Subject: Microsoft and related subjects From: "Curt A. Monash" <0006058685@mcimail.com> Date: Thu, 8 Jun 95 12:01 EST

Message-Id: <84950608170148/0006058685NA3EM@MCIMAIL.COM>
Sender: owner-online-news@marketplace.com
Precedence: bulk
Status: RO
X-Status: 

I'm going to take a bandwidth risk here.   Namely:  I've gotten a lot of
e-mail requests for my Microsoft article.  I'm going to act as if it is of
general interest, and post it here.  The presumption is that software
monopolies are enough like big economy-of-scale information providers to
offer useful insights.  Please forgive any netiquette violations.

There are three parts to this article.  The first covers software buying
behavior.  The second reviews IBM in its monopoly days.  The third is about
Microsoft.  Many of you will only care about Part 3.  This article, BTW, was
written just before Judge Sporkin disallowed the Microsoft/Justice Dept.
settlement of the anti-trust action.

This article is excerpted from the Monash Software Letter, a monthly
(approximately ) analysis of issues in the enterprise software industry. 
Some portions assume more software industry knowledge than will be possessed
by many readers of this list; please accept my apologies for that and skip
to the more interesting portions.


1.  FEAR, GREED, AND THE USER EXPERIENCE

Virtually all business decisions are motivated by some combination of greed
and fear.  In the case of computers, greed drives the desire for
productivity-enhancing, feature-rich new technology - at the best possible
prices.  Fear creates a hunger for product reliability and customer support.


The application and ADT (application development tool) markets are typically
driven by greed.  Users want the latest, greatest, and coolest technologies,
to help them solve business problems and enhance their resumes.  Certainly
buyers are concerned and fearful about a broad range of pitfalls, but they
focus first and foremost on "the user experience".  Whether it's an actual
feature list, a look-and-feel, or just general simplicity and ease of use,
buyers want the "right" product, and what's right for one is wrong for
another.  Indeed, it's literally impossible to create one product that's
right for everybody.  For example, suppose you could match everybody's
feature wish list.  Then you would wind up with a product that's cumbersome
and top-heavy, both in functionality and in its "footprint" and other
performance attributes.  For example, consider Oracle CDE vs. PowerBuilder,
or a "real" accounting package vs. Intuit's Quickbooks. 

The market for operating systems and other similar operating software
(transaction processing monitors and other middleware, some aspects of DBMS)
is very different, driven primarily by fear.  There's fear of failure (i.e.,
unreliability) and fear of incompatibility (your hardware, peripherals,
operating software, and applications might not work together, at least not
without a lot of hairpulling, risk, and expense).  As one operating system
becomes dominant (IBM 370, Microsoft Windows), there's also career risk
avoidance.  If you don't know the dominant platform, it's hard to get hired;
furthermore, if you use the industry-standard machines, you may not get
fired even if they do crash.  Since most people need the same things from
their operating systems anyway, it is perfectly natural for one system to
become a broad industry standard. 

Just as they have different product requirements and market share dynamics,
the two types of software have different support requirements as well.  The
operating software market requires, simply, absolute support.  Above all
else, a platform must work, and it is the vendor's job to see that it does. 
IBM's fabled support is the most obvious case; HP's Idaho printer hotline is
a more modern example.  By contrast, it is impossible to give absolute
support for a spreadsheet or accounting package unless you're willing to
give a free telephone MBA in accounting.  Users understand this, and opt for
training and consulting where hotline and on-site support break down.  One
vendor which is acutely aware of - and vocal about - this distinction is
Oracle.  It is beefing up on-site support to enhance DBMS reliability, while
simultaneously building a booming consulting business focused on application
implementation and development. 

2.  IBM,  LONG AGO

For 20 years, IBM dominated the computing platform business with the 360
operating system and its descendants (leading to what are now MVS and VSE). 
Eventually, however, minicomputers and PCs developed a massive
price/performance advantage vs. mainframes, which made those platforms able
to support a totally different type and quality of application than IBM
mainframes could (at least for anything resembling an affordable price). 
IBM's hegemony was then finally broken.  

For most of this period, IBM was unstoppable in some branches of software
(operating systems themselves, TP monitors) and hopelessly inept in others
(applications, ADTs).  Its products in the latter category were classic
examples of cumbersome bloat; examples range from PROFS and DisplayWrite, to
mercifully forgotten ADTs such as ADF and CSP, to IMS's programming
interface.  IBM's share in DBMS - which of course combine elements of both
operating and application development software - oscillated wildly, from the
early 1980s when IDMS, et al., clobbered IMS, to the mid-80s when DB2
blasted the independents out of the water.  

Of course, without Justice Department intervention, there might not have
been an independent software industry to begin with.  In 1969, IBM was
forced to "unbundle" software package pricing, giving the independents a
chance to compete on a somewhat level price playing field.  Even with
unbundling, IBM got large market share with a number of inferior products,
either because its customers didn't know any better, or because they made
the "safe choice".  With bundling - which would have made these products
seem "free" - IBM's market share might have approached 100%.  

In the early and mid 1980s, IBM considered getting serious about software
sales.  Not much happened, however, except for DB2.  In retrospect, there
were several reasons:

        ---Overall, IBM gained more in hardware sales from ISV support than
it lost in software opportunity.  

        ---IBM recognized its software development problems.  

        ---IBM had enough market share in enough segments that it didn't
want to undercut independents' pricing.  

        ---The leader of the software effort, Bob Berland, died tragically
from brain cancer.  

