1
PC Week Extra!        ZNT:LET-14

This document was typed from a photocopy.  Some portions of the memo were
illegible and were left blank.

Challenges and Strategies

Bill Gates May 16, 1991

Prologue:  The Reason for this memo

Every year I set aside at least one "think week" to get away and update myself
on the latest technical developments-reading PhD theses, using competitive
products, reading books, newsletters and anything I can get my hands on. Several
valuable thoughts have come out of these retreats (tables for Word, outlining in
Excel, treating DOS as more of an asset), however the complexity of the industry
and it's technology means that a lot of my time is spent just trying to keep up
rather than coming up with new product ideas.  It is no longer possible for any
person, even our "architects", to understand everything that is going on. 
Networking, processors, linguistics, multimedia, development tools, and user
interfaces are just a subset of the technologies that will effect Microsoft.  My
role is to understand enough to set direction. I enjoy these weeks a great
deal-not because I get away from the issues of running Microsoft but rather
because I get to think more clearly about how to best lead the company away from
problems and towards opportunities.  A lot  of people choose things for me to
read.  By the end of the week I make an effort to synthesize the best ideas and
make our technical strategy clear.

This year I decided to write a memo about overall strategy to the executive
staff.  As we have grown and faced new challenges my opportunities to speak to
each of you directly has been greatly reduced.  Even the aspects of our strategy
that remain unchanged are worth reinforcing.

In the same way that DEC's strategy for the 80's was VAX-one architecture, one
operating system-our strategy for the 90's is Windows-one evolving architecture,
a couple of implementations.  Everything we do should focus on making Windows
more successful.

A source of inspiration to me is a memo by John Walker of Autodesk called
"Autodesk: The Final Days" (copies available from Julie G.)  It's brilliantly
written and incredibly insightful.  John hasn't been part of Autodesk management
for three years and hasn't attended any management meetings for over two years,
so he writes as an outsider questioning whether Autodesk is doing the right
things.  By talking about how a large company slows down, fails to invest enough
and loses sight of what is important, and by using Microsoft as an example of
how to do some things correctly he manages to touch on a lot of what's right and
wrong with Microsoft today.  Amazingly his nightmare scenario to get people to
consider what's really important is Microsoft deciding to enter the CAD
market-something we have no present thoughts of doing because it would stretch
us too thin.  Our nightmare-IBM "attacking" us in systems software, Novell
"defeating" us in networking and more agile, lower cost structure,
customer-oriented applications competitors getting their Windows act together is
not a scenario, but a reality.

Recently a long time employee mentioned that we seem to have more challenges
facing us now than ever before.  Although I agree that it feels that way I can
say with confidence that it has felt that way every year for the last 15.  We
decided to pursue a broad product strategy from the very beginning of the
company and that means we have a lot of competitors.  Our success is incredible,
not just within the software industry or computer industry but within the
history of business, and the combination of this with the incredibly competitive
nature of our business breeds challenges to our position.  I think it is
critical to divide these challenges into different categories.

Category 1

This category contains issues of great importance but which I judge should have
little effect on how you do your job in the future.

Apple lawsuit:  This is a very serious lawsuit.  If the judge rules against us,
without making it clear what we have to change or asks me to eliminate something
fundamental to all windowing systems (like overlapping windows) it would be
disastrous.  At the very start of this lawsuit we decided that Bill Neukom and I
would give it very high priority and that the rest of the executive staff could
focus on their jobs without learning about the complex twists and turns of the
lawsuit.  Microsoft is spending millions to defend features contained in every
popular windows system on the market and to help set the boundaries of where
copyrights should not be applied.  I think it is absurd that the lawsuit is
taking so long and that we are educating the third federal judge on the case.  I
am pleased with our work on this case. Our view that we will almost certainly
prevail remains unchanged.

Federal Trade Commission:  It must be surprising that our two most vbisable
problems are in this category.  Certainly I take the FTC inquiry seriously and I
am sure it will use up even  more executive staff time than the Apple lawsuit
has.  However I know we don't get unfair advantages in any of the markets we are
in.  As Ruthann Quindlen stated recently in InfoWorld (supported by many other
editorials like Business Week's) our combination of products is similar to that
of every other high technology company and our success is based on having great
products.  I hope we can quickly educate the FTC on our business.

