It’s Time to Kill Surface

“The question that needs to be asked and answered is why hardware.”

To Satya Nadella’s credit, he provided not just the answer, but the question as well. And, looked at narrowly, there were good things seen – and not seen – at Microsoft’s Surface event. Having clearly failed as a mass market device, it makes sense to focus Surface and more clearly define its use case. And, if that use case is productivity, then it also makes sense to kill Surface mini. That Nadella allegedly did just that is a great sign. Now he just needs to kill the whole line.

I actually think a useful way to understand the Surface problem is to think about the Xbox.

It is by focusing on the console world that one arrives at the conclusion that the Xbox, all things considered, is a success and something Microsoft can hang its hat on. In fact, that’s exactly what nearly all Microsoft employees, from executives to rank-and-file, do whenever questioned if Microsoft is innovative. “Look at the Xbox! Look at Kinect!” is their refrain.

The problem is that winning at consoles is a small goal, one wholly different from the reason Xbox was created in the first place. For many years now Microsoft has been focused on “Three Screens and a Cloud,” the idea that they as a platform provider ought to have a presence on your desk, in your pocket, and in your living room, all tied together by the cloud. While that specific formulation arrived somewhere around 2009, that vein of thinking was central to Xbox’s creation; the console aspects were meant to be a trojan horse, giving people a reason-to-buy the Xbox (and not a Playstation); the more computer-type aspects would then be added over time until an Xbox was to living rooms what PCs were to every desk in every office and every home (running Microsoft software).

It was this original goal that contributed to the current Xbox One disaster; Microsoft’s newest console is not only underpowered relative to the PS4, it’s also $100 more expensive (due to the mandatory Kinect), and launched with a terrible wave of publicity surrounding its always-on nature. The Kinect and connectedness were both included to help the Xbox One fulfill its goal of being the primary box in your entertainment system, controlling not just games but also live TV with your voice. Unfortunately, it doesn’t work that well for entertainment, even as it has hurt Microsoft’s ability to compete for console buyers. Thus, Microsoft has spent the last year walking back many of the main features of the console, including its DRM system, its connectedness, and, last week, offering the console without the supposedly essential Kinect to make it $100 cheaper. Now it’s the same price as a PS4, but still less powerful and with the same dark cloud.

Here’s the bigger problem though: even if the Xbox One worked perfectly as an entertainment center, it would still cost $499.1 For those not good at math, that’s $400 more than an AppleTV, and completely unapproachable for anyone who does not care about gaming. In short, it is not enough to consider how the Xbox is doing relative to consoles; the Xbox must be evaluated based on how it is aligning with and contributing to Microsoft’s overall strategy, and in that light, it is an unmitigated disaster.2

So what about Surface?

What the Xbox example illustrates is it is not enough to consider whether or not Surface in isolation is a successful (i.e. profitable) product (although, like the Xbox for much of its existence, it’s not). Rather, we need to consider the overall goals for Surface. As best I can tell there are three:3

  1. Surface was the physical manifestation of Windows 8. Back when I was a category manager for the Windows 8 app store, trying to explain Windows 8 to developers,4 Surface was incredibly useful for explaining Microsoft’s vision of moving seamlessly between work and play with one device – and why you might want two operating systems on one device.5 Not that Surface was made for my personal benefit, of course; rather I believe it was intended to help sell Windows 8 to all of Microsoft’s stakeholders, including OEMs, developers, enterprises, and end customers.

  2. Microsoft did not believe their OEM partners were capable of competing with Apple. Certainly there are no public statements to this effect, but I can tell you that internally most of Microsoft’s OEM partners were viewed with disdain, being resolutely focused on the bottom line and unable or unwilling to make something of Apple-like quality. Correctly understanding that things like fit-and-finish were that much more important in a personal device like a tablet, Microsoft decided that if they wanted hardware done right, they had to do it themselves.

  3. By making their own tablet, Microsoft could reap absolute margins similar to what they enjoyed while providing software for PCs (although the margin percentage would be lower). Microsoft traditionally captured about $115 per PC (through Windows and Office licenses), but in the tablet category, $115 renders OEMs uncompetitive, especially compared to using Android. Instead, Microsoft could sell Surface for $499, with a 25% margin, which comes out to $125, about the same amount Microsoft is accustomed to earning on a device. Were this strategy successful, it would allow Microsoft to maintain its bottom line (and significantly increase its top line) even as PC sales declined in the face of tablets.