Ultimately, IBM decided to go with software industry "partners" rather than
push its own software (maybe it was following the great example - ahem - of
the Microsoft/DOS deal).  The upshot, of course, was a series of highly
visible failures, from the AICorp/Intellect/natural language fiasco
(multimillion dollar investment, including 2-page ads in Time; 1900 trials;
7 sales) to the AD/Cycle debacle.  

3.  GOOD MICROSOFT, BAD MICROSOFT

Now I'll finally return to the questions of Microsoft, its market impact,
and what the government should do about it.  There are actually two
Microsofts:  Good Microsoft, the application vendor, and Bad Microsoft, the
system software company.  Good Microsoft has helped propel a tremendous
explosion in PC capacity and unit shipments.  Bad Microsoft, by reducing
product differentiation, has also helped foster PC hardware vendor
competition.  However, the abysmal quality of the core product has created
enormous hassles and excessive usage costs, hampered PC industry growth, and
postponed the benefits of the computer revolution.  Good Microsoft has
competed ethically, except where it has taken unfair advantage of Bad
Microsoft's OS market power.  Bad Microsoft has the morals of a 1970s Latin
American generalissimo.  

Thus, Good Microsoft and Bad Microsoft should be separated before Bad
Microsoft corrupts Good Microsoft any further.  Both these powerhouses
should then be honored as national treasures.  Good Microsoft should operate
under few unusual restrictions, if any.  It should even be allowed to launch
the Microsoft Network and purchase Intuit.  Bad Microsoft should suffer the
same restrictions IBM did in the 1970s, for the same reasons.  

How can we distinguish between the Good and Bad Microsofts?  Good
Microsoft's application and "content" businesses cater, as they should, to
users' greed.  Unlike IBM, Microsoft has been early and aggressive in using
cool new technologies, from graphical applications to CD/ROM.  It has been a
price leader, notably with Microsoft Office.  While it has had product
delays and preannouncements, it has been no guiltier of these misdemeanors
than many other software companies.  Microsoft has, of course, indulged in
operating system-related shenanigans, such as the OS/2-Windows flip-flop,
which wrong-footed most of its competition.  Otherwise, however, Microsoft
has earned its market share fair and square.  

Bad Microsoft's operating system business legitimately caters to users'
fear.  However, it skipped the part where you start with a quality product
offering, well-supported, at a good price.  Windows PCs are notorious for
their unreliability.  Few people outside Microsoft even know what Windows
actually sells for.  As for support - dream on.  Instead, Microsoft was
handed market share by IBM, and has enhanced it by excellent marketing and
ruthless machinations.  Well-known examples include the Macintosh OS
license, Microsoft's refusal to support the large OS/2 base, its strong-arm
pricing and negotiation tactics, and the multiyear preannouncement of
Windows 94/5/6.  And of course, Microsoft is just as guilty of unethical FUD
(Fear, Uncertainty, and Doubt) sales as IBM ever was:  "Nobody will ever be
them out."

Even worse, Microsoft has an understandable desire to bring its entire user
base forward to each new release.  Thus, instead of writing a solid new
client OS product and releasing it promptly, Bad Microsoft faithfully
reproduces DOS design flaws in 32-bit technology.  Meanwhile, Windows NT,
which is a solid client OS product, by all accounts has a deliberately
crippled user interface.  Similarly, Microsoft is foot-dragging on
object-oriented technology, confusing VBX controls with class libraries, and
OLE's enhanced cut-and-paste with true transaction processing middleware. 
Please note:  Microsoft is not alone in suffering the Curse of the Installed
Base.  But what's good for near-monopolist Bad Microsoft is far from healthy
for its users, let alone the rest of the computer industry.  

Thus, it is imperative for the future health of the computer business that
Good Microsoft and Bad Microsoft be separated into distinct entities.  Bad
Microsoft will still be a powerhouse, but without the applications business
it may compete more aggressively on technical innovation.  Good Microsoft
doesn't need the OS umbrella to prosper; and if a breakup DOES knock $1/2
billion off the value of the non-existent Microsoft Network, who suffers
except a few Microsoft and TCI shareholders?  Software tools like system
management, DBMS, and languages could go with either part of the company;
probably the operating system side would be a better fit, subject to the
same "unbundling" strictures IBM faced.  

Would such a breakup amount to dismantling a national treasure?  Hardly. 
General Electric didn't need the electric utilities; AT&T didn't need the
Baby Bells; and Good Microsoft doesn't need the operating system monopoly to
extend its impressive record of innovation, evangelization, and leadership. 


Curt A. Monash
President, Monash Information Services
212-315-3120
Fax:  212-399-3268
61 West 62nd Street, NY, NY 10023

Since I'm new to this list, I'll give a bit of background on myself for
people who've read this far.  I've been a software industry watcher or
participant since 1981, first as an investment analyst, then as an
entrepreneur, now as an information provider.  My company publishes the
Monash Software Letter, and also Application Development Tools from A to Z,
a lengthy product selection guide.  I'm on this list because I'm exploring
how to bring my paper-based publications on-line.  Also, I'm just interested
because I'm an industry-watcher.

Within the past year, I've consulted for Microsoft, Computer Associates, AND
Oracle, but most observers would agree I keep my independence anyway .


From crobato@kuentos.guam.net Thu Oct  3 02:50:00 1996
Received: from saba.kuentos.guam.net (root@saba.kuentos.guam.net [198.81.233.14]) by cnj.digex.net (8.6.12/8.6.12) with SMTP id CAA25388 ; for ; Thu, 3 Oct 1996 02:49:50 -0400
Received: from saba.kuentos.guam.net by saba.kuentos.guam.net with smtp
	(Smail3.1.29.1 #9) id m0v8hb6-002F3RC; Thu, 3 Oct 96 16:49 GST
Message-Id: