Microsoft Moves Windows 8 Forward

by Mark W. Hibben

Divergent Reactions

Microsoft released a preview of Windows 8.1 today, 6/26, which contained some desirable new features and improvements. Is 8.1 going to "save" Windows 8? No, but Windows 8 really didn't need to be saved.

The reactions to 8.1 have ranged from Adrian Covert's characterization of it as a "huge improvement" to fellow Motley Fool blogger Sam Mattera's as "fundamentally broken". After downloading and installing 8.1, I remain in the middle of the two extremes. I see Windows 8 as an innovative, but unfinished, experiment. Win 8.1 moves the OS a little closer to what I consider its final destination, and that's a good thing for Windows users.

The best new feature is a new Start Button. The Start Button doesn't quite work the way many had hoped, since it's not a reversion to the desktop Start Menu of Windows 7. Instead it serves as a shortcut to the Home Screen, which is still Microsoft's intended path for launching all apps.

What's nice about the Start Button is the fact that it works the same regardless of where you are in the OS, or in an app, or in desktop mode. The uniform start button provides a sense of convergence that had been missing in the Windows 8 split personality of Modern UI (MS-speak for the tile based interface) and the more traditional Desktop.

This split personality has always been my major qualm about Win 8, both from the standpoint of user interface design and as a business strategy. Microsoft set out to design an OS that would work on both tablets and non-touch screen portables and desktops. One has to wonder if this was not a bridge too far, but Windows 8 does anticipate a future of ubiquitous touch screens.

That touch screen future is something I fundamentally accept, despite being a supposedly biased iOS developer, and I admire Microsoft for being willing to reach for that future. In the process , Microsoft created an OS in which the Desktop mode and Modern UI (MUI) merely coexist. It was as if there simply wasn't time and money enough to merge the two modes in a more thoughtful and complete way.

The down turn in PC sales that was identified in the reports of IDC and Gartner do indicate some negative impact from the unfinished character of Windows 8. On May 7 Microsoft announced that Windows 8 had sold 100 million copies, a respectable number that to me indicates that Windows 8 doesn't need to be "saved", but the number wasn't great news either. As Woody Leonhard pointed out in InfoWorld, Win 8 monthly sales have fallen to about 10 million copies/month from first month sales of 37.5 million. In contrast, Win 7 sold about 20 million copies/month for its entire lifetime.

8.1 is unlikely to impact sales numbers significantly, although it does makes transitioning between the MUI and Desktop modes a little less jarring. But true convergence is still in the future. Is it even possible to converge the modes into a single unified OS and user interface that works well in all scenarios? I believe it is and that Microsoft, Apple and Google are all working towards it.

Backing off from MUI

For Microsoft, it may be a matter of backing off somewhat from the radical departure that is Modern UI. In creating Modern UI, Microsoft dumped two fundamental user interface paradigms that have been difficult to satisfactorily replace:

1) Resizable, movable app windows. Yes I know they're available in Desktop mode, but they're banished from the MUI. Instead, MUI allows split screens, and with 8.1, even multiple split screens. But it still feels too restrictive, and the screen can only be split into narrow vertical panels. MUI's split screen mode is a poor substitute for resizable windows. Such Windows were banished from mobile operating systems because of screen size constraints and lack of processor power to run concurrent apps. But resizable windows are perfectly achievable within a large touch screen interface.

2) Pop-up menus. Menus actually work very well for mobile operating systems, and both Apple's iOS and Google's Android make extensive use of them, yet they are also banished from MUI. The UI designers at Microsoft aren't budging on this one, which is why the new Start Button doesn't bring up a menu.

Google Keeps it Simple

Google has perhaps demonstrated the greatest flexibility regarding integrating touch screen and traditional cursor modes. Track pad or mouse driven cursors have been available on Android tablets for some time. Google makes it look so easy and natural that you have to wonder why the other guys make such a fuss over it. Touch the screen, the device responds in normal tablet mode; touch the track pad, and a cursor instantly appears.