These goals must have been important, because Surface came at a significant cost. There was the actual cost of hiring the right employees, the tooling, the factories, but more importantly, there was the cost of competing with Microsoft’s most important partners – the OEMs (and Intel, when it came to the ARM-powered Windows RT). Sure, said partners perhaps hadn’t invested as much in R&D as they could have, but part of it was because Microsoft (and Intel) had long since bled them dry; surely Surface would further demotivate them.

So then, how has Surface fared?

  1. Windows 8 is a failure, rendering reason 1 moot. In fact, from what I understand the Windows group is laser-focused on simply making Windows acceptable to their enterprise base, which has been very vocal in their displeasure.

  2. Every version of Surface has been a high quality device, on the same general level as Apple. So I guess reason 2 is a win. The problem, though, is that Surface’s sales numbers show that device quality is not the primary sales driver for Microsoft customers. In other words, Surface is the tablet equivalent of the HTC One: it is high end hardware in a market where Apple has already taken the high end. Both Surface and One are thus stuck in the middle, appealing to no one.

  3. Contributing in a meaningful way to the bottom line entails selling at much greater numbers than Surface has to date. Remember, volume was always the goal; that’s why Microsoft made so many Surfaces that they had to eventually take that $900 million writedown on unsold inventory. Clearly this goal has failed as well.

Meanwhile, Sony is leaving the OEM business, Dell is restructuring, HP can’t decide whether to sell or not; Acer is barely afloat; only Lenovo seems to be prospering (and now Surface Pro is aimed directly at their Thinkpad lineup). When you consider the original goals, none of which have been met, and the original dangers, all of which have come to pass, the only conclusion is that Surface is a failure.

So then, the question must be asked, as Nadella did, “Why hardware?”

We are not building hardware for hardware’s sake. We want to build experiences that bring together all the capabilities of our company from our cloud infrastructure to our application services to our hardware capability to build these mobile-first productivity experiences. That’s the mission.

This here is the greatest danger of forgetting your original goal; you start making up new ones, that are basically “because we need it to exist.” The hardware capability that Nadella claims Surface leverages only exists because of the decision to make Surface. Nadella is basically saying Microsoft needs to make Surface because Microsoft makes Surface. With that sort of reasoning, you can continue on a wrong path forever, just like the Xbox.

To be fair, from what I understand it was Nadella himself who killed Surface Mini, likely spelling the end of the road for Windows RT.6 Assuming this is true, Nadella should be applauded for making a tough decision based on the world as it is, not the world that Microsoft wishes it were.

Moreover, Microsoft is doing just that when it comes to the cloud and application side of their business. It actually rather pains me to write something so negative, given the dramatic transformation Microsoft has undergone over the last few months.7 However, when it comes to PCs, Microsoft needs to focus on fixing Windows 8, and leave the devices up to its partners, especially Lenovo. Lenovo knows how to compete in mature markets,8 makes great hardware, and Microsoft should see them as their best partner, not a competitor (which, with the business-focused Surface, they necessarily are).

It’s time to kill Surface.

  1. Believe it or not, until a couple of weeks ago you also had to pay for an Xbox Gold subscription to even use Netflix – which, of course, you also had to pay for 

  2. There is no group more oblivious to this reality than the Xbox team. They have long seen themselves as too cool for the rest of the company, limiting access to their buildings and generally treating other divisions like crap 

  3. While I previously worked for Windows, I had no insight into Surface 

  4. A red flag to be sure 

  5. I know that’s not technically correct, but it’s how it is perceived, and fairly so 

  6. It’s actually kind of a bummer; Windows RT was a very clever operating system once you learned it (which, admittedly, was part of the problem), but it never had the app support to make it viable. I suspect that future small tablets will be based on Windows Phone 

  7. While Nadella is getting most of the credit, obviously much of this work had to have begun under Ballmer, who deserves credit for his graciousness 

  8. More on Lenovo tomorrow