2016 Was a Monster of a Year for Windows 10

Windows 10 is wonderful but controversial, and 2016 was in many ways a landmark year for Microsoft’s last great desktop platform. This story is still being told, and the ending is unclear. But Windows 10 dominated my mind-share and coverage this past year. And any discussion of this product must by nature cover both the highs and the lows. There were many of both.

To be fair to Microsoft as well as the people who build this product, I will state up front that I am overwhelmingly and unhesitatingly a fan and enthusiastic user of Windows 10. I believe in the vision for this product, and that it meets the needs of a diverse user base across a wide variety of PC and device form factors.

That Windows 10 arrived during a transitional era of personal computing is no one’s fault. And that our perhaps too deeply and interconnected world of social media, where everyone on earth can have a platform for their opinions—moderate, insane, or somewhere in-between—is likewise a matter of circumstance. But the volume of complaining that occurs today is at a level that has never been possible before, and we—me, you, Microsoft—are all still trying to adapt to this new reality. I naturally recoil at what I think of as the “everybody’s a winner” mentality that pervades today, and I would ask all of you, all of us, really, myself included, to remember that noise is not the same thing as viable or valuable feedback. It’s not the same as truth.

By which I mean, it’s important to put these events in perspective. There is a lot of complaining about Windows 10 these days, but not all of it merits the attention and outrage—faux or real—that we’ve seen, either from individuals on Twitter, perhaps, or the publications that pander nonsense in a mad bid for page views.

The thing is, transitions like the one that is happening all around Windows 10 are difficult. They’re messy, and they’re uncertain. The heart, or core, of what Microsoft has done with Windows 10 is both pure and important. But Microsoft, a big company full of well-meaning people, I’m sure, can likewise make incredible mistakes. Mistakes of strategy. Mistakes of hubris. Just … mistakes. And Windows 10 has been molded by both ends of this spectrum. By the good and the bad. By the yin and the yang.

And so when I look back on 2016, which I’ve spent an inordinate amount of time doing this past week, I can see both sides of this story. I have thought about how I might pull some narrative out of this, some pithy way to explain away it all, to make sense of the jumble of events through which we we all just navigated.

And I can’t, not really. But what I can do is recap some of the stories I see through that fog, and hopefully this, collectively, will provide a clean picture of what happened to Windows 10. I see a few major stories. And then several smaller events too.

And it goes like this.

One billion? Not so fast

At a January 2015 press event, Microsoft’s Terry Myerson infamously pledged that there would be over one billion Windows 10 devices—PCs, phones, Xbox, Surface Hub, HoloLens, IoT, and so on—“within two or three years.” Mr. Myerson made this statement publicly to remind both developers and users that Microsoft had a new platform that was big enough to be in the same conversation as Android and iOS.

He needn’t have bothered: Windows 10’s astonishing growth, at least over its first year, resulted in the best-selling version of Windows that Microsoft has ever released. It would have been—should have been—a great story. By January 2016, there were 200 million active Windows 10 devices, which amounts to real world users. That number jumped to 270 million in March, 300 million in May, and then to 350 million in June. By September, Microsoft could claim over 400 million active Windows 10 users/devices.

But this news was completely offset by a bizarre admission on July 15, 2016, two weeks before the Windows 10 one-year anniversary and the end of the free upgrade offer: Microsoft stated that it would not be able to reach the one billion milestone by the middle of 2018, which is the end of that “two or three years” period that Mr. Myerson had previously noted.

Two weeks prior to that admission, I had openly questioned whether Microsoft would ever meet that goal. “July 2018 could be tight,” I wrote after laying out why I thought they’d never make it. But I also added a bit of perspective I still feel is important: Regardless of the speed at which Microsoft reaches the 1 billion figure, doing so is in fact assured. As is doing so faster than with any previous Windows version. So Windows 10, by any measure, is an incredible success story. And coming as it does during a time in which PCs are in decline and Microsoft couldn’t sell a phone if its very existence depended on it, this level of uptake is all the more impressive.

The problem, of course, was that Microsoft had promised to hit one billion active users/devices on a set schedule. And in doing so, and in failing to achieve that, the company unnecessarily opened itself to outright derision from critics. “I worry about the impact this will have on Microsoft and Windows—and Terry [Myerson], frankly—and that this will trigger another Windows Vista-like round of timidity,” I wrote at the time.

Point being, 400 million is an incredible success story, especially coming as it does in a world in which personal computing is indeed moving to more modern mobile systems. But in betraying its internal hopes and goals, Microsoft opened itself up to criticism that even that 400 million number was, in some ways, bogus: After all, Windows 10 was free to upgraders for that first year, and was thus adopted more quickly than would normally or otherwise be the case.

