Bonum Certa Men Certa

Microsoft Confidential (on OEMs)

Category:OEM Date:



Microsoft Corporation Tel 206 882 8080 One Microsoft Way Telex 160520 Redmond, WA 98052-6399 Fax 206 883 8101

Microsoft Memo

To: Chris Smith, Bernard Vergnes From: Joachim Kempin Date: July 11, 1991 Subject: OEM Sales Directions for FY'92 and Beyond

The attached memo was distributed to all OEM managers, to use at their discretion within OEM Sales. I want to share the information with you, but would like to keep the distribution as limited as possible.

OEM Sales Directions for FY'92 and Beyond

While the first half of FY'91 was dominated by an overpowering enduser vote in favor of Windows 3.0 - disregarding DRI's new version of DOS 5.0, the last half of the FY could be named the breakthrough for Windows applications and the breaking up of the 10-year MS-IBM alliance. With Lotus, Word Perfect and Borland poised to introduce their key Windows apps this FY and IBM pushing OS/2 2.0 as the new standard through newly formed alliances, it is time to analyze the MS OEM situation and give you some advice in how to win during the years to come.

The MS-IBM Relationship Revisited

Formed 10 years ago, this alliance created the largest compatible computer market ever. Approximately 70M PCs and well over 400M units of compatible software sold. At the same time, IBM's market share eroded to less then 15% of the total PC market while Apple, through it's MacIntosh, was able to defend 10% overall market share. While IBM allowed competitors to clone their architecture, Apple kept its proprietary. MS' role in the compatible PC business grew from the language and OS supplier in the beginning into the number one applications supplier for both compatible IBM PCs and Macs while enhancing its number one position as the key OS supplier for PCs in introducing Win 3.0.

Neither UNIX nor OS/2.1 x ever gained critical mass on IBM compatibles with combined sales of less then 1.5M units compared to 5M units of Windows and close to 70M units of MS DOS. The Mac with approximately 8 - 10M units installed taking and overall second place. Win 3.x run rate for this FY is predicted to add another 6 - 8M units to its installed based and should therefore - assuming a 2 0M unit run rate for the Mac - gain second place in FY92 in overall PC OS installations. This outlook, which will put us even more clearly into the OS driver seat, cannot please IBM or Apple understanding that both companies are making most of their money by selling proprietary solutions.

Back in 1987, MS and IBM announced the next generation of operating systems called OS/2. This new OS was supposed to take over from DOS and Win 2.x by allowing ISVs to develop more complex and sophisticated applications to enable PC architectures to compete successfully with mini-systems. OS/2 was plagued from the beginning by the late deliveries and poor code, demonstrating again that excellent software can only be written by small teams, at a single site, sharing a similar culture. DOS, Windows and the Mac OS are prominent examples of this. In understanding this best, MS convinced IBM to change the OS/2 development method by giving IBM the responsibility for OS/2.1.x and 2.x, while retaining the ownership of DOS, Windows, and OS/2 3.0

When Win 3.0 success demonstrated to IBM that the endusers were drifting away from OS/2 even further - the IBM culture could not cope with it. They started to throw more people (estimated to be 1000) at its standard and extended edition development and at the proprietary application efforts still to be delivered. OS/2.1.x and 2.x for IBM were meant to regain control over the overall computer industry and the Win 3.0 success was blamed for their eroding market share and was seen as loosing control over their customer base. Apple answered with a lawsuit and drastic price reductions, enabling them to hold onto their market share at the cost of their profit margins. IBM reduced prices as well - not as aggressively as Apple - and decided to throw even more good money after bad money by flexing its marketing muscle in promoting the next vapor-version of OS/2 more aggressively. Instead of actively selling Windows and riding the wave as a lot of clone manufacturers did successfully, they decided to go to war and terminate the MS-IBM alliance in public.

Knowing that if you can't win a war with a better strategy, th second best way to win as by alliances. IBM started to search for them. Early signs go two years back when they acquired rights to NEXT software technology followed by the Methaphor joint development agreement. During the last 12 months, the pace of new alliances increased by endorsing GO and Novell, teaming up with Micrographix, Borland, and Lotus, and more recently, with Apple. During my 20 years in this business, I have seen a lot of comparable announcements being made without any success. Just look at IBMs situation to understand the nearly impossible tasks they are facing.

The users are routing for Windows and DOS, IBM is promoting OS/2 2.0 and while promising to make this new OS successful in integrating Apples' Mac technology, Borland's object technology, Novell's networking, and Lotus' groupware hooks, they want to keep AIX alive and are announcing a new portable and scalable OS they want to develop together with Apple to dominate the industry once again. All this Will be done by integrating the Patriot object work as well as migrating to three different architectures, the 68040, 386, and RS6000. Sure enough, Micrographix will guarantee Windows 3.x compatibility to make the user transition easier than ever. Who is going to believe that mighty IBM can do all this and deliver on time? And again, it contradicts our belief that successful OS are written by closely knitted teams, in single site, and not by companies with completely different cultures - being competitors at the same time - and large integrated committees who take longer to decide than to develop.

Common sense is telling me they will not succeed - but nevertheless, lets stay alert to make it even harder for them to win even a single design-in. MS OEM Sales is the only sales force out there who can stop their efforts right from the beginning by being nice to our customers and strengthening our bonds. All of the clone manufacturers, as well as the ACE consortium, will need us to defend their market position and should see us as their natuaral ally. The IBM alliance has nothing to offer than FUD (Fear Uncertainty, and Doubt), and be assured they will try to seed a lot of it until they have something to sell Our action plan of action is straight forward!

