I know you’ve been waiting for a comment on the Microsoft Surface impairment charge ever since Microsoft’s earnings report. Well, I’m going to give my view of the problem and how Microsoft turned what should have been a shining moment in its transition to a Devices and Services company into a debacle. There will be no dwelling on the Surface RT hardware (and particularly ARM) debates nor on Windows 8 shortfalls, except for limited reference. Those things play a role, but I believe a fairly minor role in how Microsoft messed this up. The problem, at its surface (pun intended for a change), is that Microsoft let a desire for Apple-like margins on hardware cloud its judgment on how to enter the PC hardware business.
The other day a friend of mine made two astute comments. The first was that the only real problem with the Surface RT was that it lacked applications. The second was that if Microsoft had given away the millions of unsold Surface RT systems implied by the $900 million write-down then the application problem would have been solved. So why didn’t Microsoft do exactly that? That is, sell several million Surface RT systems at bargain basement prices to seed the market? I’ll get into that as I walk through this story.
Before getting into Surface specifics I want to discuss a general point about Microsoft over the last decade and moreover the last few years as it started the transition to a Devices and Services company. As a (near) pure-play software company Microsoft has had absolutely enviable margins. The kind of margins that make for business school case studies. Put simply the Microsoft software business model is very high fixed costs and near-zero variable costs. So as you drive volumes higher you reach a point where every sale is nearly pure profit.
Software profits have been so high over the years that they allowed Microsoft to invest in many unprofitable products (or even outright failures) while still maintaining profit margins that were the envy of the entire business world. In the 90s no one even noticed. In the 00s they noticed, but were more critical of the lack of new successes than of the margin impact of trying to find one. In the 10s you are seeing more calls to split or drop all the new endeavors and remake Microsoft as an Enterprise-focused software company, thus returning it to historical margins at the risk of the likeliness that the “Consumerization of IT” will eventually render it irrelevant.
The problem for Microsoft today is that if you transition the company away from its primary revenue stream being sales of perpetual software licenses to a Devices and Services revenue stream you stop being able to hide behind the tradition software license sales margins. So Microsoft, and I’ll put this very much (and very appropriately) at Steve Ballmer’s feet, is extremely focused on maintaining margins as much as possible during the transition.
Take Office 365 as an example. Without getting into all kinds of details of the dynamics of the Office business and how a services offering changes them, I’ll just point out that Microsoft is adding a lot of variable cost (running the service) to what used to be a purely fixed cost business. And the fixed costs (e.g., Engineering) don’t go down at all. So when you look at the complex set of offerings under the Office 365 umbrella what you see is Microsoft trying to address cost-sensitive parts of the market, value for $ parts of the market, and high-end information worker parts of the market and trying to get the mix of subscriptions to end up margin neutral. Or at least close enough that when the transition is done shareholders will be happy with the resulting margins. So far they seem to be on a path to success with this one.
Bing is another case of a focus on margins. While there are very strategic (survival even) reasons for being in the Search space, the business case is that once you achieve a particular market-share the advertising revenue hits the knee of the curve and you quickly get to very high margins. So, if you have the money, you can pour immense amounts into scratching your way to that magical market-share (whatever it may be) knowing the eventual positive bottom line impact will be enormous. If, and it remains a big if, Microsoft hits the knee of the curve on advertising revenue then the many billions spent to get there will quickly be forgotten. For now they are getting close to a break-even run rate while making good use of Bing as a platform technology.
But Surface and future mainstream device offerings are a far more risky venture for Microsoft. Before Microsoft announced that it was becoming a Devices and Services company you could look at the Surface and declare it a “North Star” product. That is, it wasn’t really intended to make money but rather to excite customers and show OEMs the way. But if you are a Devices company then you need to make money selling devices! And because devices are very high revenue per unit, they can quickly overwhelm your software revenues and have significant margin impact. You can look at Apple, and to some extent Samsung, and see that it is possible to get high margins from devices. And that led Microsoft, which has always had a bit of Apple-envy in its DNA, to fall into a trap.
Microsoft, in entering the devices market, could either go for market share or margin. They went for margin. According to IHS iSuppli, the gross margins for the Surface RT were (based on original pricing) higher than the Apple iPad. And they did that at exactly the same time that overall tablet prices were being driven down by the introduction of 7-8″ devices, often with a subsidized price. If Microsoft had actually been able to sell several million Surface RT’s at its original pricing then this would have looked like a brilliant move and Steve Ballmer would have been a candidate for CEO of the year. But it turns out the “go for margin” strategy completely failed.
I’d originally expected a competitive, but not aggressive, list price for the Surface RT that, when combined with a service subscription, made the device seem inexpensive. Microsoft didn’t do that. I’d also expected a price in the ball park that the Surface RT came out at with a price reduction once supply became plentiful. Microsoft never lowered the price. Actually my main expectation was that after the holiday season Microsoft would effectively lower the Surface RT price by package the Touch Cover with it at the same price as the Surface RT-alone had been for the holiday season. They didn’t do that. Instead they maintained the effective price at about $100 more than the iPad through FY13, and then took a write-down as a result of a re-pricing for FY14 that did nothing more than take the Surface RT to the price it should have been at introduction!
