Steve Ballmer, Stop Beating Yourself Up Over Hardware (Premium)

Former Microsoft CEO Steve Ballmer this week expressed regret that he didn’t make the software giant a “world-class hardware company.” Steve, please. You did the right thing.

Our desire to rewrite our personal histories is natural: We all have that one moment we wish we could take back or handle differently, that one retort we wish we had summoned at just the right time.

For many of us, this moment is something truly tragic and personal. For Steve Ballmer, it is apparently his time as Microsoft’s CEO.

But count me among Mr. Ballmer’s fans. He was bombastic and fun, and you will never meet anyone who loved and cared about Microsoft more than he. But some see Ballmer in a more critical light for whatever reason. Granted, he once tried to buy Yahoo for over $44 billion, and, yes, he did orchestrate the disastrous purchase of Nokia.

The bigger issue, perhaps, is that Microsoft’s stock price waged a miraculous battle against change during Ballmer’s tenure, and it sat, unmoving, for basically that entire time period. This is a legit complaint, I guess, but I don’t see what Ballmer could have done to have changed that for the better.

But whatever your feelings about Ballmer, whenever he discusses his history at Microsoft, we will, of course, sit up and pay attention.

Before getting into this, here’s a warning: Ballmer’s comments about his time at Microsoft came during a horrifically painful Recode interview which I strongly recommend you do not watch. The interviewer is awful, and appears to know very little about the topic. I took this bullet for you, folks. Don’t be a hero.

(Most of the interview is not about Microsoft. If you do want to skip ahead, jump to about the 27:00 minute mark. But seriously. Don’t.)

OK, here’s what Steve Ballmer said. And why he shouldn’t beat himself up about the past.

Taking a cue from an “Interviewer 1.0” YouTube video, Ballmer was asked about the one thing he did wrong as CEO of Microsoft.

“I’d say I have more than one thing I regret,” he starts off. “Companies get successful with an idea, with talent around the idea, and with capability as a company to execute on that idea.”

Here, he is clearly talking about Windows, the economic engine for Microsoft’s hyper-growth between 1990 and the antitrust ills of the early 2000s. Windows wasn’t just big as a single entity, it fed innumerable side businesses such as Office and Server, which have major implications for Microsoft’s future (in the form of Office 365 and Azure).

More to the point, Microsoft’s single-minded approach to feeding the Windows economic monster was correct. And you can see this same sort of network effect driving other tech giants like Apple (iPhone) and Google (Search/advertising). You do that one thing right and then you just milk it to death.

“But it turns out, if you want to have a second idea that is different from the first,” Ballmer continues, “you may need new talent, but you also need new capability. And you don’t get new capability overnight. So if you’re lucky enough to invent the first thing and invent the second thing before anybody else does, then you can build up your capability.”

Put simply, Microsoft’s success was all built off of Windows. Either with products and services that literally required Windows, or with products and services that emulated the Windows business model. If a product or service arose within Microsoft that could negatively impact Windows—or, over time, other core products—it was killed.

This made sense at the time. And that’s the thing I think the history rewriters don’t understand. It’s like the Millennial who watches Star Wars for the first time today and doesn’t see what all the fuss is about: The acting is bad, the special effects are terrible, and the storyline is derivative. But Star Wars changed the world in 1977, and was an epiphany for a generation of fans, all of whom walked away from that experience changed. The movies and other entertainment that a Millennial enjoys today would never have happened without Star Wars.

Yes, in this comparison, Windows is like Star Wars. It doesn’t seem that special today, sure. But it changed the world. Back to Steve.

Now he gets specific.

“I think I was too slow to recognize the need for new capability,” he says. “Particularly in hardware. I wish we built the capability to be a world-class hardware company because … one of the new expressions of software is essentially the hardware.”

And to this, I will simply say, bullshit.

Bullshit.

Sorry, Steve. But you are beating yourself up over nothing.

If you were to go back in time 20 years, 15 years, or even 10 years, here’s the history of the personal computing industry you will find: That hardware makers created what we now call PCs, but these PCs were not useful without software. Hardware, unlike software, has a lot of expensive overhead, and becomes obsolete quickly. Software is created once and sold over and over and over again. Hardware was Apple’s (Steve Jobs’) model. Software was Microsoft’s (Bill Gates’) model. And Bill Gates won.

Now, things have obviously changed in the past ten years or so. And today, Apple is the biggest tech company in the world because of its hardware business, the iPhone. But had Apple been successful at anything else during the 1990s or early 2000s, the iPhone would never have happened. You only change tactics that much when what you’re doing (the Mac) isn’t working.

What Microsoft was doing during this time period was working: Windows, again, was the economic engine that drove Microsoft’s success through the mid-2000s. Steve Ballmer was Microsoft’s CEO from 2000 to 2014, so for the first half of that time period, he was simply following the successful strategy that was pioneered by his friend and partner, Bill Gates. During the second half of that time period? Ballmer got Microsoft into hardware.

