New questions in mobile

Seven years into the smartphone world, it seems like it’s time to change the questions. The questions that we asked and argued about for the last few years have now mostly been answered, become irrelevant, or both, and new problems and puzzles are emerging.

Hence, the first phase of the platform wars is over: Apple and Google both won. Apple now sells around 10% of all the 1.8bn (and growing) phones sold on Earth each year and Android the next 50%, split roughly between say 2/3 Google Android outside China and 1/3 non-Google Android inside China. Over time this will expand such that smartphones take almost all phone sales - perhaps 400m or 500m units a quarter, with Apple taking the high-end and Android the rest, and there'll be close to 4bn smartphones on earth. And though Apple sells a minority of devices, its positioning and execution means it has a much larger share of traffic and a majority of content and ecommerce revenue in developed markets, so its ecosystem is perfectly sustainable, as is Google's.

So Apple and Google have both won, and both got what they wanted, more or less, and that's not going to change imminently. Within that framework, what happens next?

What happens to Android OEMs?

After several years in which Samsung was able to use scale and better execution to take half of the Android market by volume and almost all the profits, the OEM space is changing, with Samsung's share of Android sinking fast. It isn’t so much that the ‘traditional’ branded OEMs are doing better, although some are (Motorola, LG) as that the whole Shenzhen ecosystem is starting to go global. Most obviously Xiaomi has taken real share in China and is trying to do the same in India, but there is a swarm of smaller local players around emerging and middle-income markets leveraging the Shenzhen phone manufacturing complex to build brand and share in their own country. How is this going to play out? Will Xiaomi build a global mid-range brand? Will other Chinese OEMs succeed in doing the same? How many local players will there be? Handsets have always been a scale game, but is it possible to be a small local brand but leverage the scale of the whole Shenzhen complex to do without scale of your own? Will someone finally make a premium range that can take on Apple (if that's possible on Android)?

What is Android going to be?

One could argue that the future of the OEMs doesn't much matter except to their shareholders - it's just the label printed on the back of the phone. But companies like Xiaomi point to a second Android question - what is Android itself going to be?

So far 'non-Google' Android has been constrained into two areas. On one hand, since Google is effectively absent from China almost all Android devices there come without Google services, but this isn't a broader strategic problem for Google. On the other Amazon has tried to make its own sub-platform (a 'fork') with the Kindle Fire line, but the Fire tablet is a niche product and Fire Phone has flopped, partly but not entirely because it consequently lacked any Google Services. Meanwhile the attempts of the mainstream Android OEMs to add their own layer of differentiation on top of Android have mostly failed (see my discussion of this here).

Now this seems like it might be changing. On one hand Xiaomi, unlike other OEMs, has managed to create a pretty appealing set of differentiated services on top of Android. For now, outside China it uses Google's services as well, but that may not last. In parallel, our portfolio investment Cyanogen creates the scope for many more companies to start experimenting with what kind of 'Android' they ship.

The lynch pin of Google's control of Android has been Google Play Services, the set of APIs and apps that it layers on top of Android - which really means Maps and Apps. Google uses these rather like Microsoft used Office and Windows - mutually supporting levers, complete with clauses meaning you cannot sell any official Android phones if you also sell a forked one, which prevents big players from experimenting.

We tend to assume that it's not possible to sell a phone outside China without Google's Maps and app store - that such a phone is effectively just a featurephone (at least to a normal consumer). But then, no-one has really tried apart from Amazon (and Nokia very briefly), so far, and there are some strong indications that Google Maps actually has a pretty small share of iOS use. I suspect that many of Google's mobile services have user bases that are very broad but also very shallow. And with the systemic unprofitability of the major Android OEMs' attempts to sell 'official' Android, and the proliferation of smaller players riding on Shenzhen and with little or no existing business to lose if Google cuts them off for forking, more people might try.

This however leads to a deeper Android question - what will Google be offering in 5 years? Will we, say, be buying Chrome phones on which Android is a legacy run-time? Google already has Android apps running on Chrome (and vice versa) - does that makes sense as the future of the platform? And if so, does that mean more or less freedom of action for people trying to build their own platform on top?

Interaction models, messaging and layers of aggregation

Part of the reason for Google to change what 'Android' means is control, but a deeper one is to change the smartphone interaction model. It's always been obvious that apps were a structural problem for Google, since their content is hidden from search and more importantly unlinkable from search results, paid or not. HTML5 'web apps' turned out to be a blind alley for a range of reasons but the underling problem they addressed remains: on the web you can link to any arbitrary resource and on mobile you cannot - everything is inside silos.

Quite what the answer to this might be is opaque. One answer is to go back to the web, and in Android 5 'Lollipop' Google is clearly trying to blur the differences between apps and web views. However, I'm not sure that just going back is ever the answer: tech tends to move forward, solving existing problems and creating new ones but not going back to the old solution and the old problems. We didn't solve the problems with the web by rebuilding AOL, but with Google and Facebook.