Retirement of key executives:  The retirement of Jon Shirley and Jeremy
Butler-absolutely two of the finest executives anywhere-are significant losses
for Microsoft.  Last year's "think week" was my worst because Mike Hallman
called me to say Jeremy was planning to retire.  I had Jeremy fly out and meet
with me for hours to try and change his mind.  I am sure more people will be
retiring in the future.  However, I am confident that we are developing a lot of
great people internally  and that we are hiring the best people from the outside
the company.  Just look at some of the recent additions to our executive
staff-people like Brad Silverberg, Jeff Raikes, and Gary Gigot. Consider the
talent pool right below the executive staff level-Jim Allchim, Pete Higgins,
Patty Stonesifer, Rob Glasser, Mike Murray, Mike Brown and so many others.  I
love working with people of this caliber-not only do they do a good job but they
keep me doing my best.  I certainly have no plans to back off from my dedication
to the company.

Printer business unit:  Generally when we enter a product category, we innovate.
Even if our first version is not a winner, we establish a position from which we
can make further improvements.  Our entry into the printer software business has
not succeeded.  Steve is considering what strategy we should pursue to make the
best of our errors.  Our problems have educated us to consider carefully the
importance and synergy of doing new things. Offering a cheap Postscript turned
out not only be very hard but completely irrelevant to helping our other
products.  We overestimated the threat of Adobe as a competitor and ended up
making them as "enemy", while we hurt our relationship with Hewlett-Packard and
focused on non-Windows specific issues. Selecting TrueType as our font solution
and building it into the system was an excellent decision despite the immense
resources that has cost us. TrueType- our font format-separate from
TrueImage-our Postscript clone. Printing is critical and we will be involved in
printing software, but in a different way than we have to dare.  The caution we
have shown in making acquisitions is reinforced by this experience.

Category 2

These are problems that are serious, but solving them correctly will provide
growth so they can be thought of as opportunities.

Dislike of Microsoft/openess:  Our applications have always succeeded based on
their own merit rather than on some benefit of unfair knowledge of system
software.  We need to explain our hardware neutral approach and the benefits
that has generated for end users.  We need to have visable events on a regular
basis where we solicit the input of anyone who wants to influence our future
direction.  If we can institutionalize a process that the world feels
comfortable with, we will strenghten our position incredibly. This is going to
require a lot more creativity than even the "Open Forums" we are discussing.
UNIX has OSF and X/Open-we also need clear ways for organizations of all types
(hardware, ISV, IHV, corporation, universities) to feel like they have something
invested in our approach and can affect our course.

IBM:  IBM is proposing to take over the definition of PC desktop operating
systems.  This would be a new role for them-their previous attempts: Topview and
the 3270 control program, did not succeed.  The barriers to their success are
not only technical but structural.  Why are they willing to lose so much money
on systems software?  The answer is that they have a plan to design the
operating system so that their hardware (MCA) and applications are tied in. Our
disagreements with IBM over OS/2 were that we wanted to push 2.0 and they wanted
to push 1.3.  Now they have switched to the strategy we proposed-even using our
marketing slogan  "better windows than Windows."  We will not attack IBM as a
company even our public "attacks" on OS/2 will be very professional. Our
strategy is to make sure that we evolve the Windows API and get developers to
take advantage of the new features rapidly, while IBM has a poor product with
poor Windows functionality.  Amazingly they are not cooperating with us on our
compatibility approach called WLO, but are trying the approach we did not choose
of using Windows code itself.  Their lack of cooperation limits WLO
effectiveness and the Windows approach has contractual and technical problems
for them.  We will do almost no work specific to OS/2 2.0- we will rely on their
1.3 compatibility to run our applications and most of our networking software. 
Our focus is on OS/2 3.0.  If a customer buys OS/2 2.0, the problem for us is
that they will expect to get Extended Edition and perhaps some PM16 applications
that may not be on 3.0 so we will have "lost" that customer. Other than
usability, making sure Windows is the winning OS is our highest priority. 
Eventually we need to have at least a neutral relationship with IBM.  For the
next 24 months it may be fairly cold.  If we do succeed, then we will be done
forever with the poor code, poor design, poor process, and other overhead that
doing our best to do what IBM has led to (for the last five years).  We can
emerge as a better and stronger company where people won't just say we are the
standard because IBM chose us.  In the large accounts IBM will retain a some of
its influence-this is where our risk is highest.