And not just into hardware. It was Ballmer who invented Microsoft’s current direction, even though Satya Nadella gets all the credit. That “mobile first, cloud first” thing? That started under Ballmer. He called it “devices and services.” It is the exact same thing.

“I think we came to [hardware] later [that we should have],” he continues. “It’s got a different business model. It’s got a different delivery model. Right now, Microsoft is making the transition, in terms of the new capability being the cloud, that one I think we did right.”

That we can all agree on.

When asked if “capability” is really just a word for “mobile,” Ballmer corrects the interviewer.

“Hardware, not mobile,” he says. “The thing we missed with phone is that we tried to use the old techniques, [like] software licensing. The same techniques were never going to get us there.”

That’s only true retroactively. And everyone reading this probably has a pet theory about why Windows phone failed. All of you are, in fact, wrong about that theory. But here’s the kicker: Many of you are also right, because Windows phone didn’t fail from one mistake. Instead, it was death by a thousand cuts.

So let’s think about the reason Ballmer cites, though to be fair, his comment about “techniques” suggests that there are other factors too. If Ballmer is suggesting that Windows phone somehow failed because Microsoft licensed the software, rather than just giving it away, then I will once again have to call bullshit.

Bullshit.

No one gives away a mobile OS. No one.

Apple’s iOS is paid for when users buy an iPhone. It’s not a charity: There is a direct cost to the research and development, and it is paid for.

Google, likewise, does not give away Android. To get Android, hardware makers need to pay Google. Yes, hardware makers can get a free subset of Android called AOSP, but this is specifically designed with a high cost of entry, so that the hardware maker will need to duplicate the Android features (the apps store, Maps, and more) that they’re not getting. And that, too, is not free.

There is no such thing as free. And Steve Ballmer knows this.

“We had the wrong monetization model, we had the wrong delivery model … because we didn’t build new capability.”

When Microsoft announced Windows Phone 7 in 2011, it was perceived as a breath of fresh air. Here was Microsoft, learning from its Windows Mobile mistakes, and addressing the iPhone and Android with a system that wasn’t just different, but better.

That was the theory. But at the Windows Phone 7 launch event that fall, I warned Joe Belfiore that the carriers (the delivery model) would betray Microsoft, because their business model relied on them selling new phones to consumers every few years, not updating existing phones for free. He assured me that this would never happen.

But Windows phone had many issues, in fact, and some of them took years to fully understand. I’ve often discussed how hubs, for example, seemed like a great idea. But no third party services ever signed on because they naturally wanted to promote their own brands, not become a faceless part of an integrated system alongside other services. And Microsoft changed the underlying system architecture several times in just a few years, undercutting developers and customers alike. It kind of goes on and on. Again, death by a thousand cuts.

The rewriting continues.

“In cloud, I’d give you exactly the opposite answer,” he says. “We built the new capability and it served the company well.”

Here, I will argue that transitioning from on-premises servers to the cloud is both obvious and easier than building a truly new capability (a software company building hardware). And that this business makes tons of sense. Not just for Microsoft, but for its customers. In fact, Microsoft is uniquely positioned to transition its customer base to the cloud, and continue a relationship it has had with the enterprise for years to come. It’s like when the music industry changes formats, from cassette tape to CD, or whatever: You get the sell customers the same thing all over again in a new form. It’s the perfect business, really.

“Now, [Microsoft] has done a good job of building out the hardware capability,” he says, referring to work that he started with Surface. Describing a Microsoft Research meeting, he describes being told that, in the future, “there’s only going to be silicon and the cloud. Everything on the client will get built into silicon, and everything else will be in the cloud.”

This is a bit fanciful, even today, given the security vulnerabilities of connected devices and Microsoft’s oft-stated need to keep these devices as up-to-date as possible. There will be silicon, of course. There will also be software. Always. And that was Microsoft’s role for many, many years. It arguably still is today, and if you accept the fact that the cloud is just software running somewhere else, as it is, then you know that it is Microsoft’s role going forward too.

By which I mean: Not only is Ballmer wrong to retroactively declare that he shoulda, coulda, woulda. But he is missing the central point: Microsoft’s future is software. This hardware stuff is just a distraction. And claiming that getting Microsoft into hardware earlier would have been better for the firm is wrongheaded. Not only did history happen as it did, but it better prepared Microsoft for its natural future. A future that I believe will be quite successful.

This is entirely too much hand-wringing from a guy who did a great job. I hate to see him throwing himself at the mercy of an alternate future, and doing so in such a terrible setting. He’s better than this. And I, for once, appreciate what he did for Microsoft and for the industry.