Microsoft's software profits have been in decline throughout
this year, and its bottom line is being propped up only by
impressive returns on its investments. Multiple trends have
converged to obsolete the MS business model, but future historians
are likely to credit Linux with being the catalyst that undid the
MS money machine at the start of the new millennium.
A lot of people in the Linux community don't want to hear
anything about what's happening with Microsoft; but the truth is
that anything remotely resembling "world domination" will only be
achieved, by Linux, "over Microsoft's dead body" -- that's the way
MS is playing the game.
I've been waiting for the past two years to write a story about
Linux putting a serious dent in Microsoft's bottom line. If Linux
is going anywhere at all, eventually it has to show up in
Microsoft's revenues. I have no doubt at all that Linux is going
somewhere, so I've seen it as only a matter of time before MS would
put out that historic quarterly report that showed its profits
shrinking -- or at least growing at a rate far enough
below the 30% per year investors have come to take for granted it
would send shockwaves through the financial markets. This is
inevitable, if MS revenues drop much below the market's
expectations -- that's the way the financial community is playing
The first signs of Linux having any effect on MS
financially appeared in the Times (UK) in February of '99, under
the headline, "Finnish programmer costs Microsoft dear." The
article speculated that an $11 billion drop in the value of
Microsoft stock could be attributed to Linux -- specifically, to
IBM's announcement that it would offer Linux on some of its server
But then, quarter after quarter through '99 and up through the
quarter just ended on 30 September, Microsoft has just kept on
putting out the "same old, same old" report -- the key phrase that
always describes it: "beats analysts expectations." Other than an
opinion in The Register about "How Linux could screw MS in Q4 2000"
(the upcoming Christmas season), you really have to sift through
the fine print in the trade and financial press to find
any negative opinions about Microsoft's financial
prospects, at least "once they get past this antitrust
So how long will it take for Linux to impact Microsoft's bottom
line? Well, a closer examination of the last three quarterly
reports out of Redmond reveals it is already happening.
But Microsoft being the unique organization Microsoft is, it hasn't
made headlines and quite possibly never will.
The January-March, 2000 Quarter
Microsoft's fiscal year is out of synch with the calendar year --
January-March is the first quarter on the calendar, but the third
quarter for MS accounting. To keep things straight, I'll just refer
to MS quarterly reports by the months they cover. For the
January-March, 2000, quarter, Microsoft earnings jumped to $2.39
billion, up from $1.91 billion for the same period in the previous
year. This was an increase of $480 million, or 39% -- in
earnings, or net profits, not just gross revenues. Same
old "blows away the analysts" report, right? Look a little
Most coverage of the MS results just looked at these numbers and
ran with headlines like IT-Director's "Microsoft makes much money."
Tony Smith of The Register took a closer look at the numbers adding
up to that $2.39 billion net and pointed out $885 million "came
from the sale of Microsoft investments." Let's do a little math
here: MS realized $885 million in net income from its investments,
but their bottom line only went up $480 million -- without the
investment income, their net profits would have gone down
by $405 million for the quarter, in comparison with the previous
I took a quick look at the MS report, but never did figure out
where the $405 million in negative numbers was coming from. Since
nobody in the financial press was picking up on this and making a
big deal out of it, I figured there were "charge-offs" for one
thing or another only accountants understand that accounted for the
discrepancy. After all, if the MS net on its "core business," its
software operations, were actually declining that would be really
big news and somebody would pick it up.
The April-June, 2000 Quarter
The general tone of the stories on Microsoft's quarter that ended
in June were decidedly less positive than what has been standard
for most of the past decade. ZDNet's Larry Dignan headlined his
story on the MS quarterly report, "Does Microsoft matter
Dignan didn't mention the numbers at all, but focused on the way
"Microsoft has lost its luster" -- the MS conference call about its
quarterlies happened to coincide with the same for "Intel, Commerce
One, i2, Apple, DoubleClick and RealNetworks to name a few." And
Dignan concluded many analysts opted to skip the MS conference call
in favor of some of the other companies, something that would have
been unthinkable up until very recently. In Dignan's view, future
prospects for MS depended on .NET, and analysts weren't buying the
"You could listen to Microsoft drone on about its .NET strategy,
but it appears to be a hosting strategy a little behind the curve.
In many respects, .NET could be called .Vaporware. Microsoft is
pursuing a strategy that's already applied by a host of others.
