Programming Windows: Success (Premium)

Windows 7 couldn’t have arrived at a better time for Microsoft: 2009 was a terrible year for the PC industry and, thanks to its financial reliance on the Windows business, it was a down year for Microsoft too.

One day after announcing that Windows 7 and Windows Server 2008 R2 had reached the RTM (release to manufacturing) milestone in July 2009, Microsoft also announced its quarterly earnings, which sunk 17 percent year-over-year (YOY) because of “weakness in the global PC and server markets.” Even its annual earnings—Microsoft’s fiscal year ran from July 1, 2008 to June 30, 2009—were down, in this case by 3 percent.

The next quarter wasn’t much better: In October 2009, Microsoft reported that its earnings were down 14 percent YOY, though an incredible $1.47 billion—or 10 percent of its total revenues otherwise—were deferred as part of the Windows 7 Upgrade Option program and sales of Windows 7 to PC makers and retailers ahead of the product’s general availability (GA). Were it not for this deferral, revenues would have declined just 4 percent YOY.

But Microsoft had weathered this difficult time with the understanding that Windows 7 would right the ship. And despite it being a minor functional upgrade over Windows Vista—a set of updates the previous Windows organization might have given customers for free—Windows 7 was indeed the cure for Microsoft’s ills. And with most enterprises moving at a glacial pace, they could now prepare massive migrations from Windows XP to Windows 7, skipping the poorly received Vista entirely.

Microsoft was ready to move on as well. And with the Windows 7 launch completed, it did just that. At a mid-November annual shareholders meeting just three weeks later, CEO Steve Ballmer revealed record sales of Windows 7, which he said had “already sold twice as many units than any other operating system we’ve ever launched in a comparable time.”

The software giant also got some good news from the European Commission (EC) in December when it revealed that it had accepted Microsoft’s proposed solution to its anticompetitive product bundling practices in Europe: among other changes, Microsoft would allow PC makers and end-users to disable Internet Explorer and make another web browser the default. But the software giant was eager to put that nightmare behind it too.

During his CES 2010 keynote address in early January 2010, Steve Ballmer claimed that Windows 7 was “the fastest-selling Windows version ever” and “the fastest-selling operating system in history,” with a 94 percent satisfaction rate. But he provided no hard numbers beyond the 300 million PCs sold in 2009—75 percent of which shipped before Windows 7 arrived—and the over 4 million available Windows applications. He also noted that it took 3,000 Microsoft engineers and 8 million beta testers to bring Windows 7 to market.

But the hard numbers would come. And they would exceed everyone’s expectations.

In April, Microsoft revealed that Windows 7 had sold over 60 million units in its first quarter of availability. As a result, Windows revenues rose 28 percent YOY in the quarter and Microsoft announced record quarterly revenues of $14.5 billion, up 6 percent YOY. Windows 7 sales would then rise to 90 million units just two months later.

During all this time, the Windows team kept quiet as it plotted the next major release of Windows. But there were minor related releases like Windows Live Essentials 2011 and Security Essentials 2, which Microsoft marketed as “completing the Windows 7 experience.” And there was muted speculation about a Service Pack 1 (SP1) update for Windows 7, though that would end up being only a collection of previously released updates and no new features. (The SP1 release for Windows Server 2008 R2, Windows 7’s sister product, would be more substantial, with major new virtualization and Virtual Desktop Infrastructure features.) Microsoft finally released the first beta of Windows 7 SP1 during its Worldwide Partner Conference (WPC) in July, and the update was finalized the following February.

Just over one year after Windows 7 shipped to the public, in November 2010, Microsoft announced that it had sold 240 million licenses of the product, kicking off a curious and legally dubious period in which it would routinely sell exactly 20 million Windows 7 licenses every single month for the duration. But no matter how Microsoft smoothed its accounting, the reality was clear enough: Windows 7 was—and would remain—a huge success, and it was popular with Microsoft’s consumer and business customers. Microsoft revealed that there were over 1.2 billion Windows PCs in the world and claimed that the PC market would see double-digit growth in 2010 and 2011.

A lot had changed in three years.

In late 2007, the year after Microsoft had finally shipped Windows Vista, the software giant told me it was “realigning” for the post-Windows XP Service Pack 2 (SP2) world and would more aggressively push hotfixes via Windows Update rather than wait for service packs, a tacit acknowledgment that Vista needed help as quickly and as often as possible. Worse, Vista had shipped with over 80 “enterprise blockers,” issues that would prevent key business customers from upgrading, and there were only 250 Vista logoed applications. These issues were fixed by late 2007—when there were over 2000 logoed applications and all the enterprise blockers had been fixed—but it was too late: most businesses had already decided to skip that version.

“This won’t shock anyone, but the mood at Microsoft this time around is completely different, and is almost giddy,” I wrote in late 2010. “Customers of all kinds are excited about Windows 7, and in workplaces, employees are actually demanding it. Based on our own research here at Windows IT Pro, and via Microsoft and its analyst group partners, businesses really are deploying Windows 7, and at an ever-increasing rate of speed. Comparing Windows 7 one year later to the situation with Vista is like night and day. There really is no comparison.”

Internally, of course, Steven Sinofsky’s Windows team was plotting the next major release of Windows, which was imaginatively codenamed Windows 8. This was happening secretly, but I was the recipient of a massive dump of internal documentation in June 2010 that revealed some of the team’s plans. The documentation dated back to a set of April meetings between the Windows team and its PC maker partners, and it indicated that the Windows team would begin coding Windows 8 in June, after a few more months of internal discussions.

