Microsoft, Please Pay Attention to What Apple is Doing with iOS (Premium)

A second report claims that Apple will slow down the addition of new features to iOS in order to focus on quality. This is exactly the strategy that Microsoft needs to adopt. In fact, it’s years overdue.

Bloomberg hired Mark Gurman for one reason and one reason only: His insider access to Apple. And it’s paying off: This highly-accurate Apple watcher has registered scoop after scoop, and for this reason, I am taking his reports about Apple’s iOS strategy changes as gospel. And he’s now written on this topic twice.

In late January, Gurman reported that Apple was pushing some key new iOS features past this year’s iOS 12 release. The reason? Software quality has nosedived in the most recent iOS release as Apple has stepped up the addition of new features.

Today, Gurman has provided a second report explaining how Apple intends to fix this problem. It is adopting a two-year development approach in which major features are being split between iOS 12 and iOS 13 (due in late 2019), which will give its developers some flexibility, not to mention more time when needed. The “renewed focus,” Gurman says, is on “quality.”

Exactly right. As it should be.

More to the point, this is the strategy that Microsoft needs to adopt for Windows. Again.

I’ve been preaching something I call “finishing the job” for years. For example, in 2012, I wrote that “Microsoft has a problem. It simply doesn’t know how to finish the job.”

“The Microsoft culture has rewarded shipping a product or product version above all else, and if that product happens to be attached to some huge revenue stream, all the better,” I wrote at the time. “The problem is that no software products and services are perfect and bug-free. And the dark half of that Microsoft culture is that there’s little if any reward for those whose job it is to set things right.”

This culture persists today, of course. But the problem has been amplified by two things: A near-fervent desire in the Windows organization generally, and in specific teams within Windows, to justify their collective existence by continually pumping out major new features. And the epic failure of Windows as a Service, a make-believe world in which we pretend that Microsoft’s oldest legacy code base can somehow be treated like a modern online service.

So let’s examine both of those points in a bit more detail.

In a culture in which new features are valued over solving existing problems, the A-teamers will always be scrambling to make sure they’re involved with something new and exciting. This ensures that problems—for example, the Windows 10 Copy and Paste bug—are never fixed, or are fixed very slowly.

It also ensures that users will be inundated with pointless new features—-3D, Windows Mixed Reality—that few people even want. The employees responsible for actually fixing problems—let’s call them B-teamers for lack of a better term—have little incentive to do more than the minimum. But fixing problems also applies to modernizing infrastructure: Windows 10 ships with some bizarre collection of games today, for some reason. But we’re still using the same legacy file management system, called File Explorer, that’s been in Windows for many years. Fixing File Explorer just isn’t as sexy, I guess, as a soccer ball shooting fire. Which is something we do have in Windows 10 today. Hooray.

(To be clear, new features, even flaming soccer balls, are fun. But they should not take precedence over quality.)

Back in 2012, I explained that Microsoft planned to update Windows like an online service. (Yes, this was three years before Microsoft announced “Windows as a service.”) I also expressed reservations.

“Whether locally installed end-user software such as Windows and Office can successful be updated this way is debatable, but we’ll see, since that’s the plan,” I wrote. “Nothing has changed: There’s no real reward for tweaking an existing product, adding that missing feature, or fixing the tiny incomplete bits … Microsoft hasn’t learned a thing. I’d like to see the team responsible for this mess spend the next three years cleaning it up. It’s time the creators at Microsoft took a step back and finished the job.”

(Yes, I even used the word creators. I’ll look for my royalty check.)

Anyway, Microsoft has formalized its process for updating Windows as a service. And since they are so creative, they call this—wait for it— Windows as a Service. It’s been a failure on every level. Worse, in the face of this failure, Microsoft has actually accelerated the schedule for updates.

Yes, it’s insanity.

I’ve written and spoken to this topic many times, but here’s the quick recap: Where Google and Apple update their simpler, more modern mobile platforms—Android and iOS—on an annual basis, Microsoft updates its older, legacy Windows code base biannually. That’s twice a year. (Don’t worry, I had to look it up to be sure.) And no, that doesn’t make any sense.

Windows doesn’t need more major new features twice a year. In fact, when you think about it, the addition of Store-based apps to Windows 10 ensures that users will be swimming in new features throughout the year because those apps can be updated at any time. As Mary Jo Foley first opined on Windows Weekly, what Microsoft should be doing is issuing new feature releases once a year, and then using that other major release as an “R2” (release two) update that focuses solely on quality. But I’m starting to think that even that schedule is too aggressive.

Some of you will argue that Windows as a Service is working. After all, the last two major Windows 10 feature updates, the Creators Update (1703) and the Fall Creators Update (1709), were relatively error-free. Microsoft was so confident in its software quality, in fact, that it actually sped up the deployment of the latter release. In some ways, they’re on a roll.

Right?

Wrong. Microsoft’s ability to release software, to jam unwanted features down our throats, isn’t the issue. One would expect that a so-called software giant would be good at—wait for—releasing software. That’s the minimum. The failure of Windows as a Service can be seen succinctly in Microsoft being forced to twice extend the support cycle for various Windows 10 versions (1511 first, and then 1607, 1703, and 1709). And the reason they had to do this is so obvious that you have to wonder why no one saw it coming.

Oh, right. I did see it coming. In 2012.

“Microsoft might be surprised to discover that consumers and business want different things from Windows,” I wrote back then. “Maybe this product needs to be split up accordingly, with a Metro-less business version that offers only the wonderful new desktop updates in Windows 8. You won’t know until you start asking questions.”

I’m not really tooting my own horn here: Everything I wrote then, as now, is obvious. Just not to Microsoft, apparently. And as should have been predicted by anyone, its biggest and most lucrative customer base—the enterprise—is pushing back against Windows as a Service.

The enterprise doesn’t want to update Windows twice a year. They also don’t want to update Windows once every two years. In fact, the enterprise never wanted to update it once every three years, which was the schedule up through Windows 8. What the enterprise wants, Microsoft, is to just keep using the software that works. They want quality and value, not quantity (of updates) and unpredictability.

Consumers may be able to swallow a faster update schedule than the enterprise, but let’s all agree on two things here. Twice a year is too damned much for anyone. And the enterprise is right to favor quality over quantity. We should all be so demanding.

Well, I am being that demanding.

Microsoft needs to stop this nonsense, copy Apple again, and back off from Windows as a Service. Contrary to Microsoft’s marketing nonsense—and that is all Windows as a Service really is—this pace of development is doing more harm to the platform and its user base than good.

End the insanity, Microsoft. And tie the release of major new features to your corporate support schedule. Today, that would be major new feature once every two years, a sane schedule for a product like this. But then, the schedule will get longer than that: It’s only a matter of time before enterprise push-back will force a three-year Windows 10 version support lifecycle. And at that point, we’re pretty much back to where we started.

Ironic? Maybe. Predictable? Yep. If a simpleton like me saw this coming almost five years ago, this situation should have been obvious to the smarter folks running Windows too. It should be even more obvious today.