Microsoft is retrofitting its tools, applications and operating
systems so it can sell software as services. While Microsoft
retrofits for two years, other companies are there today."
As with the previous quarter's report, The Register was one
source that took a hard look at the numbers -- a very hard
look. Graham Lea noted the trend towards investments propping up
revenue growth, which had surfaced in the previous quarter, was
continuing in a big way (italics added):
"Almost without it being noticed, Microsoft has transformed
itself from a software company to an investment company.
...Microsoft's investments have been performing more profitably
than its products, with the operating income showing 30
cents/share coming from products and 14 cents from investment
income and interest."
Lea's analysis of the numbers continued, indicating just what a
precarious financial situation MS was in (italics added):
"Revenue for the quarter was almost flat... Microsoft's
ability to recognise revenue more or less as it pleases, may have
been helpful here, allowing it to show this very small increase
rather than a potentially headline-grabbing loss. ... Microsoft's
operating income was down 13 per cent for the quarter,
primarily reflecting a 10 per cent decline in the sales of MS
Office and tools..."
The July-September, 2000 Quarter
The quarter just ended in September continued the trend of the
previous two quarters, except that MS stock didn't get the bump
that has been typical after announces its quarterly results.
Apparently some analysts are getting uncomfortable with the idea of
Microsoft turning into a venture capital/banking venture. The
Register story on this quarter, written by John Lettice, was titled
"MS beats Street, ungrateful Street beats up MS" (italics
"Microsoft's Q1 results comfortably 'beat the Street' - but the
Street seems to be getting a mite suspicious. The company yanked
earnings up from $2.19 billion a year ago to $2.21 billion, but the
increase came from a surge in income from company investments...
Some analysts however churlishly declined to be impressed by
how effective a VC/bank Microsoft is becoming, and pointed instead
to a drop in profits from operations of $12 million. ... One
analyst ominously observed Microsoft was "pulling a legacy business
which accounts for 70 per cent of their revenue."
Operating income -- the net MS made off its software operation
-- declined again, by $12 million. And again it was the income off
of MS investments, which increased by an impressive 105% (from $550
million to $1.13 billion), that allowed the Microsoft quarterly
report to be anywhere in the ballpark of what the financial
community has come to expect.
The Motley Fool saw the MS decline in operating income "as the
result of higher operating expenses, mainly in research and
development and sales and marketing." And Fool writer Richard
McCaffery noted that "...margins may very well [continue to] shrink
as the company faces tougher competition from Linux and the
Meanwhile, MS Has Been Boosting Prices in a Big
What makes these quarterly reports more ominous is that the
declines in MS software income have been occurring at the same time
MS has been systematically increasing its prices. As far back as
November of '98, a story appeared on the AP wire with the headline,
"Microsoft Software Costs Inching Up," reflecting MS tinkering with
licensing to boost prices.
In October of '99, the Gartner Group warned that changes in
licensing that MS had quietly announced in conjunction with the
release of Windows 2000 would impose on enterprise customers a
"50 per cent hike in their Microsoft budgets each year until
2002" (italics added):
"Although Microsoft pricing will remain ostensibly flat,
enterprises using Microsoft software should expect to pay 50 per
cent more a year as a result of changes to terms and
In February of 2000, on the same day Microsoft officially
released Win2k, Gartner released another report on MS pricing.
ZDNet noted that "the 'significant' new charges for Windows would
affect 90 percent of large firms over the next two years, no matter
what other software products they used."
The full report on the Gartner website gave the "gory details"
on exactly what was going to cost more, and included some
straightforward price increases: for example, client access
licenses, required to allow Win2k desktop clients to connect to
Win2k servers, would cost 82-122% more than with NT 4, and the
upgrade of a Win9x desktop to Win2k would cost 84% more than to
WinME (which MS effectively wouldn't allow anyway). But one item
that Gartner flagged showed just how "innovative" MS could be when
it came to figuring out new ways to extract money from its
"...if an OEM ships new desktops pre-installed with Win2000
Professional and an enterprise overlays this with its own system
image based on NT v.4 Workstation or Windows 9x, a UA ['Upgrade
Advantage'] fee of $157 will be due for each machine overlaid."