I’ve argued that this logic is flawed for a variety of reasons—the key one being that in-place upgrades have historically been very difficult, and something that normal users have never done—but reality and perception rarely intersect. The stigma remains and, sorry, but it’s all Microsoft’s fault. All they had to do was communicate that one billion figure more vaguely—“we’ll hit one billion active devices faster than any previous version of Windows,” or whatever—and we wouldn’t even be having this conversation.

Well, except for one thing. Which is the next big Windows 10 story from 2016.

Upgradegate

Microsoft could have simply let the Windows 10 upgrade speak for itself. After all, the had a strong upgrade story and were offering the product for free to Windows 7 and 8.1 users.

They didn’t do that.

Escalating a strategy that started in 2015, Microsoft in the first half of 2016 consistently undermined the credibility of both Windows 10 and itself by forcing the upgrade on its users, often in deceptive ways.

I spent much of the first half of the year complaining about the software giant’s tactics, and for once I wasn’t myself criticized for being a complainer: I’m pretty sure no one was happy with what Microsoft did.

“This entire episode has been indefensible,” I wrote at the time, “with Microsoft introducing a non-stoppable, non-hideable advertisement on several hundred million PCs from around the world. And then upgrading that advertisement to thwart those who do seek to remove or hide it. It has changed the language of the ad, made no clear cancel choice available, and jammed it into the ‘recommended’ updates that auto-install via Windows Update.”

In a move that perhaps led to its July admission about missing that one billion devices goal—I suspect Microsoft internally expected the Windows 10 upgrade to happen even faster, the real reason it missed the goal—Microsoft in May 2016 silently made yet another change to the “Get Windows 10” advertisement on users’ Windows 7/8.1 PCs. And once again, they went too far.

Previous to the change, users who viewed the Get Windows 10 advertisement and then closed the window would simply be shown the ad again later. That is, closing the window didn’t do anything, didn’t trigger some action. You simply closed the window, as you would normally with any window.

But that May change to the advertisement silently introduced new, indefensible behavior: Now, by closing the window, the user was implicitly agreeing to the upgrade, without realizing they had done so. It was literally that deceptive.

“The violation of trust here is almost indescribable,” I wrote. “It’s bad enough that Microsoft has been training Windows 7 and 8.1 users—i.e. most Windows users—to not trust Windows 10 because of this horrible, unstoppable advertisement. But now they will not trust their own sanity because all they’ll remember is that they dismissed the advertisement by clicking the Close windows box. Why on earth did Windows 10 just install on my PC?!?”

Why indeed.

But when asked about this change, Microsoft simply admitted to it. Just … admitted to it.

“As we shared in October [2015], Microsoft has been helping customers who received the Windows 10 upgrade as an ‘Optional’ (and now ‘Recommended’) update, to schedule their upgrade,” a Microsoft statement noted. “The scheduling UI that customers are seeing began on February 1st 2016, and has evolved over time based on customer feedback. Once a customer’s upgrade is scheduled, they will receive a notification that states the time their upgrade is scheduled for, with options to reschedule or cancel if they wish. If the customer wishes to continue with their upgrade at the designated time, they can click ‘OK’ or close the notification with no further action needed.”

Microsoft eventually reversed this change—as you may recall, the outrage was both loud and, in this case, quite deserved—and it even took steps to dismantle all the bad stuff in the Get Windows 10 advertisement in late June. You know, 11 months later than they should have. And then in September, it finally killed “Get Windows 10” for good. So today, Get Windows 10 is just a vestigial nightmare in the collective memories of those Windows 7 and 8.1 users who never wanted the upgrade in the first place. But I bet some of those people still wake up with night sweats just because they had the audacity to want to stick with something that was already working fine for them.

Upgradegate, as I dubbed this event, is impossible to defend, and explaining it is a fool’s errand. The best I can muster here is that Microsoft saw an existential threat in the mobile platforms that were—and still are—slowly killing the PC. And that the ends—quickly establishing Windows 10 as the third major personal computing platform alongside Android and iOS—justified the means. Microsoft will never admit to that publicly of course, and the strategy was so user-adverse it’s hard to even imagine smart people making the decisions that led to this disaster.

In other words, there’s little in the way of nuance on this one.

The future is ARM and I feel fine

In the last month of 2016, Microsoft in effect saved the year from a Windows 10 perspective by providing fans of the platform with an unadulterated feeling of hope for the future. Microsoft announced that it would make the full, mainstream versions of Windows 10 available on ARM-based mobile PCs. Not the mongrel Windows 10 Mobile, mind you, with the Windows phone stink of death still hanging over it like an unwelcome reminder of past defeats. But full-blown realWindows 10, including the Enterprise version with all of the capabilities that businesses demand.