Usability/support:  If there is any area we have not paid enough attention to it
is usability/support.  It is really embarrassing that people have to wait so
long on the phone to talk to us about a problems in our products.  The number of
customers that get a bad impression because of this must number in the millions
worldwide.  Why weren't we hiring at full speed and picking new site every day
for the last three years?  Why did people keep talking about support as a profit
center?  The creation of support as a channel hid its costs from the product
groups.  As CEO I take full responsibility for these mistakes.  Our product can
be far more usable and the product groups are focusing on this
opportunity-particularly the Windows and Windows application groups.  We will
spend what it takes to have the best support (without an 800 number.)  I think
we can cut the number of phone calls generated by our products to less than half
of what it is today and use training and technology to cut the length of phone
calls.  However, we shouldn't assume this in our plans to solve the problem. 
Excel 3, with Word 2 and our EBU products have started to move in the right
direction.  Hopefully Windows 3.1 will generate lot less calls.  The handwidth
of communication between the product groups and PSs is going up dramatically,
but there is still lots of room for creativity. I insist that we are able to use
quality of support as a sales tool.  surveys like the JD Power survey done on   
will become important-asking people How many times were you confused?  How many
times did you have to call?  How good was the service you received?  Fixing this
problem will cost us a lot of profits and we should make that clear to analysts.
With this problem fixed we can really start building some lifetime customers. 
Only really usable software can be used by the "rest of the people who have not
bought PC's", so making software more usable expands the market.  Likewise it is
the usability of software that will determine how many people decide to use only
a WORKS- like product or move up to a larger package and it will determine how
many large packages they can easily work with.  Usability is incredible
stuff-once it is designed it is easy to implement, saves money, wins ,market
share, makes customers happier and lets them buy more expensive software!

Networking:  We knew it wasn't going to be easy, but it has been even harder
than we expected to build a position in networking.  You will see us backing off
on some of the spending level but don't doubt that we are totally committed to
the business.  Our strategy is to build networking into the operating system. 
Some of the services will not be in the same box but they will have been
designed, evangelized, implemented and tested as part of each operating system
release.  What this means is that we will define operations on and attributes of
entities like files, users, machines, mail, printer or services that users or
applications can have access to directly inside the system software.  Although
we will allow connections to different systems we will make ours the easiest to
use by          some of them and making all of them seemless.  Architecturing
the extensions for these entities including our evolution of the file system and
how we tie in with standards like Novell and DCE will be Jim Allchim's
responsibility even though the implementation of several of these will be in
other parts of the company (for example OS or Mail.)  We are in a race to define
these extensions because Novell's dominance and DCE's popularity could allow
them to usurp our role unless we get a strong message, good tools and great
implementations done fairly quickly.  We will embrace DCE as a weapon against
Novell although we don't know exactly how to relate to DCE quite yet.  Our
strength will come from Windows, including the advanced implementation based on
NT.

Technology:  Technical change is always a challenge for the current companies in
a field.  Even if they recognize that a change is taking place, they are tied to
the past.  New companies will move to exploit the opportunity.  Our gain in
applications is in no small part due to the failure of the existing leaders to
listen to what we and other people were saying about    GUL. Technical change
can be a new hardware platform like NeXT, a new type of machine like Pen or
Multimedia, a new software platform like Patriot Partners, a new category, a
redefinition of a category or a much faster development methodology.  Many of
the changes that will take place in PCs can be anticipated (performance, memory,
screens, motion video), however, understanding when and how is still quite
complex.  Other changes like linguistics, reasoning, voice recognition or
learning  are harder to anticipate.  We will reduce our technical risk by
strengthening our relationship with the research community and having some
projects of our own in areas of greatest importance (development environments
and linguistics, for example.)  Nathan (and Kay Nishi before him) has pointed
out that the transition of consumer electronics to digital form will create
platforms with systems software-whether it's a touch screen organizer or an
intelligent TV. The need to work closely with Sony, Philips, Matsushita,
Thompson and other Japanese consumer electronics companies will require people
people in both Tokyo and Redmond working with both the research and project
groups in these companies.  We should have an annual exchange of research
thinking with most of these companies similar to what we want to do with MIT or
Stanford.  We have the opportunity to do the best job ever in combining research
with development in the computer field largely because no one has ever done it
very well (although Sun Apple are also working hard on this.)  Nathan's kickoff
memo talks about about having the research group use our tools and including
program management inside the research team.  Our proposition is that all of the
exciting new features can be accommodated as extensions to the existing PC
standard.  Others propose that start-from-scratch approaches are cleaner and
therefore better.  This is the essence of the debate with Go, NeXT and Patriot. 
To win in this we have to get there early before significant development
momentum builds up behind the incompatable approach.  The key to our Macintosh
strategy was recognizing that the graphics and process of the PC would not allow
us to catch up soon enough to prevent Mac from achieving critical mass so we
supported it.  Sun presents a particular challenge to us because they have
significant development backing and high end features to go with their RISC
performance.  ARC is the most evolutionary way to get to RISC and it will
require alot of good execution by us and others for the strategy to succeed.

Our evolutionary proposition should be quite marketable to users-combined with
hardware neutrality the message is "Our software runs today's software on all
(almost) hardware and both today's and tomorrow's software on all (almost) of
tomorrow's hardware."