At a high level, the partners were told that Windows 8 would follow the same release schedule as did Windows 7, with a total time to market of three years and the same basic milestones: M1, M2, and M3, after which the product would be feature complete, a single Beta, at which it would achieve a higher quality bar, a single Release Candidate (RC), at which all language versions would be ready, and then Release to Manufacturing (RTM).

Microsoft would likewise try to emulate what it had done successfully with Windows 7—it was particularly proud that it had “consistently adhered to” its release schedule and had released high quality builds—while fixing some of the issues that had come up; for example, the secretive disclosure model employed by Sinofsky limited the availability of PCs designed for Windows 7 during the 2009 holiday selling season, a mistake the team would try not to make with Windows 8 by engaging with them even earlier in the process.

One of the more interesting functional changes coming in Windows 8 was a new Windows Store, sometimes called the Windows App Store, that would help consumers discover relevant and trustworthy applications and provide flexible licensing and monetization for developers.

Tied to the Store, Windows 8 would provide a new application model that would leverage Windows capabilities in what it called “tailored web applications.” The goal was to make Windows app development more approachable to hobbyist developers than was possible with the traditional Windows SDK, and to make the resulting apps easy to maintain and distribute. These tailored apps, Microsoft believed, would have higher engagement and higher reach than traditional apps like Skype, Word, and Acrobat Reader.

This was an interesting shift. To this time, Microsoft had pushed the notion that so-called “rich” native applications were more powerful than web or mobile apps, and that they were thus technically superior. But the thinking now was that “engaging is the new rich,” meaning that they were shifting to engaging users with apps that were closer to web and mobile apps in nature. These types of apps were seen as being simpler to install, update, and remove, and they were often powered by Internet services.

What form this app platform would take was at that time was undecided. But Microsoft was investigating two primary options: HTML 5 plus JavaScript (web apps), and Silverlight, a .NET technology that had started life as Windows Presentation Foundation/Everywhere (WPF/E) and had found great success to date on the web. Microsoft had also used Silverlight for Windows Phone development. (Because that platform was driven largely by former members of the Windows team, Sinofsky’s team ended up going in a different and incompatible direction.)

Related to this, Microsoft claimed that only 40 Windows applications—or 3 percent of the total—had over 2 percent usage share across the entire userbase. And that over 50 percent of the top 100 applications were non-web-related “by nature,” including productivity apps, imaging and printing apps, and games. “Application churn in the tail,” it said, showed “user desire for application discovery and trial.” In other words, Microsoft was making the case that a new app store and platform made as much sense in Windows as it did on mobile.

Windows 8 would also sync Windows settings and apps via a Microsoft or corporate account. That way, when users signed into a PC with such an account, their settings would be applied immediately and they would gain access to apps they had purchased or downloaded elsewhere. It would include the “Orion” location platform with Wi-Fi triangulation and GPS capabilities, and better support for modern sensors like infrared (IR), ambient light, accelerometer, human presence/proximity, system orientation, and more. It would also support new connectivity standards like USB 3.0 and Bluetooth 3.0, and Microsoft was considering deprecating IEEE-1394 (Firewire) and wanted to see PC makers phase out USB 2.0.

Microsoft would also introduce new ways to sign in to a PC with Windows 8: the “three-fingered salute”—typing Ctrl + Alt + Delete—was not appropriate for certain new form factors. This would include fingerprint improvements, facial recognition, and presence detection. That latter feature could work in tandem with the first two and detect when you walked up to a PC, turned it on, and walked away from a PC, signing you out and putting it to sleep.

Microsoft was also pushing PC makers to care about all segments of the PC market—“from low-end to high-end, there is a Windows offering for everyone”—and to embrace modern and new form factors like Slate PCs that would emulate the iPad, as well as emerging markets, like China, which Microsoft said would be the world’s biggest market for PCs by 2013. It was also showing what was then a surprising interest in the nuts and bolts of the PC building business and asking about ways to decrease the cost of making PCs, a line of inquiry that should have alarmed these partners. (They would later come to understand why when Microsoft launched its own line of Surface-branded PCs in 2012.)

Microsoft also offered some predictions for 2012, when Windows 8 would ship. Netbooks would grow in size and capability, blurring the line between this low-end product and more traditional laptops. Dual-core PCs would still be the highest volume part of the market, but quad-core and even 6-8 core PCs would increase. And the majority of new laptops and desktops would ship with more than 4 GB of RAM.

In conclusion, Microsoft told its partners that 6 trends were guiding its Windows 8 planning:

Explosion of form factors. New innovations would trigger new form factors and power unique new consumer experiences.

Assumed connectivity. The software + services world would power end-user scenarios and the health of the PC industry.

The collision of enterprise and consumers. The consumerization of IT was leading to a world in which users would have a seamless experience across their personal and professional lives.

Content is personal. Content consumption experiences would be highly tailored to user devices and context.

Diverse and vibrant ecosystem. Windows was uniquely poised to help PC makers grow and healthy hardware + software ecosystem.

Fundamentals. Customers would continue to make buying decisions based on their perceptions of quality, performance, reliability, and security.

This all seemed reasonable at the time. But what Microsoft’s partners didn’t realize was that Steven Sinofsky and his Windows team would take the threat from Apple’s iPhone and iPad a bit too seriously. And that, as a result, Windows 8 would not be a well-regarded and thoughtfully designed release like its predecessor. In fact, Windows 8 would be nothing short of a disaster, an embarrassment to Microsoft, and the primary cause of a years-long decline in the PC market.

We will get to all of that soon.