The Steve Ballmer Post

5 minutes estimated reading time

At the end of last week Steve Ballmer announced his imminent retirement and the formation of a committee to find his successor. The narratives that went out with the media painted a picture that was a decade or so of opportunities squandered.

As with most narratives it hides a more complex truth that doesn’t neatly fit in media storytelling.

When did Steve actually take control of Microsoft?

That’s the big question that I don’t feel was answered or really taken into account by the media reports, Ballmer had the financial responsibility when he was appointed CEO in 2000, but Gates set the direction for the company for at least another six years until his retirement in 2006. As one of Microsoft’s largest shareholders; Gate’s still holds sway over the company that he founded.

What did Steve Ballmer achieve?

Steve Ballmer is a hyper-achiever – he went from being the first business manager at Microsoft, to heading up some of the companies most important projects:

  • The Microsoft sales and marketing machine
  • .NET development environment

He even headed up crucial versions of operating systems development. Ballmer played a key role in the success during the 1980s and 1990s.

Gates was the architect but Ballmer was a master-builder.

Gates’ design also laid the foundations of weakness in the Microsoft model. Their winner-takes all approach to partners meant that:

  • PC manufacturers innovated in process to try and claw back hollowed out margins rather than being ready for the kind of disruptive innovation that Apple brought in the hardware space

This is what a consumer PC offering looked like in the late 1990s / early 2000s
Free to Good Home
This is what Apple’s offering looked like
Imacs
I think the pictures tell you everything that you need to say on that front.

If you look at Microsoft’s mobile strategy, there is a succession of screwed over partners including i-Mate and Sendo. Other sectors that Microsoft’s tries to enter look at Microsoft’s history and become very wary. Nokia themselves have admitted that Microsoft’s ownership and bundling of Skype with it’s Windows Phone software was hurting carrier relationships that were key to shift units. The fear and distrust killed competitors at the business plan stage as Silicon Valley kept out of the way of the Redmond juganaut. But it’s also the reason why we have a vibrant open source community and open standards; to try and counterbalance Microsoft’s dominant position.

So Steve Ballmer was playing with less of optimal hand than the media would have you believe. In spite of this Ballmer managed to keep Microsoft growing at an enviable rate of knots.

From Ars Technica:

Under his leadership, Microsoft’s net income has increased to $23 billion, with annual revenue climbing from $25 billion to $70 billion, with an average annual profit growth of over 16 percent.

Now those numbers depend on how long you think Ballmer was actually the shot-caller, but the trend is undeniable.

Secondly, in many key areas like mobile and tablet computing Ballmer was hamstrung by Microsoft having invested too early. Robert X Cringely in his book Accidental Empires likened successful CEOs of a technology companies to surfers. Knowing when to hit a wave and when to transition to the next one. Microsoft had tablet computer products for a decade, smart TV/ set-top box and mobile / PDA software way longer.

Yet according to Ars Technica:

Ballmer is responsible for expanding Microsoft’s reach into a number of new areas, including the heavy push into the “post-PC” era with its focus on portable devices. Ballmer’s Microsoft also created the Xbox and poured tremendous resources into gaining a foothold in the home entertainment market, and it developed an actual viable search competitor to Google.

You could argue that search was a mistake, but as an IBM advert in the latest US edition of Wired magazine says:

80% of the data currently produced is unstructured – coming from sources like images, videos, tweets, posts and e-mails.

That statement alone shows how important search capability would be for Microsoft across their business lines and whilst the Online Services division has been a spectacular under-performer; the company simply cannot afford not to have a dog in the search fight. The work done on search will also pay dividends in enterprise products and emerging areas such as machine learning as a service.

Does Steve still matter?

Ballmer as a retired executive is still a multi-billionaire; he could still make a difference through his investments, probably more so than other retired Microsoft executives have done previously. So it is worthwhile keeping an eye on what he does next.

Why now?

Your guess is as good as mine:

  • Microsoft is making a lot of transitions, it is has re-organised for its next iteration as a devices and services company. That road will be bumpy. If one looks at the likes of Yahoo! one can see the benefit of shareholder goodwill that a new CEO gets. Ballmer doesn’t have that, he has set the course but won’t be able to see this through
  • Microsoft could be taking pre-emptive action, since they have seen the way Apple is in the midst of being Icahn-ed. They already pay a dividend, so changing leadership would be next most likely demand getting ahead of that activism keeps the board in the driving seat
  • A change in CEO would allow Microsoft to put an engineer back in the driving seat. Microsoft at its heart is an engineering culture, which is the reason why its marketing historically has been blunt but effective
  • A new CEO, particularly one from outside the company has much more leeway to make big choices like breaking up or spinning off parts of the company

More related content here.

More information

Microsoft CEO Steve Ballmer to retire within 12 months | Microsoft News Center
Sendo: why it went titsup | The Register
Nokia’s CEO Talks About How Skype Affects Carrier Relations | BusinessInsider
Microsoft, Ballmer, and the end of the PC era | I, Cringely
Microsoft CEO Steve Ballmer to retire within 12 months | Ars Technica