This flexibility also carries over into Chromebook Pixel, which also has a touch screen and a track pad. On screen controls can feel a little small for fingers, but Pixel has all the essential elements of a converged touch screen experience: app windows that can go into full screen mode if the user wants, efficient use of menus, and touch screen functionality combined with a trackpad.

Apple's Deep Integration

Apple has probably done the best job of integrating touch-based gesture support of all three of the major PC OS players. Here, I'm talking about Mac OS, not iOS. Apple's Magic track pad works better than any other track pad I've tried, and the gesture support is comprehensive and intuitive.

Apple has even integrated into Mac OS full screen app capability which works along side normal windows. I consider Mac OS to have solved the problem of uniting a full screen app experience as in MUI with the traditional windowed app approach. Any app capable of full screen mode can be instantly switched back to windowed mode.

In almost every way, Apple has provided in Mac OS the converged UI solution, except one. They won't take that last step and actually build a touch screen equipped Mac. But they're thinking about it, as demonstrated by a recent patent application filed by Apple designer Duncan Kerr for a laptop with a removable screen.

Merely thinking about it isn't good enough, and I hope that the touch screen potential of Mac OS is not held back much longer. Unfortunately, I saw nothing encouraging at WWDC in this regard.

Despite objections from Apple fans that Macs don't need touch screens, I consider Apple's adoption of touch screens for Mac OS to be inevitable. And Windows 8, in its own halting and inelegant way, will have led them into the future.


Do Gartner's Smartphone Numbers Add Up?

by Mark W. Hibben

Virtuous Cycles

Gartner released a report yesterday (6/24) detailing its projections for computing device sales for 2013 and 2014. Up until now, I've considered Gartner to be a reliable source of information on mobile device sales and market share, but after taking a closer look at their latest numbers, I have serious doubts about Gartner's data for 2012 as well as their projections for 2013-14.

I've been writing about the Mobile Device Wars for several years and have kept running tallies of device shipments as reported by Google, Apple, as well as Gartner. I follow the reports closely because I consider market share to be key to winning the Mobile Device Wars.

Numerous virtuous cycles take hold as the pool of users for a particular ecosystem increases. The larger the user base, the more prospective customers there are for third party apps and accessories. This increases the level of interest in the ecosystem by businesses engaged in app or accessory development. App and accessory availability increases, thereby making the ecosystem more attractive to users, and so on.

As much as Apple management would like to downplay the importance of market share, both Apple and Google benefit from the virtuous cycles taking hold in their respective ecosystems, while competitors find establishing healthy ecosystems increasingly difficult.

Thus, when I read over Gartner's report, I was struck by a couple of disconnects in their 2012 data. In the report, Gartner states that number of Android devices shipped in 2012 was 505.509 million and that the number of combined iOS/MacOS devices shipped was 212.878 million.

Falling of a Log

Let's take the Apple device number first. Getting this right should be falling off a log easy because Apple publishes exact sales numbers every quarter for iPhones, iPads, iPods, and Macs. But Gartner doesn't even get this right, as summarized in the table below (all numbers in millions).

The table shows the number of iOS devices and Macs Apple sold each quarter, as detailed in their quarterly earnings reports. Not only did Gartner not get the iOS device total correct, it looks as though they completely forgot about Mac sales. The correct number for combined iOS and Mac sales should be 234.77 million.

Is there any way to rationalize the discrepancy? I doubt it. Although Gartner's report doesn't explicitly state it, their numbers represent shipments to end users, as I'll discuss below when I go over the Android data. My understanding is that is what Apple reports as well. But even if Apple were adding in shipments into retail channels, I doubt this would be a significant difference when summed over an entire year.

Real Disconnect

But the real disconnect that struck me was for the Android numbers. Gartner's assertion that 505.509 million Android devices had been shipped in 2012 is at odds with Google's announced Android activation data. Google periodically announces total numbers of Android devices activated to date, as when Larry Page announced that 750 million Android devices had been activated as of March 13 of this year. The number of devices shipped in 2012 is the difference between the cumulative total as of the end of 2012 and the total at the end of 2011. That difference is about 364 million, not the 505.5 million reported by Gartner. Gartner's error for 2012 Android device shipments is at least 39%.

