Benedict Evans

Until recently, the personal computer (PC) ecosystem was the centre of gravity of the tech industry: it was where the investment and innovation was centred. It took that role away from mainframes, mini-computers and workstations slowly and in stages over the previous 30 years or so, and it did it with scale – PCs could be sold to so many more people that their economies of scale became overwhelming.

The same thing is happening now as the ‘mobile’ ecosystem supplants the PC ecosystem. Last year, probably around 275 million PCs and 1.5 billion smartphones were sold worldwide. In a few years, annual unit sales of smartphones will be around 10 times the sales of PCs – more than 2 billion versus perhaps 200 million to 250 million PCs. There will be 4 billion to 5 billion smartphones on earth versus 1.5 billion PCs. Almost everyone in the world is going to have one of these.

This is where the economies of scale are and where the investment is focused. Today, if you are innovating in sensors or cameras or radios or pretty much any other component, you target mobile. And over time, these economies of scale mean mobile will supplant the PC just as the PC supplanted everything before it. Crucially, that doesn’t mean ‘mobile phones’ replace PCs. A data centre today is hundreds of thousands of PCs – but they are not actually ‘personal’. That’s just the label for an ecosystem that sits behind every kind of computing. The same happens with ‘mobile’ – ARM, iOS and Android are winning, not a screen size or input method, and will take over pretty much all computing in time. So ‘mobile’ isn’t a new part of tech, it is the new centre of gravity.

Consumers are already moving – ‘mobile’ devices are half of traffic in the developed world.
But again we need to ask what we mean by ‘mobile’. Around half of traffic from these devices is on Wi-Fi – at home or at work, not walking down the street or waiting in a queue. So we shouldn’t think of mobile use as limited to certain times or locations – rather, mobile use is ubiquitous. Mobile expands the internet from ‘every home and every desk’ to ‘every pocket’. Just as important, we still have a lingering mental model of mobile devices as limited – ‘yes, you can take them everywhere but the PC has the full internet’. We should reverse that. An iPhone 6 has over 600 times more CPU transistors than a Pentium from 1995 – 600 times more computing power in crude terms. And a PC only has the web, while a smartphone does much more – with permission it knows where you are, whether you’re walking or sitting, who your friends are and what your credit card is. It’s the PC internet that is limited and constrained, and mobile that’s the complete internet.

All this poses new problems. Smartphone apps unbundled the web browser into lots of new experiences – we are post-Netscape, and therefore we are also post-PageRank. We have left behind the monolithic browser, mouse and keyboard model that lasted 20 years. We have also left behind search as a single unifying way to get users. Using an app store today feels like using Yahoo in 2000 – a manually curated home page in front of a hierarchical directory with millions of entries – that once seemed like a good model but that’s now far too large to browse. And we don’t have Google for that, yet.

We are post-Netscape and post-PageRank, but we haven’t settled on a new model. We have two platform providers, computing energetically and creating new interaction possibilities every year, and creating and breaking business models as they go.

Facebook (and perhaps Amazon) wants to create its own layer on top, and so does Snapchat and a dozen or more startups. Really, the hunt is for the next runtime – after the web and apps, how else can we develop, and can that new step give us new ways to acquire users?

Of course, that’s not a new problem. The internet may have removed physical barriers – no more printing costs or limited retail inventory. But people have to hear about you – they have to know you’re great – and that’s a bigger challenge than ever. That’s a marketing problem, not a tech problem.