The write-down, aka impairment charge, is an abomination in my opinion. Why did Microsoft take it? Well, we are back to the margin point. By re-pricing the Surface RT to where it should have been in the first place Microsoft eliminated the possibility that it would ever be a high margin product. They could have taken the hit in reduced overall margins for the first two quarters of FY14, which would have been a fair representation of the impact becoming a Devices company is likely to have on the business in the short to medium, if not long, terms. Instead they pulled the entire margin impact into Q4FY13 and took the PR and investor hit all at once. This will make FY14 look artificially good. Most likely from Microsoft’s perspective they hope to attain good margins on Devices in the second-half of the fiscal year and just wanted to avoid more margin discussion until then. But personally I’d rather see a continual true representation of the impact of Devices on the business quarter-to-quarter instead of using accounting tricks to hide it.
Back to the Surface RT debacle itself let’s focus on the weakness of introduction strategy. The uniqueness of the Surface RT tablet is that it is (a) tablet for running Metro apps, (b) runs Microsoft Office, and (c) has a keyboard and track pad built into a thin cover. So basically, along with other Windows RT tablets, you have a tablet that will let you run Office and be more productive on content creation tasks than the iPad or Android tablets. But at introduction it had almost no Metro apps, and nine months later the library is growing fast but is still not compelling. So effectively you would pay $100 more (than an iPad) for a tablet that can run Microsoft Office, but can’t run any other applications.
Microsoft Office is indeed a compelling capability on a tablet. So is a keyboard and track pad. But, Microsoft missed on the tablet version of Maslow’s Hierarchy of Needs. Near the base of pyramid that represents the hierarchy is “Run the applications I want”. A few levels up is “Does real work” (which for many people means runs Microsoft Office). And the way Maslow’s hierarchy works is that you have to satisfy the needs nearer the base before people care about things more towards the top. A tablet that doesn’t “Run the applications I want” is a non-starter despite the ability to run Microsoft Office. And one that requires I pay a premium for the privilege of not meeting my base needs is truly DOA.
Microsoft apparently expected that the Metro application library would grow so quickly that Maslow’s tablet hierarchy would be a non-issue. But they shot themselves in the foot with their strategy to achieve this. No, I don’t mean that Windows 8 sucks so app vendors are ignoring it. On the contrary, app vendors offer near universal support for Windows 8. The mistake Microsoft made was to assume that support for Windows 8 meant the creation of Metro applications while Microsoft made it perfectly acceptable for Windows 8 support to mean desktop Win32 applications and IE10-tuned websites. Microsoft gave App vendors (and I use the modern term broadly to also include traditional ISVs and website operators) almost no real incentive to create Metro apps rather than continue down the same path they were on for Windows 7.
Think about Microsoft’s initial spiel to developers: There are already hundreds of millions of PCs installed that will be capable of running Metro applications (because they are upgradeable to Windows 8). That takes away the disincentive of the market for Metro apps being too small, but does nothing to incent developers to create Metro apps. So today there are about 100 Million PCs actually running Windows 8, and perhaps 98 Million of them also run Win32 Desktop apps. So the real market, the one you can’t reach in other ways, for Metro apps is only 2 million systems.
The reason we have 110,000 Metro apps in the Windows Store is because of the business advantages the Store model brings for small developers. Such as not dealing with installation, uninstallation, payment/billing, automatic updates, etc. But for medium and larger developers, the incremental market reach of doing a Metro app pales in comparison to the costs associated with maintaining a desktop application (for Win 7 for example, or for power users) in addition to the Metro app. So the Windows Store is largely populated by apps from small developers, with only a modest subset of apps from larger developers and website front-ends. This is a speed bump for Windows 8, because users will continue to use the desktop applications. But it is devastating for the Surface RT and other Windows RT systems as they cannot run Win32 apps.
Think about other examples of confusing messages from Microsoft. The IE team is on a big push to get websites to support IE10 and be touch friendly. But a good touch friendly IE10-compatible website reduces the incentive for a front-end Metro application. Microsoft itself, in tweaking Office for touch and then including it in Windows RT, showed developers how they could avoid creating Metro apps in favor of tweaking their desktop Win32 apps. So you have developers claiming Windows 8 support and pointing customers to either their website or their Win32 apps.
Now we get to where Microsoft’s desire for high margins and its need to incent developers to create Metro apps crippled each other. The main incentive for developers would have been a large number of systems that they could only effectively reach with Metro applications. The best way Microsoft could have made sure this happened would have been to price the Surface RT very aggressively so developers saw a rapidly growing population of Metro-only devices and moved rapidly to jump on the bandwagon. Banks, newspapers, airlines, social networking companies, Information Worker ISVs, etc. would pay a lot more attention to 10 million devices and growing fast then they are to 2 million devices apparently going nowhere. Of course, since Microsoft equated Windows 8 unit growth with Metro app library growth they thought they could have their cake and eat it to. Windows 8 would drive Metro app library growth and they could go for high margins on the Surface RT. It didn’t work.
Microsoft now finds itself with the worst of all worlds. They’ve had to write down the Surface inventory. They will have to be more price competitive (and accept lower margins) on the next generation of devices. And they still don’t have enough Metro apps to make Windows RT devices successful. Moreover, the reputation of Windows RT may have been harmed to an irreparable degree. This is important because, as I’ve pointed out before, Windows RT was supposed to be the future of Windows.