I consider that the 2012 total activations represents an absolute upper bound to the number of Android devices shipped to end users in 2012, and certainly an upper bound to the total number of Android smartphones shipped in 2012, since activations includes all types of Android devices.

This is why I included the Gartner reported Android smartphone shipments to end users in the first column. The total for Android smartphones of 446.8 million, according to Gartner, is well in excess of Android activations for 2012. However, the smartphone total is consistent with the total of 505.5 million Android devices in Gartners 6/24 report, since Gartner estimates that smartphones represent about 90% of all Android device shipments. This also confirms that the 6/24 report data represented shipments to end users.

Systematic Error

In order to see if there has been a consistent pattern of overestimation of Android device shipments by Gartner, I went back through their smartphone shipment reports as far back as 2009 and have plotted the data below. Note that the curves represent cumulative device shipments or activations.

Also note that I've provided a curve fit to the Google activation announcements in order to interpolate between announcements, since Google management make announcements at irregular intervals. There's always some uncertainty about the time phasing of the Google activation announcements. For instance, Larry Page's blog of March 13 didn't quite come at the end of the quarter, so there's still some uncertainty about when the 750 million activations became effective. For simplicity, and because the potential errors are small, I always treat the announcements as effective as of the end of the relevant quarter.

The cumulative Android smartphone shipments according to Gartner are way above the total Android device activations as of Q1 2013. Gartner's errors are more difficult to detect on a quarterly basis, but it becomes apparent that these errors are systematic when you plot cumulative results. Conversely, Gartner slightly, though consistently, underestimates iPhone shipments.

The Take-away

In order to accept that Gartner's Android shipment data are valid, one has to assume that Google is deliberately low-balling their activation number announcements. This strikes me as highly implausible, given how proud Google management are of their activation tallies when they do announce them.

In fact, I regard Google's activation claims as somewhat suspect, since, as I have previously pointed out, there is a potential to activate the same device multiple times. But I certainly don't believe numbers that are substantially in excess of the announced activations. It would be great for investors if Google and Microsoft followed Apple's lead and announced product shipment data on a quarterly basis.

Google's Android is still trouncing Apple's iOS in market share, but the margin of victory is smaller than Gartner would have us believe. As of the end of Q1, Apple had sold a total of 523 million iOS devices, while at most 750 million Android devices had been sold.

If Gartner's numbers for Android device shipments in 2012 were in error by such large amounts, should we believe their prediction of 866.8 million Android devices shipped in 2013? In a word, no.

Adobe's Candy Coated Earnings Report

by Mark W. Hibben

Don't Ask, Don't Tell

Until Adobe's fiscal Q2 earnings were released last Tuesday, I had never before encountered a presentation so short on traditional financial metrics but long on invented metrics. It was as if Adobe's management wanted to distract investors from certain inconvenient truths. Judging by the market reaction the day after the report was released, Adobe succeeded, if briefly.

Adobe's slide presentation contained the bare minimum of hard financial data: quarterly revenue and earnings per share (GAAP and non-GAAP). Investors interested in operating profit, gross margin or any kind of year over year comparison were forced to pore over a supplementary table. I doubt that many did. Even Microsoft, which at times I have accused of obfuscation, provides more concrete information in their earnings slides.

It's as if Adobe regards investors as mere children, to be distracted or placated with colorful lolipops of data. These lolipops took the form of estimates of annualized subscription revenue for things like Creative Cloud, Adobe's cloud-based successor to Creative Suite. Adobe could easily have provided actual quarterly revenue for Creative Cloud, as it did for Document Services and Marketing cloud, but this was never revealed even in the supplementary data table.

These estimates of Annualized Recurring Revenue (ARR) were scattered liberally throughout the earnings slides and even the supplementary data table as if they were accomplished fact rather than forward-looking estimates. Even if we accept these ARR's as accurately predictive, it virtually impossible to know what they really mean, since Adobe provides no basis for comparison.

