Now that mobile is aging, Benedict Evans predicts the Next Big Thing

“Hello, this is the future. Please hold.”
“Hello, this is the future. Please hold.”
Image: AP Photo/David Cantor
We may earn a commission from links on this page.

Now that mobile is maturing and its growth is slowing, everyone in tech is turning to thinking about what the Next Big Thing will be. It’s easy to say that “machine learning is the new mobile” (and everyone does), but there are other things going on, too.

On one hand, we have a set of profound changes coming as a result of new primary technology. Electric and autonomous cars will change cities, virtual and mixed reality will change the entire computing experience, and machine learning is changing the kind of questions that computers can answer. But each of these is also just beginning, especially relative to their potential—they are at the bottom of the S-curve where smartphones are now getting towards the top. On the other hand, I think we can see a set of changes that come not so much from any new technology as it does from shifts in consumer behavior and operating economics. These changes are potentially just as big, and might be starting sooner.

Electric and autonomous cars

Electric and autonomous cars are just beginning. Electric is happening now but will take time to grow, and autonomy is five to 10 years away from the first real launches. As they happen, each of these destabilizes the car industry, changing what it means to make or own a car, and what it means to drive.

Gasoline is half of global oil demand and car accidents kill 1.25 million people a year, and each of those could go away. But as I explored here, that’s just the start: If autonomy ends accidents, removes parking, and transforms what congestion looks like, then we should try to imagine changes to cities on the same scale as those that came with cars themselves. How do cities change if some or all of their parking space is now available for new needs, dumped on the market, or moved to completely different places? Where are you willing to live if access to public transport is everywhere and there are no traffic jams on your commute? How willing are people to go from their home in a suburb to dinner or a bar in a city center on a dark, cold, wet night if they don’t have to park and an on-demand ride is the cost of a coffee? And how does law enforcement change when every passing car is watching everything?

Virtual reality and mixed reality

Virtual reality and mixed reality are also some years away from mass-market adoption. We have some VR products in market today and some very early MR, but for both, it feels as though we are in the 2005-2006 phase of multitouch smartphones—almost, but not yet. Once these really come to market, they may change the world just as much as the iPhone. Mixed reality in particular could change things a great deal: We may all have a pair of glasses that can place something in the world in front of us as though it was really there. Predicting what this could be today reminds me of trying to predict the mobile internet not in 2007, but in 1999; “stock tips, news headlines, and the weather” don’t really capture what has happened since then.

Machine learning

Machine learning is happening right now and rolls through, or perhaps underneath, the entire tech industry as a new fundamental computer-science capability—and of course enables both mixed reality and autonomous cars. Like relational databases or (in a smaller way) smartphone location, machine learning is a building block that will be part of everything, making many things better and enabling some new and surprising companies and products. I don’t think we quite understand what it means to say that computers will be able to read images, video, or speech in the way that they’ve been able to read text and numbers since the 1970s. But though we are creating machine learning now, again, it’s still very early to see all of the implications. It’s at the beginning of the S-curve.

So, we have these hugely important new technologies coming, but they’re not quite here yet. At the same time, though, we have a set of more immediate changes, that have much more to do with consumer behavior, company strategy, and economic tipping points than with primary, frontier technology of the kind that Magic Leap or Waymo are building.

E-commerce and physical retailers

First, e-commerce, having grown more or less in a straight line for the past twenty years, is starting to reach the point where broad classes of retailers are having real trouble. It’s useful to compare physical retail with newspapers, which face many of the same problems: a fixed-cost base with falling revenues, the near-disappearance of a physical distribution advantage, and, above all, unbundling and disaggregation. Everything bad that the internet did to media is probably going to happen to retailers.

The tipping point might now be approaching, particularly in the US, where the situation is worsened by the fact that there is far more retail square footage per capita than in any other developed market. And when the store closes and you turn to shopping online (or are simply forced to, if enough physical retail goes away), you don’t buy all the same things any more than you read the same things when you took your media consumption online: When we went from a corner store to a department store and then from a department store to big box retail, we didn’t all buy exactly the same things but in different places—we bought different things. If you go from buying soap powder in Wal-Mart based on brand and eye-level placement to telling Alexa “I need more soap,” some of your buying will look different.

TV, cable, and on-demand streaming

In parallel to this, TV, which so far has not really been touched by the internet, is also starting to look unstable. Again, this is especially important in the US, which is very over-served by pay TV: Almost everyone has it, and the average spend is much more than people in other developed markets typically pay, so there’s a lot of pent-up desire for change. The US TV market reminds me of those diagrams of three gear wheels interlocked such that none of them can turn: Netflix and Amazon (and others) are trying to unlock them. But though this tension is probably strongest in the US, it applies in most developed markets: On-demand has a new user experience, a new value proposition, and, with no legions of customer support agents and installation engineers, a new cost structure. The tipping points are getting closer.

The result?

All this change affects huge pools of capital. There is retail, TV, and then there is the whole world of advertising. A third of the $500 billion global ad business has now moved to the internet, and Google and Facebook are more than half of that—but TV advertising has hardly changed at all yet. The internet has offered neither the inventory nor the experience to draw TV ad budgets. Indeed, since neither Netflix nor Amazon run advertising in their TV products today, ad budgets so far have stayed with legacy players, even as viewing has shifted. This will probably change, and the more that viewing shifts, the more that ad budgets will be reconsidered.

More deeply, though, the more that buying shifts, the more that ad budgets might change. Will all of that $500 billion be spent in the same ways by the same brands on the same formats to drive the same sales if both physical retail and TV start tipping over? Google and Facebook, as we know and hear all the time, dominate internet traffic and internet ad revenue, and that dominance only seems to get stronger, first from mobile and now machine learning. How much do they capture of this, how much ad spending does Amazon take, and how far can Amazon apply its maxim, “Your margin is my opportunity,” to advertising itself and remove that cost? And what about the $500 billion that’s spent on marketing, in addition to that $500 billion of advertising?

Finally, let’s go back to cars, mixed reality, and machine learning. How much, really, do AVs change shopping or the cost of home delivery? And what happens to your buying choices when machine learning means a pair of glasses can look at your living room and suggest a lamp based on your taste, and then show what it would look like in situ?