Should Microsoft Spin Out Windows & Office?

On Monday December 27th, I received a message on my Twitter account from Jordan Novet, a reporter at CNBC. He was writing a story on CellarTracker, had figured out that I was a long-time CellarTracker customer, and he wanted to get a few quotes from me.

We ended up having a very long phone conversation, as he had a lot of questions about Microsoft and the Internet, and so I tried to share a few stories and provide a few insights. One off-hand observation I made is that Windows and Office are holding Microsoft back, and he was surprised by my comment. I said surely others have made that observation, but it was a novel idea to Jordan.

I got another message on Twitter from Jordan yesterday (January 10th), saying his editor had encouraged him to write up an article about my comments on Azure. We traded several more messages, then I moved on to other things.

So I was not completely surprised to see the CNBC story this morning: Microsoft should sell Office and Windows to boost cloud business, former executive says. But I do wonder (as several members of the Microsoft Old-Timers group on Facebook wondered): are my opinions about the structure of Microsoft today newsworthy? 😉

Here are a few words to provide a bit more context and rational for my thinking that Windows and Office are (40+ year) old technologies holding Microsoft back. Just as the mainframe held back IBM from the PC, chemical film held Kodak back from digital photography, and the success of the pre-Internet Blackberry killed RIM. [Not to mention the failures of AOL, Blockbuster, CDC, Compaq, DEC, Netscape, Palm, Xerox, …]

As you may know, I worked at Microsoft 1985-1999, built the first three versions of Internet Explorer, worked on Win95, MS-DOS 6, and OS/2, and wrote a few memos that have proven to be quite accurate predictions of the future:

1995: Web kills Windows Apps (

1998: “The Windows Service; predates Slack and Teams by 20+ years (

Here are a few examples of where Windows/Office has already held Microsoft back:

1) The Windows group had too much power within MSFT, and so throttled the “mobile OS” group, which was unable to compete with Apple/Google. Android should have been a Microsoft offering, eh? IBM’s mainframe group pulled a similar gambit on their PC business.

2) Internet Explorer hit a peak of 95% browser share in 2004 (before mobile, obviously). But SteveB felt that IE existed only to support Windows, so Microsoft dropped development of Macintosh and *nix versions and didn’t pursue mobile.

Here is a longer post about how Microsoft “fumbled the future” (see Xerox) on web browsing:

[Should I mention that SteveB turned down the opportunity to buy Google for $400M in 1998?]

P.S. This is not the first time I have suggested that Microsoft should split itself apart.

On August 23, 2013, in a front page article in the New York Times by Nick Wingfield — Ballmer Exit Brings Microsoft a Chance for Reinvention — he quoted my Facebook post from that morning (when Ballmer suddenly announced he was stepping down as CEO of Microsoft):

Others believe Microsoft is not governable in its current form. Ben Slivka, a 14-year employee of Microsoft who left in 1999, said the company should split up into five independent companies he calls “Baby Bills” devoted to Windows client software, Office applications, servers, Xbox and the Web.

“Give each of them (say) $5B for a rainy day, but not much more,” Mr. Slivka wrote in a post on Facebook after the news of Mr. Ballmer’s retirement. “You want them to be hungry. Return most of the cash hoard to shareholders.”

Author: benslivka

19 start-ups, software, hardware, biotech, space launch, neurodiversity, learning, free markets, food, wine, cycling, walking, Seattle, Microsoft, Northwestern University, Garfield HS, DreamBox Learning, IBM, Amazon.

Leave a Reply

Please log in using one of these methods to post your comment: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s