July 30, 2005

Divided Loyalty

Microsoft is keeping itself in the news with the Kai-Fu lawsuit. As a former Microsoftee, it makes me think back to a bigger legal event when I was there: the Jackson ruling.

Normally, when you work for a company, you are rooting for its victories. But sometimes you find that you are cheering for the opposing side.

For example, what happens when a company improves profits by squeezing benefits? What happens when a company retains talent by suing its own former employees? It is not so clear who you should be rooting for, The Company or The Employee. It doesn't matter if it is Wal-Mart or Microsoft. When a company works against the interests of its own employees, it can divide loyalties.

The most spectacular, and most revealing, case where "Microsoft company" interests ran counter to "employee" and "shareholder" and "customer" interests, I think, was the Microsoft breakup case before Judge Penfield Jackson back in 2000. When Jackson ordered the breakup, you might think that legal defeat would have sent shudders of disappointment through the rank-and-file.

However, employees did not get demoralized. In fact, although it was a subtle shift, in some corners of company you could almost smell the atmosphere change for the better. It was the smell of hope....

Rooting For a Breakup

It was not a universal feeling among employees. However, I was one of the people inside the company who thought that Jackson was right.

"Abusive monopoly" or not, I believed that it would benefit everybody involved with Microsoft - shareholders, employees, customers, and the industry - if it were broken into two or three separate companies. My reasoning was simple, and I was not alone. The thinking was completely self-interested:

1. The innovative parts of Microsoft were (and still are) being held back by its fiduciary duty to milk existing Windows and Office revenue; similarly the "business milking" parts of the company are held back by the drive to expand the company's scope. Even though as a programmer you want to figure out 'how do I make the best product?', the question 'how do we protect the Windows franchise' never goes away. Often the answers to the two questions are exactly the opposite.

2. From its position of comfort as a monopoly, Microsoft has become too fat, lazy, insular, and blind. It is considered a mark of shame, for example, to run any product development on Linux, or to use open-source tools. In a time where much of the excitement of the world is open-source, that bias is like telling a world-class journalist to work under oppressive censorship. No perl or python or Apache for you!

3. In a breakup, us programmers would have individually been more free of 'bureaucratic red tape' and would get to work on something more exciting than Just Yet Another Piece of Armor to Defend Windows. Smaller companies are easier to understand, and you can have a bigger impact.

It may be surprising, but really, it should be no surprise that I was not alone. Many engineers and several top executives privately believed that a breakup would be better for Microsoft.

I still believe that. A broken-up Microsoft would be a far more formidable force in the industry than today's glued-together monopoly. And it would be far more valuable on the stock market, and it would be a better place to work. As a shareholder, I would love Microsoft to be broken up. And of course, customers would love it; customers always hate a monopoly.

If employees like me were rooting for a breakup, and if it would be good for shareholders, then why did the company fight it so vigorously? What is a company if not its employees, shareholders, and customers anyway?

Who is Microsoft?

To understand Microsoft, there are two financial things you need to understand. Please do not be scared by the following. You do not need to be a financial nerd to understand or appreciate it.

The first thing to understand is that Microsoft is Bill Gates (and Steve Ballmer). To make my point more clear (since Microsoft has diluted its shares since then), let's rewind a decade to 1995 when I joined Microsoft. Here is a quote from the 14(a) filing that year.

                         Amount of beneficial ownership of common shares
                                           as of 9/8/95        Percent of class
                                       --------------------    ----------------
William H. Gates......................      141,159,990              23.9%
Paul G. Allen.........................       55,893,020               9.5
Steven A. Ballmer.....................       29,952,764               5.1
(others are all under 1%)
Executive Officers and Directors
  as a group (22 persons).............      234,167,557              39.4

In 1995, Microsoft was already an international behemoth, and 24% onwership in the hands of a single person was really quite remarkable. The 39% insider ownership meant that Microsoft was a very tight-knit little group in 1995 - it was the Bill, Paul, and Steve show.

From the most recent 2004 14(a) you will find that Bill Gates' stake has been sold and diluted down to 10.1%, and Ballmer is only down to 3.8%. While Paul Allen has been divesting (3/06 update: an article on the topic), Ballmer has not been selling. Steve Ballmer is clearly the #2 person.

Still, more than anything else, Microsoft is Bill Gates. Financially, culturally, historically, the company is and was founded, owned, controlled, and operated by one man.