1. DOS 5.0

Secure long-term contract with OEMs, whereby the standard contract length should be three years instead of two to deny entry to any attempt of IBM to license DOS directly. Never lose on price - escalate fast in case you see some competitive pressure. This should ensure a steady revenue flow to enable us to invest in advanced OS development.

Growth areas for DOS remain to be in penerating more smaller manufacturers, the embedded system market, and manufactures in development countries. For al lthree market segments we will introduce new pricing policies to help you gain share.

2. Windows

Agressively pursue long-term licensing oppurtunities - in particular, for 386SX systems and beyond. The more happy users we will have, the harder it will be for any other OS to prevail. In addition to the basic system, we need to explore licensing of Pen Windows and MM Windows Extensions to deny any other architecture to be more popular. Remember, GO is an alliance partner of IBM, and both Apple and IBM have announced that they will license their MM technology. Any activity of these competitive offerings in your accounts should be reported immediatly. Windows pricing - because we do have a retail sales alternative - will not be as aggressive as DOS pricing, meaning it will remain your personal challenge to convince PEMs to bundle and pre-install it as an integral part of their offering. We will start to offer incentives for pre-installations ot OEMs to reduce both our support cost. Try hard - never give up on OEMing Windows!

During Q4/92 we will introduce a Windows networking component, as a peer-to-peer instant network solution - without a need for dedicated network servers. As soon as product plans get further developmed, we will aggressivly promote this solution through the OEM channel, whereby the target OEMs will be te ones who sell to the small and medium business customers. This should compliment our LAN MAn offering through a new concept and enable portables and desktop systems to wirk in a more efficient way than today by sharing files and linked models transparently. Please di not discuss this with OEM's until further notice.

Windows 4.0 is probably two years away and needs DOS 6.0 to run. It will be developed for todays DOS/Win 16-bit user and enable him to run 32-bit applications an Intel OCs. Expect both systems to be more integrated than today and to enjoy installable file systems. As soon as development plans get further developed, we can use the information to give our OEMs a better understanding of the future for DOS and Windows for the Intel architecture. I am willing to give customers rights to both version for signing up for more than 3 years with us.

3. NT for RISC and 386/486/586

In its ultimate destination - say beyond the year 2000 - this advanced OS kernel technology with its possible DOS, Windows, and OS/2 subsystems is designed to be the "DOS of the year 2000", enabling us to return to one standard. To position it this way today would be a mistake, like it was with OS/2. The first goal for this new system is to establish it with its Windows subsystem as the dominant workstation architecture for MIPs and high-end Intel PCs and to compete against SUN's UNIX workstations. It is therefore important to sell it on a per-system to all ACE manufacturers and leave the SCO or possible SUN UNIX offerings to the retail channel. I am convinced that certain system features like 3-D graphics will be brought first to NT mode Windows before you will see them in Win 4.0. Expect the NT kernel - over time - to contain network server calls as well to truly compete with Novell's offerings beyond Windows instant networking features. While ddevelopment plans get firmed up, we in OEM should focus on licensing NT to the ACE community on a per-system basis without speculating on the future development.

4. Networking

The effort to sell LAN MAnager 2.x to OEM should continue, complimented by our retail efforts. We no longer believe that IBM's LAN Server will be sold seperatly from EE and be based on 2.1 code. The next step in FY93 will be to sell NT server technolohy bit our product plans need to settle further before we should spend any time on this today.

5. OS/2 3.0

The development of this OS as contractually agreed upon between MS and IBM will continue as planned. It will be based on the NT kernel, support 32-bit aplications written for OS/2 and Windows, and will probably have a DOS emulation subsystem as well. Under the currrent situation it is completely unclear if IBM will ever use this OS or if we will eveer activly market it. We therefore will currently make no attempt to seel its component beyond the kernel and the WIn subsystem to OEMs.

6. OS/2 2.0

We will start licensing the OS starting this quater aggressively. Despite the fact that we do not believe in it's sucess - it is still better for us to take the lead and get the design-ins instead of IBM getting an inroad into our OEMs. Undeer the current agreement with IBM, we have the rights to binary and the source codes and are getting weekly "software updates". We believe that the software will be released by the year end, but will lack a lot of promised features and will not be properly tested.

We are planning to pass the software through to our OEMs without adding additional value. An exception wil lbe the inclusion of the LADDR drivers. The same will be true for the documentation. The IBM original wil lbe delivered as is and all localizaton will be done by IBM without any IPG involvment. I expect our OEMs not to be too pleased with this, but at the same time all blame should be directed towards IBM - as promises not kept, and as a result of their non-cooperation with the industry.

Note: Please leave the discussion of the IBM/Apple-MS relationship to the MS executives. Do not bash IBM or Apple in front of customers, they will hurt themselves by not delivering anyway. Instead, tell your customers that it is time to get even closer to us when being threatened by the "big boys" and that we will take care of them. Reflect this "take care" attitude in dealing with our OEMs to help us form new alliances to replace the old ones.

And on a more practical basis, you should compare the IBM/Apple "openness" with the SUN situation. Who do you think will spec any new system, or get its hand on new chips and new software first? IBM/Apple's competitors - you know better, stay with us!

Full Exhibit