Meanwhile, as we all know, the search model that Google brought to the web does not really work on mobile - mobile is both 'post-Netscape' and 'post-Pagerank'. But in the same theme, I don't remember the pre-mobile, Google web as a prelapsarian paradise lost - there were plenty of problems with Google's dominance.

What we can see happening now are shifts in the layers of aggregation and discovery. We've gone from everything being bundled within the web browser, and then accessed through Google (itself a bundle) to app icons unbundling specific content from the web. At the same time apps bundle all the content from one site and make it impossible to link directly to a specific page. Hence we now have both deep linking, which lets you get into the silo, and rich actionable notifications unbundling pieces of content out of the silo. And we also have cards floating around (delivered in many different ways) as another unbundling metaphor.

This in turn creates new problems. Deep linking doesn't necessarily fit an app interaction model, and it's also not clear who should be doing it anyway - is this something that should be layered on top (Facebook, Google on iOS) or built by the platform owners themselves? Meanwhile notifications are becoming a new message stream with notification panels on iOS and Android starting to break (in different ways on each) under the volume.

This reminds me of Zawinski's Law, if one expands mail to include messaging:

Every program attempts to expand until it can read mail. Those programs which cannot so expand are replaced by ones which can.

We can see one route for messaging as the new interaction model in China, as laid out here (the Chinese internet is always worth looking at to challenge your assumptions as to what's inevitable). What's really going on in China is a relocation of the aggregation and discovery layer from the app store and home screen into apps from internet platform providers (though just to make life easy there are also at least half a dozen important app stores on Android). I discussed some of these issues here, focusing on Baidu maps. That is, content is unbundled from the web browser into an app, which is actually a new bundle, and then unbundled again into messages and notifications, which you then bundle into a messaging app - or just the notifications panel. And this is without even thinking about the future of email (a topic for another post) and things like Slack, which unbundle workflows from email, move them into an app - and then give you notifications. This reminds me of another quote, in which a PG Wodehouse character explains what he will do when he finds a man he objects to:

"I propose, if and when found, to take him by his beastly neck, shake him until he froths, and pull him inside out and make him swallow himself"

So, how many times can we pull the apps/web/mail/messaging model inside out and make it swallow itself? Right now it's being shaken.

Facebook and Amazon

Clearly, as I mentioned above, there is a question as to whether fundamental platform-level interaction models can be changed by third parties or whether only Apple and Google can really do it. But Amazon and Facebook also play a role here, or rather they would like to. The web browser was (in hindsight) a neutral, unmediated platform and the smartphone is not. The platform owners do things that affect how you discover and engage with, well, anything. This is why Android itself exists - Google was scared of what Microsoft might do - and it’s a big reason why Amazon and Facebook keep circling around the interaction model with things like the Fire Phone and Facebook Home. So far they’ve not found away to insert themselves at that key, gate-keeping level of the stack - at the primary aggregation layer that control of the home screen gives you. But they're not going to give up. Most obviously, Facebook is trying to build a meta-layer on top of both Android and iOS to allow deep linking and sharing to work within apps. How far will this work? Is it (and deep linking in general) really something for the platform owners to build? What will Amazon (which is also in the search and links business) do next?

Wearables - end-points both for cloud and for messaging

My final question, at least today, is around wearables. I wrote here that the Apple Watch reminds me of the iPad in that if you want such a thing it's well done but it's not clear that we want such thing at all. But it is clear, at least for now, that wearables generically are end-points for cloud services, but watches, simply because of the screen size, are largely messaging and notification devices. To the extent that they take off, this means they will necessarily feed back into the broader question of what the broader interaction model looks like and where messages and notifications sit within that. Pushing further out, of course, things like Oculus and especially Magic Leap have the potential to make much more fundamental changes to prevailing interaction model, but we're a few years away from that.


Across all of this, and far more important, we are now well on our way to having some 3.5bn to 4bn people on earth with a smartphone - there are probably 2bn today, and close to 4bn people with a mobile phone (the number of duplicate SIMs makes the number of active connection closer to 6bn). This compares to around 1.6bn PCs, of which roughly half are consumer and half corporate. So there will be something like 5 times more smartphones than consumer PCs, and those devices are always with you and, with all their apps and sensors, are much more sophisticated than PCs ever were, seen as internet devices.

It’s hard to express just how much of a change this is. For the first time ever, the tech industry is selling not just to big corporations or middle-class families but to four fifths of all the adults on earth - it is selling to people who don’t have mains electricity or running water and substitute spending on cigarettes for mobile. This means we have accelerating complexity in three ways: we go from middle class families to everyone, we go from the PC to mobile, and we go from the simplicity of the web browser, mouse and keyboard model that lasted twenty years (thanks Marc!) to these complex, multi-layered devices where everything is still changing. And this is why it's so much fun.