For instance, how does the ARR for Creative Cloud compare with revenue in previous years for Creative Suite? How does Creative Cloud revenue for the current quarter compare with Creative Suite revenue for the current quarter? I guess it's best not to ask.

The Inner Core

Best not to ask, because the inner core of reality is just so much worse than the candy coating. But it's there if you really look for it in the data, which I've summarized in the chart below.

Quarterly revenue has fallen year over year by 10%, but what is worse, operating income and net income are down by 63% and 65% respectively. Operating margin has collapsed from a reasonable 27% in Q2 2012 to just 11%.

Adobe was very proud of the fact that Creative Cloud subscriptions jumped from 479 K in Q1 to 700 K in Q2, but it's not clear why, since it comes on the heals of Adobe's announcement that it is pulling the plug on future licensed versions of Creative Suite. The announcement provoked resentment among current Creative Suite users, as documented in a recent CNET article. More than 30000 people have signed an on-line petition urging Adobe to change its policy on

I'm pointing out the deficiencies in Adobe's financial performance not to argue that Creative Cloud is a failure but that it's success has not been demonstrated. It could work, since it has the advantage of lower initial cost for users. What hasn't been demonstrated is whether it's revenue neutral, let alone revenue enhancing. Adobe is engaged in a massive experiment that other software companies, notably Apple and Microsoft, have been unwilling to undertake.

Are Subscriptions the Future?

It's one thing to distribute software over the Internet. This is something Apple pioneered with the App Store for iOS devices. But here Apple has the advantage of its relatively secure and protected ecosystem of devices and can provide a seamless buying experience with one click shopping and almost instant downloads. At WWDC in June, Apple celebrated the App Store's 5th year. The App Store has over 575 million accounts, most with credit cards and simple one-click purchasing. Customers have downloaded 50 billion apps, and Apple has paid developers $10 billion.

Clearly, this is the future of software distribution. Apple is currently shifting Mac OS software to Internet distribution as well with the Mac App Store, eliminating optical disc drives from its most recent computers, as well as distributing its most recent Mac OS version exclusively via download.

But not a single app is sold on a subscription basis, excluding things like on-line magazine subscriptions. Apple now reports all digital content sales revenue, including iTunes, the App stores and traditional software media in its iTunes/Software/Services (iTSS) category. In Q1 iTSS made $4.1 billion, about what Adobe expects in revenue for its entire fiscal year.

At WWDC Apple announced a beta of iWork in iCloud, but it isn't clear whether Apple even intends to charge for the service.

Although Microsoft is still in the process of establishing its mobile device app and content stores, most of its extensive software offerings are availabe for digital download. Where Apple has tended to shy away from web apps, Microsoft has embraced them fully with Office 365 as well as its developer tools for cloud based apps, Windows Azure.

Office productivity suites such as Office 365 have relatively modest computing requirements, so they've become a popular focus of web app development. But Microsoft hasn't tried to cajole Office users to abandon conventional licensing for cloud subscriptions. Given what happened with Xbox One DRM, I doubt that Microsoft has the stomach for any more consumer backlash.

In its last Q1 earnings report, Microsoft didn't have a lot to say about Office 365, except that it had "momentum", a word that frequently stands in for less flattering characterizations in MS-speak. I doubt that Office 365 subscription revenue will be reported by Microsoft any time soon. But for Microsoft this is okay. It's understood that MS isn't betting big on 365.

Assessment Difficulties

Adobe, however, is betting big on the cloud, and Adobe investors owe it to themselves to make the best assessment they can of the success of this bet. Adobe has made this even more difficult by a recent reorg, so now there's effectively no way to perform year over year comparisons by segment. The best investors can do is continue to track Adobe's fundamentals and watch for a trend, either up or down. Both operating income and net income were up slightly from last quarter, but I don't consider the change large enough to be significant. However, it income begins to recover to 2012 levels in Q3 and Q4, that would be a good sign that Adobe's Big Bet has begun to pay off.