Can I get a hallelujah, people?

For two important reasons—it just happened, and we don’t actually know yet whether this scheme will even be successful—I can’t really recap this event as thoroughly as I did with the previous two. But make no mistake: Announcing full Windows 10 on ARM is one of the biggest Windows 10 stories of 2016 just as assuredly as the first availability of this product will be in 2017. This is a game changer in the sense that it gives Microsoft a mobile strategy that both makes sense and answers real customer needs.

Terry Myerson made the announcement at his WinHEC 2017 keynote address this past week, but rumors of x86 emulation on ARM had popped up a month earlier. At that time, I noted that emulating x86 on ARM was “a sort of holy grail: An as-yet unobtainable goal that would help us bridge the legacy PC past with the mobile/ARM future.”

Many people assumed then that x86 on ARM was for phones, though I at least had the presence of mind to add that “there’s no reason that Windows 10 Home or Pro couldn’t make the ride too, in something similar to Windows RT.”

Turns out that’s exactly what was happening: This was for mobile PCs—realPCs–not phones. ARM-based PCs that will be able to run 32-bit Win32/x86 desktop applications—Apple iTunes, Adobe Photoshop, Google Chrome, whatever—directly and unchanged, and with excellent performance.

“What we’re really providing here is choice,” Microsoft executive vice president Terry Myerson told me earlier this week. “And Qualcomm chipsets have two major advantages that our PC maker partners and customers have been asking for: Incredible battery life and efficient, integrated cellular connectivity.”

Now, this doesn’t signal an end to the Intel era, as Myerson told me. Instead, PC makers and customers will be able choose between Intel and ARM based on their needs. Naturally, those with high-performance needs—gaming PCs or mixed reality, for example—will continue to choose Intel (or AMD). At least for the short term.

If you don’t believe in this vision for the future, fair enough: We’ve been burned before. (I briefly thought that Windows RT had the change to be the next NT, for example.) But be sure to watch the first public demonstration of this technology working.

Me? I’m often pigeonholed as negative, but I think this is one of the best developments I’ve seen for Windows in years, second only, perhaps, to Windows 10 itself. And part of the reason is that Intel, for all its decades of experience with microprocessors, just can’t see to get mobile right. They’ve been holding us back.

“Intel has failed and failed again trying to move its performance-focused x86-compatible chipsets into the mobile world, most notably in mid-2016 when they finally gave up on an Atom processor that was designed for smartphones and tablets,” I explained. “Intel, like Microsoft with Windows phone, simply ceded the market at that time, in this case to ARM, the chipset family that powers all viable mobile platforms today. And in August, Intel accepted their fate in this market by—gasp—licensing ARM.”

In my estimation, ARM with x86 compatibility is “the Peanut Butter Cup principle, two great tastes that go together.” Assuming it works as expected, this is what no compromises computing looks like.

We’ll see.

Happy anniversary

The Windows 10 Anniversary Update was a big topic all year, in part because it was developed openly (and was originally codenamed “Redstone 1”) and in part because the rollout—which upgraded Windows 10 to version 1607—went so terribly.

It started out innocently enough, and at the time of its release—and believing that Windows Insider testing and telemetry were enough to catch all the major issues ahead of time—I declared that this release “makes the best version of Windows yet even better.”

And to further distance myself from blame, I’ll just note that Microsoft told me that “the Anniversary Update is the most heavily-tested release of Windows yet, with over 25 complete builds of the OS shipped to testers since late 2015. The firm reports that Windows Insiders—those individuals who have elected to test these pre-release builds—have collectively spent over 50,000 years working with Windows 10 and have delivered over 75 million individual pieces of feedback, resulting in over 5,000 improvements to the OS.”

Mission accomplished, right?

Not quite. Shortly after the Anniversary Update began rolling out, users began claiming that the update broke the most common webcam in the world. And as I later realized, I had in fact experienced this issue myself, but attributed the problems to Skype, not Windows 10 version 1607. Oops. After other problems materialized, like a Kindle crashing bug, I confirmed my suspicions: Microsoft had slowed the rollout because there were so many issues.

This led, as it so often does, to me calling on Microsoft to improve matters, in this case with a formal Reliable Computing Initiative, similar to the previous Trustworthy Computing Initiative. But after thinking on this for a while, it occurred to me that Microsoft slowing the rollout may have proven the system was working. That is, rolling out the Anniversary Update in a measured fashion was part of the plan.

