsoft ecosystem firing on all cylinders, with attractive offerings for customers and developers. Developers are the hardest nut to crack. And they are absolutely not looking for a new mobile platform to support.

So what is Microsoft doing to win developers today? A few things:

• Aggressively courting them with Azure.
• Taking a leadership role in the evolution of the web’s technology and standards.
• Creating the small beachhead that is Visual Studio Code (free open-source code editor for Mac, Linux, and Windows).

Notice I didn’t even mention “building a modern, universal platform”. They’ve done that. But without users or the opportunity to build something new and unique, this means nothing. It’s a prerequisite for success, not a strategy.

What can they do? I look at the problem three ways:

1. Bring all your users to the table. They don’t have a meaningful number in mobile. But they have 130+ million on Windows 10 PCs, and it’s a fair bet they’ll march that number of up to a billion over the next few years. Find ways to get these users buying apps, and you’ll attract developers. They know this. They’ve just struggled to figure out how. Windows 10 is trying some new tactics here that they’re going to be dialing up a bit soon, and I hear early indications are that they’re having some success with this (meaning more users are buying apps on from the Win10 store). But just driving users to the store isn’t enough. They need to encourage a PC app renaissance, and I have a bunch of ideas for how to do this (perhaps for another post).

   Also, adding in 15 million and growing Xbox One users who love to spend $$ into the mix will not hurt.

2. Let developers do something unique. This is what drew developers to the iPhone, and Apple is actually really great at this. HoloLens could be a long-term play for this, but that’s clearly not enough. Kinect might’ve been a big missed opportunity here, and Xbox might still be the key place to pull it off. That’s because letting developers do something unique generally means giving them unique hardware to play with.
3. Find developers where they live. Microsoft is really strong at developer tools. Apple is kind of atrocious at them. So what’s the opportunity here for Microsoft? Here’s a bold vision they could adopt:

   Make Microsoft’s tools be the best way to build an iOS app.

   What could that look like? Here are some more specific ideas:

   1. Make the Surface Pro/Book the best device for building iOS apps. Make Visual Studio into the best Objective-C and Swift IDE. Make the best emulator experience (newsflash: it doesn’t even need to run iOS!). Provide the best cloud-based build environment (a la PhoneGap Build) running on some Macs in a Microsoft datacenter, to get around Apple’s licensing restrictions.
   2. Make Windows 10 on a Mac the best iOS development experience. Buy up Parallels, and give developers a VM with Windows 10 and Visual Studio to run on their Mac. Make some awesome plumbing so that VS can connect across the VM host to the iOS emulators and build tools in OS X. Bring the game to Xcode’s home turf and crush it.
   3. Build a full Visual Studio iOS development experience for OS X.

I really like the second option there, actually. It’s technically doable. Apple would be powerless to stop it. You could make developers love Windows without them having to give up their Mac. Then later you show them why they’d love it even more on a Surface with a touch screen (those phone emulators really light up when you can, you know, touch them).

Is this a plan for cracking the “mobile ecosystem”? No. It’s a plan for attracting developers into the Microsoft ecosystem. Once they’re there, Microsoft will have a way easier time directing their attention to mobile.