In June of 2000, MS announced across-the-board price hikes on
its server apps. ZDNet's Mary Jo Foley quipped, "The company's new
motto, at least for its 2000-generation back-office server
products: You get more, you pay more." MS justified the increases
with claims of increased functionality:
"Our prices are higher across the board," acknowledged Stan
Sorensen, group product manager for server applications, while
speaking about changes to the price of Exchange. "But we're adding
tons of new functionality."
In August, Gartner revealed the full extent of the "innovative"
price hiking that had only been previewed in February: a
"reimaging" charge. Reimaging is what MS enterprise customers have
been doing for years: they take the PCs they buy from a hardware
vendor and overwrite the vanilla Windows software that the vendor
ships with a customized "image" of the operating system that has
been configured for the organizations specific needs in accordance
with the IT department's standards.
They've already paid for Windows and they're just overwriting it
to save the time it would take to configure it the way they want on
each PC. They've been doing this for years, and now MS suddenly
says they have to pay an extra charge of $117 to $157 per
PC, effectively just about doubling the cost of the
In September a Register story reported that the reaction from
Microsoft's biggest customers was so negative -- The Reg termed it
a "revolt" -- that MS had amended the reimaging policy to exempt
the big guys ("Select" and "Enterprise" customers, in MS licensing
jargon). But Gartner calculates that the cost of the amended policy
to the small and medium size businesses still affected "will net
Redmond an extra $11 billion in revenue" (italics
Is Linux to Blame?
If Microsoft's bottom line software income is indeed in trouble
(relatively speaking -- no longer growing at historic rates), what
evidence is there that Linux is the cause? Well, there's virtually
no hard evidence at all to make this connection, although most
Linux supporters probably have a sense that it's true.
A Register story by John Lettice, posted right after the MS
report for the April-June, 2000, quarter was titled "Linux server
growth threatens MS revenue supply." But Lettice was referring to
an IDC report that had just appeared, forecasting that "Linux
server shipments will climb faster than the rest of the market,
resulting in 2004 in Linux shipments of around 4.7 million, finally
within spitting distance of Microsoft, at 5.7 million." So any
impact on MS revenues was only a future possibility.
The problem with open source software is that it's very
difficult to get hard numbers on the actual number of installs that
are out there. With proprietary software, each copy licensed is
recorded and feeds into the vendor company's financial statements
-- counting the copies is fairly easy, as is reckoning the
software's "market share" in monetary terms. It's fairly easy to
see that MS isn't making as much money as it used to on its
software, but we can only estimate the total number of Linux
servers and desktops that are out there that represent a lost sale
of Microsoft Windows.
What's even harder to do -- virtually impossible -- is to
determine how many corporate IT departments are putting some part
of the Win2k/Office2k upgrades they would have otherwise done by
now on hold, while they evaluate Linux & open source office
suites for possible use in select areas -- especially Linux &
NDS vs. Win2k & Active Directory on network servers.
It's hard to counter the argument that there are
multiple trends underway that are responsible for the
decline in MS software profits, and that the Linux impact so far
has been minimal. And I have to agree that, whatever impact Linux
is having on the MS bottom line, it is having it in conjunction
with these key trends (and probably a few others) that are
independent of the Free Software Foundation and the open source
The fastest growing market now is low-cost devices, not PCs --
on a $200 product, there's no margin to pay MS $50-100 for an
operating system (even if it works!).
Businesses are balking at the unending upgrade cycle MS depends
on -- the real business gains from upgrading all your PCs just so
you can upgrade to Win2k are pretty marginal, compared with the
benefits of upgrading from Windows 3.x to Win95.
The focus of software development is shifting from the desktop
to the Internet -- this environment is orders of magnitude more
complex, and the cost of proprietary software development here is
much higher, than what MS has been geared up for historically,
and its desktop monopoly no longer gives it that much of
an advantage against competitors like Sun and IBM.
The antitrust trial broke Microsoft's iron grip on its OEMs --
it doesn't matter what the outcome of the appeal is, MS has lost
the tight control it used to have over the vendors who rely on
Windows to sell their products.
But when analysts look back on the year 2000 from the vantage
point of five or ten years in the future, I suspect that they'll
credit Linux with being the catalyst that dramatically
accelerated the impact of all the above trends on the MS bottom
line. And I can imagine all sorts of sage opinions about how
brilliantly Microsoft executed its transition from selling software
to providing venture capital and banking services!