“When problems come up—as they must have with the webcam issue, or the even more recently discovered Kindle issue—then Microsoft can also shut off the spigot on what are now known-bad configurations,” I wrote. “And do so until those issues are fixed. This results in a high-quality experience for everyone. Those people who have known-good PC configurations will get the update and should see positive results. And as more data about all upgrades arrives, more PC configurations will be added to that known-good list. And yes, some may be added to a known-bad list because Microsoft has found issues too. The result? A slow but measured and reliable roll-out.”

Today, the Windows 10 Anniversary Update still isn’t completely rolled out, because there are some PC configurations out there with some mix of incompatible device drivers. But I suspect that we’ll finally be done with this release by the end of the year, for the most part. Because we’re already looking forward to the Creators Update for early 2017.

I’m sure that one will go just fine.

But wait, there’s more

So those were, in my mind at least, the biggest Windows 10 stories of 2016. But there was so much else going on.

Somewhat related to the Anniversary Update woes noted above, I opined that Windows 10 servicing overall, or what Microsoft calls Windows as a service, just isn’t working. “We’re all in a perpetual beta, where the speed of these updates and the explicit understanding that they will always be followed my more updates, means that quality control can lapse,” I wrote. “If Microsoft screws up an update, no worries: They can and will just patch it again, because they can. And patch it and patch it and patch it. Which they have.” Countering this, Microsoft explained that Windows as a service was in fact necessary, and that, as a result, Windows 10 version 1607 is now the most secure Windows version yet.

Microsoft unveiled its audacious plans for the future of Xbox gaming in 2016, and a major component of that plan was to pull a classic strategic move and change the rules of the game. “Leverage Windows 10 and the firm’s decades-long dominance of the PC market to give another platform a new lease on life.” For gamers, this means all kinds of things, including, best of all, the availability of Xbox Play Anywhere games you can interchange between the Xbox One console and Windows 10 PCs and tablets.

Crapware continued to be an issue in 2016. I’ve argued that crapware continues to ruin the Windows 10 experience, and saw how the Signature PC program at Microsoft presented an obvious answer to this problem. But Microsoft has a way of surprising me sometimes. Concurrently to the release of the Anniversary Update, Microsoft provided a new free tool called Refresh Windows, which works just like the Reset this PC tool in Windows except that it provides a clean version of Windows, not the one loaded down with crapware that was provided with your PC. Nice.

Windows 10 still has other challenges. Microsoft is again escalating the use of advertising in its flagship OS, which is disappointing. Its UWP apps platform is going nowhere fast. And assclowns like Tim Sweeney and Forbes’s Gordon Kelly (to whom I am not linking) continue to spread FUD about Windows 10.

But there are bright spots too. The built-in UWP apps in Windows 10 are mostly excellent and are getting better all the time; they’re the one part of Windows as a service that is working quite well, thank you very much. And while Centennial is a mixed bag for UWP success, it’s fair to say that this important technology—which wraps desktop applications in a secure and reliable UWP shell—is making it easier to move legacy apps forward. The availability of full Evernote and Photoshop Elements in the Windows Store, for example, is amazing.

Speaking of UWP, Microsoft Edge remains a mixed bag. Microsoft’s new browser is today responsible for just 5.21 percent of all desktop web browsing usage, a far cry from the 56 percent commanded by Chrome, or Firefox’s 12 percent. Part of the problem is that the browser functionality is lackluster: Microsoft doesn’t update it enough, and the extensions market has never really materialized. One bright spot: Edge has a significant battery life advantage over all of its rivals.

Cortana, another key new Windows 10 feature, is likewise in a strange holding pattern. Microsoft personal digital assistant is actually updated regularly, unlike Edge. But it’s still available only in a limited set of markets. And the PC hasn’t proven to be a great platform for voice-activated usage. To counter this, Microsoft can and should create a “Cortana cube” (or whatever). But it is not doing so, contrary to rumors. As a result, Cortana’s future is cloudy, as Brad noted. This is an important and unacceptable strategic mistake, I think.

Finally, while this didn’t get a lot of press in 2016, I think the slow uptick of another new Windows 10 technology, Windows Hello, is worrying. But there are some peripherals out there, thankfully. And in late 2016, Microsoft even started selling some through its own retail store. That’s progress.

Looking ahead

As big as 2016 was for Windows 10, 2017 could prove to be even bigger, with the Creators Update in Spring 2017 and then a second major update, called “Redstone 3,” set for later in the year. And the arrival of full Windows 10 on ARM could rejuvenate Microsoft’s mobile strategy. Best of all, some of the biggest issues with Windows 10 in 2016—Upgradegate and the problematic Anniversary Update rollout—were one-time events and shouldn’t be repeated next year. (It’s always possible that Microsoft will botch the Creators Update rollout; cross your fingers.)

But I’m hopeful for the future. I’m excited by what’s happening with Windows 10. And I’m looking forward to another go-round. I bet many of your are as well.