This article is more than 1 year old

Microsoft: Ditch your phone biz and do crazy hardware experiments

Let a million Echoes bloom?

Analysis Even if Windows 10 Mobile was flawless, the debate would rage about whether continuing to develop a phone-friendly branch of Windows is the wisest thing Microsoft could do.

Microsoft is in a holding pattern while it looks for ways of getting people to write Universal apps. But instead of maintaining a phone platform, there are all kinds of different and more interesting things it could do instead.

After using the latest, so-close-to-release-you-can-smell-it build of Windows 10 for phones, the strategy strikes me as a bad case of the Sunken Cost Fallacy. Mega rival and neighbour Amazon seems to be shrewder and braver. Microsoft could take a leaf out of its book.

Even if you don’t know what the Sunk/Sunken Cost Fallacy is, you can make a pretty good guess what it is. The amount spent on a project can never be recovered: you may as well keep going. Concorde is the most oft-cited example; the economic case for it was always dubious, but politics and pride and got in the way.

The description (it’s not an economic theory) was cited a great deal last week as the Washington DC street car was finally opened to the public. This boondoggle has been funded for more than a decade, and goes from nowhere to nowhere. But city politicians figured they’d got this far – they may as well open it. Is Windows mobile another example of the sunken cost fallacy?

Not entirely, because maintaining a mobile branch of Windows 10 isn’t simply about phones. It’s about maintaining a branch that’s suitable for tablets and two-in-ones, which is how people in emerging markets are increasingly doing their computing. A full fat Windows 10 isn’t needed here, so a cut down mobile one does the trick.

But at the same time, there’s an opportunity cost to go with the real cost of a platform nobody uses.

Think how Amazon has de-creepified its Amazon Echo voice hub. I can’t imagine why I’d have an Echo, and maybe neither do you. But people who have one really like it, and it’s succeeding with almost no marketing. Amazon expanded the range yesterday. Why didn’t Microsoft get there first, with a Cortana hub?

Cortana is just as good at the technical bits as Amazon's offering but is less impressive at implementations. One Google veteran reminded me this week of Alan Kay’s observation that to make great software, you must make hardware*. Echo is an experiment that allows Amazon to refine the practical uses. Echo should be a Microsoft experiment too.

One reason Windows 10 Mobile is being left behind because it doesn’t talk other devices in your life.

Windows phones don’t talk nicely to cars. Windows phones don’t work with wearables, except for Fitbit and Microsoft’s own Band. There is no Windows phone payment infrastructure. Even if CEO Satya Nadella could wave a magic wand overnight, plugging the gap left by the absence of thousands of minor apps, and quite major (finance, Sky, Virgin) apps, Windows phone would remain marooned. But that’s something Microsoft can fix with hardware experiments.

New consoles, new media viewers (perhaps incorporating parts of HoloLens) … yes, they’d be expensive to design and bring to market. But at least they’d help retain and win developer mindshare. And Microsoft isn’t quite so inept at this as everyone assumes. When it was launched, the Surface invited only ridicule. Now it makes money, and everyone is copying two-in-ones. Microsoft’s version of a fitness Band is seriously underrated, if a little expensive to hit the part of the market that gobbles up Fitbits.

For Microsoft’s Universal app strategy to succeed, it would make as much sense to experiment with new device categories as it would to continue pouring money down the black hole that’s a phone platform. ®

Bootnote

After 24 hours, my test Microsoft Lumia 950 (running the shiny and speedy Build 122) still refuses to configure my Microsoft Band. Practice makes perfect?

* Alan Kay's quote was the punchline to a 1982 lecture: "People who are really serious about software should make their own hardware."

More about

TIP US OFF

Send us news


Other stories you might like