Category 3

This is a category of challenges we face that I don't feel are widely
recognized.

Patents:  If people understood how patents would be granted when most of today's
ideas were invented, and had taken out patents, the industry would be at a
complete standstill today.  I feel certain that some large company will patent
some obvious things related to interface, object orientation, algorithm,
application extension or other crucial technique.  If we assume this company has
no need of any of our patents then they have a 17 year right to take as much of
our profits as they want.  The solution to this patent exchanges with large
companies and patenting as much as we can.  Amazingly we haven't done any patent
exchanges that I am aware of.  Amazingly we haven't found a way to use our
licensing position to avoid having our own customers cause patent problems for
us.  I know these aren't simple problems but they deserve more effort by both
legal and other groups.  For example we need to do a patent exchange with HP as
part of our new relationship.  In many application categories straightforward
thinking ahead allows you to come up with patentable ideas.  A recent paper from
the League for Programming Freedom (available from the legal department)
explains some problems with the way patents are applied to software.

Ridgity/Pricing:  In the Autodesk memo, Walker talks about the short term
thinking that high profitability can generate.  He cites specific examples such
as a very conservative approach to giving out free software or a desire to
maintain fixed percentages for the wrong reasons.  Microsoft priced DOS even
lower than we do today to help get it established.  I wonder if we would be as
aggressive today.  This is not a simplistic advocacy for just lowering our
prices - our prices in the US are about where they should be.  However the price
of success is that people fail to allow the kind of investments that will lead
to incredible profits in the future.  For example we have gotten away without
funding any internal or external research.  Nathan is working with me to put
together a plan that will end up costing $10M per year about two years from now.
I have no plan to reduce our spending in some other category by $10M.  Microsoft
is good at investing in new subsidiaries and even at investing in new products
(database, mail, EBU, networking).  Most of our rigidity comes where we have a
very profitable product and when the market changes.  In these circumstances we
should spend more or change less, but our system locks us into staying the same
and losing share.

My largest concern about price competition comes from Borland.  Organizations
smaller than Borland will not have enough presence or credibility to use low
price against us broadly.  I think 90% of the significant competition we will
face in productivity applications will come from Lotus, WordPerfect, Borland,
Claris, and IBM barring technical innovation by small companies.  It is amazing
how similar the applications strategies of Microsoft, Lotus, IBM or Microsoft,
so he can afford to do things we would consider wild.  For example Borland is
considering not offering their Windows word processor separately but integrating
it with Quatro for free-the technical opportunity and value would be very
strong.  This is very different than lotus temporarily offering Ami for free. 
Only immense loyalty to a product at the end user level prevents corporations
from using     buying power to         cheap site licence.  when the US
Government DOD moves software procurement to a separate contract, the price per
user of software will end up around 0.  Why shouldn't some small organization
price their product at say $1M for the entire US Government for all time?  We
would if we were small and hungry.  Fortunately most organizations don't force
cheap software on their end users.

Another price concern I have is that companies will eventually equipt all
employees that need software with a full complement of packages, and our only
revenue opportunity will be upgrades or ephemeral information.  Although this
problem is over five years away, I think it is important to keep in mind.

Summary

Readers of this memo may feel that I have given applictaions too little air
time.  I don't mean to downplay their importance at all.  Applications have been
the primary engine of growth (especially in International) over the past two
years.  Although Windows' success is necessary for Microsoft applications to
succeed it is not sufficient.  Other ISVs will be there early with good
applications fully exploiting the environment (Notes, Ami, Designer), so
exploitation is only half of the job.  The need to "reinvent" categories and the
way they relate to each other is crucial for all of our applications.  I will be
writing up some of my ideas for big changes in applications.

The simplest summary is to repeat our strategy in it's simplest form-"Windows-
one evolving architecture, a couple of implementations and an immense number of
great applications from Microsoft and others."  The evolution refers to the
addition of pen, audio, multimedia, networking, macro language, .32-bit,
advanced graphics, setup, a better file system, and a lot of usability.  The "a
couple of implementations" is a somewhat humorous references to the fact that
our NT based versions and our non-NT versions have a different code in a number
of areas to allow us to have booth the advanced features we want and be fairly
small on the Intel architecture.  Eventually we will get back to one
implementation but it will take four years before we use NT for everything.  I
would not use this simple summary for outside consumption-there it would be more
like "Windows-one evolving architecture with hardware freedom for all users and
freedom to chose amongst the largest set of applications."

Although the challenges should make us quite humble about the years to come I
think our position (best software company setting many desktop "standards") is
an enviable one and our people are the best.  The opportunity for us if we



!