We should not be surprised then the interests of The Company sometimes run counter to the interests of everybody else involved: employees, minor shareholders, whoever.

Who is Microsoft? The answer is: Bill... and a bit of Ballmer.

Of course, the company's power does not come from Bill. It comes from the business. The business is a moneymaking machine. "Better than a printing press for money", some of us employees used to say. Printing, printing, printing, ka-ching, ka-ching, ka-ching....

What is Microsoft?

So where is the money printed? Is it printed in building 17? Or in building 4? Here is another set of numbers. Again, don't get scared. They are interesting.

The following is taken from Microsoft's 2004 annual report. This page of the annual report, in my opinion, is the most interesting page in the whole filing, because it breaks out revenue and income for the individual business units. Take a read of the original.

Microsoft in 2004 (year ended June 30)
                                   Revenue (money in)  Profit (money made)
Client (Windows XP)                      11.2b              9.0b
Server and Tools (Server+SQL)             7.9b              2.2b
Information Worker (Office)              10.9b              8.1b
Microsoft Business Solution               0.7b             (0.2b)
MSN (homepage, Hotmail, search)           2.4b              0.3b
Mobile and Embedded Devices               0.2b             (0.1b)
Home and Entertainment (Xbox)             2.9b             (0.8b)
Reconciling amounts                       0.6b             (9.4b)
Totals                                   36.8b              9.0b

The first two lines represent Microsoft Windows, with SQL server and development tools (probably money-losing) mixed into the second line.

Notice that Windows and Office got 31 billion of the 37 billion of revenue for the company - about 84% of revenues. And they made more than 21 billion out of 9 billion in profits - that is, they would have been 235% as profitable on their own than as part of the company as a whole.

The 9.4b of losses that come from "reconciling amounts" is comes from various profit-sapping moves that Microsoft has made that make the company seem smaller than it is, e.g., big legal settlements in a favorable legal environment that increase shareholder value over the long run without letting the profits pump up the stock price in the short run.

Taking that stuff out of the calculation, Windows and Office on their own still make 115% as much on their own as the company as a whole.

Microsoft Is Mostly A Loss

For all the the employees working on MSN Search, Mobile Devices, XBox, and so on, Microsoft really only does two things from a financial point of view, just as it has only really done two things for decades:

1. It does Windows.
2. It does Office.

(Actually, Microsoft is a triangle, with an important money-losing third leg hidden in the second line item of the income breakout. I have written about this structure in the past, and that article is still relevant today.)

Everything else in the company is a huge multibillion dollar loss, and has been for many years.

From a financial point of view, the company would be far healthier if it were to set Windows and Office free from the rest.

And it can also be argued that both Windows and Office could make a ton more money if you separated them from each other too. Whether it would be wise to separate the giant twins can be debated another day.

Newbie Meets Reality

So the situation must be great for new businesses like Mobile, MSN, and Xbox, right? They are getting a huge free ride with billions of dollars of unearned monopoly profits being funneled into their businesses. Who would ever want to be split off from that river of free money?

The real hidden value in Microsoft is not obvious in the numbers above. The real surprise is that all the money-losing businesses would be much healthier split off of Microsoft too. The thing you do not see on the income statement is that at the "innovative" divisions suffer because they pay a huge Strategy Tax.

When I was a newcomer at Microsoft, it was already quite a big company. So as a newbie, they had lots of artificial events to make it "feel" like a smaller company. A few months after I joined (after HR figured out they had lost my initial invitation to an earlier such event), I got to attend a "meet the executives" welcome dinner. The executive at my table was Craig Mundie, who at the time was in charge of Microsoft's mobile offerings - Windows CE.

"Here is my big chance to fix Microsoft's technology for handhelds!", I thought as an idealistic newbie. So after smalltalk about old-time programming and other pleasant topics, I piped up and talked Real Company Strategy. "Why is Windows CE so much like Windows?" I expressed my opinion that Palm, the major competitor, didn't suffer under that constraint.

Strategy is an Energy Drain

I explained to Craig Mundie as a know-it-all engineer: with a simpler architecture, simpler UI, and a simpler design, Palm OS could stretch device battery life to a month or more. Windows CE measured battery life in hours. A whole day of battery capacity for Microsoft was considered a victory, which was embarassing. My proposal: "Why not walk away from the Win32 architecture and start with battery life as goal #1?"

At that point, he stopped smiling.

"We must protect the Windows Franchise," he responded. No matter what technical objection I raised to the overall approach of Windows CE, he answered with almost exactly these words. He had absolutely no interest in engaging in any discussion about how to improve the technology outside a Win32 model. In fact, he struck me as a bit of a dummy because of his intransigence.

As an engineer, I left the dinner shaking my head, deflated and ashamed of my company. I did not understand why the company would build an inferior product on purpose. But a decade later, I am older and wiser, and I know how to read an income statement breakout. I understand him now.

His job was not to build the best mobile operating system.
His job was to make Windows into the best mobile OS.

Craig Mundie had a specific charter to expand the multibillion dollar Windows franchise. Windows was the beginning, end, and middle of the strategy. No 24-year-old engineer straight from graduate school could ever change the Company Strategy. Trying an approach totally different from Windows would risk the money flow that supported the whole Windows CE effort.

Sure, you could build a new and different OS, but what if it ended up replacing Windows with a Different Business Model and as a result only made 50% as much money? The only sensible approach was one that expanded Windows.

Taxation Without Representation

I never did work inside the Windows CE effort. But most of my time at Microsoft was spent working on other money-losing "strategic bets" such as Internet Explorer and .NET. I was in the part of the company that spent the money that Windows made, in an effort to expand the scope of the company.

Paying the Strategy Tax is the worst part about working for Microsoft. It is what sets any little Microsoft team apart from a free-as-a-bird startup. The Strategy Tax means spending months figuring out how to embed Microsoft Office documents inside Internet Explorer, or figuring out how to stitch Internet Explorer into the Windows Desktop, even when no customers are asking for it. It is why DLL Hell becomes Hell for Everybody, rather than making a World Without DLLs.

The Strategy Tax is the reason some of us employees were quietly hoping, secretly wishing, for a big Penfield Jackson breakup.

Many of us knew, intiuitively that our businesses would do better free of Microsoft's Strategy Tax. And there was also evidence. Certainly people who left did quite well on their own. And the rare example of an actual spinoff, Expedia, surpassed all expecations.

But the outcome of the Expedia spinoff was said to have made Bill Gates furious. Ungrateful employees, working in their own interest instead of in Microsoft's.

Googling for [Microsoft "strategy tax"] gives you a nice selection of commentary from a cross-section of different parts of the company - Office, Xbox, Internet, .NET - that suffer under the strategy tax. A very nice, succinct summary of the general feeling comes from Erik Engstrom when talking about Xbox development:

"The strategy tax was very real.
You never knew when the tax collector was going to come.
You could be halfway done and then get hit with the tax bill."

That is why even the money-losing parts of the company wished, in their secret heart of hearts, to be free of their sugar daddy. Sure, we would lose our river of Free Microsoft Money from Bill. But better to take the risk. We wanted to Live free or Die, as they say on all the license plates in New Hampshire.

But alas, the Jackson breakup of Microsoft Corporation was a short-lived ray of hope. A few months later, his ruling was overturned, and a Bill-Gates-friendlier G. W. Bush administration in 2002 squashed the Justice department lawsuit and negotiated a softball settlement.

Thanks to Adrian Raeside for permission to post the Microsoft Octopus cartoon.

Posted by David at July 30, 2005 08:25 AM

2 things? I'd say 3 things:

1. Windows
2. Office
3. Flight Simulator

A coworker w/ more MSFT experience than I have (I work for a Seattle-based vendor) once remarked that Flight Simulator was essentially a revenue padding device that helped underwrite those strategic money losers.

Actually, Joel Spolsky posted something about making MSFT a better company by breaking it up and relocating some of the edgier, more experimental shops/groups to places like Pioneer Square... I think he has something there - MSFT has lost a lot of its soul.

(Oh, and I hope my 5-yr-old shows as much interest in 'fun' programming a year from now as your 6-yr-old ;-) )

Posted by: protected static at July 30, 2005 04:35 PM

You have a mistake, near the end, you mean the bush administration of 2002, not 1992 (end of another bush era :) )....

Posted by: Tom Z at July 30, 2005 10:04 PM

Thanks Tom. Fixed that!

Posted by: David at July 31, 2005 05:09 AM
Post a comment

Remember personal info?