Falling on their Face: Six Incidents from Corporate History
May 27th, 2000 - Chuck Flink (Feedback Appreciated!)
Last Saturday's article, "Perspective on Integration, Bundling and the Law" was picked up by ZDNET News under the title, "Freeze! It's a catch-up!" That commentary was based on my observations about recent news that Sun and Oracle were bundling their products (hardware and software) into "vertical stacks" to better serve the e-Enterprise market where they claim to have the dominant platform. The authors of the ZDNET article suggested it was "ironic" that Sun and Oracle were actively practicing "bundling" after complaining to the Department of Justice (DoJ) about just such practices by Microsoft. I took exception, claiming it was not ironic but an action that should have been expected: while Microsoft was frozen in judicial limbo, Microsoft's competitors should be expected to merge forces (and software and hardware in this case) and try to catch-up to and blow by the leader! Momentum counts for a lot in this race. To miss this opportunity would be foolish in the extreme.
I thoroughly enjoyed the feedback to the article ("The author replies: ...Wow!") One respondent, however, felt I was not credible in suggesting many Microsoft competitors simply "fell on their face", leaving Microsoft in the lead. If they fell, he believed it was more likely due to Microsoft using monopoly power to trip them up! In my original article, I said I could supply examples but didn't have time to do so before. Forty-eight hours without power (severe storms here in North Carolina) gave me a chance to put some words on paper (the old fashion way).
I love watching any battle of ideas. I love ideas and enjoy the interplay of one against another. We all enjoy the progress that results! No where else do we see more rapid idea-driven progress than in the Information Revolution. No where else do we see more clearly how the interplay of ideas can shape our future. The list of victors and vanquished is long, though the history of computers is barely more than 50 years. Ideas have flowed over this battlefield like water to the lips of the exhausted combatants, bringing renewed life to the near vanquished and washing away the grime of past defeats. Here are outlines of just six battles in this war, chosen to describe how a market leader is challenged and responds. These battles outline critical points in Microsoft history, but similar battles are fought at every level, in every micro-market where there is competition. It's a glorious game of "king of the hill" where ideas, not muscle mass, is suppose to crown the victor.
Six incidents in the history of Microsoft:
1) the IBM deal: The first incident gave birth to Microsoft. IBM was far more dominant in the Computer Industry of the time than Microsoft is today. The DoJ was on IBM's case, big time. Maybe that's why IBM management ignored the possible impact of a couple of college drop-outs that offered a ridiculously good deal: an OS and a Basic interpreter for a tiny royalty and the right to license these products to others in a market that IBM "knew" wasn't worth worrying about. The result was the end of one dynasty and the start of another. IBM had every bit of power any monopoly could have had. It "owned" the mainframe business, which was the entire computer industry of the time. But IBM let these kids with their “toys” blow by them. The towering icon that defined the computer industry for two generations “fell on its face” dealing with a challenge from two drop-outs.
2) the Apple GUI challenge: A few years later, selling MS-DOS, Basic and some games, Microsoft wasn't much, but heck, it had the bulk of the growing PC industry under it's belt and IBM's reluctant blessing. Then along came Apple's Lisa:. the Graphical User Interface (GUI) was now big news. Apple had a great product and a grand opportunity to make DOS look like the junk it really was. Microsoft went into crisis mode and hobbled together a terrible kluge of a GUI in the form of Windows 1.0 / 2.0.
Apple had every technical justification to blow by MS, but made a huge foolish mistake: it threatened to sue anyone who copied its idea and further decided to keep their GUI proprietary to their hardware. MS ignored the threat of suit; after all, the idea was really Xerox's idea anyway. Apple was just the first to bring it to the PC market. But more importantly, Apple had a highly proprietary product (making lots of enemies among competitors) while MS was making partnerships with multiple PC vendors, playing one against the other for lower platform prices and greater market share compared to Apple. DOS-based Windows eventually became workable by release 3.0 and barely decent later, but never mind the quality of the product, it was available from many PC vendors, virtually always cheaper than corresponding Apple products.
Apple, by playing the proprietary-product card, failed to meet the market demand in a market they themselves started!
(Note: I’m very impressed at the turn around at Apple. Apple fans need not flame me; I'm talking about history here! They need only look at Apple’s past and wonder if they might not have had a much better deal if Apple had sold their GUI vigorously through an array of Apple clones!)
3) In 1988/89, UNIX was growing by leaps and bounds. AT&T’s Unix Systems Labs and the Open Systems Foundation each had dozens of allies and were slinging “roadmaps” at each other, claiming to have the best view of the future of the computer industry.
It was obvious that networking, security, multiprocessing, multi-platform, and multi-vendor was the future. Obviously (to all of us UNIX-lovers), that DOS/Windows kludge was nothing but a sick joke of a toy. A prototype that should never have left the lab… worse, it was a copy of Apple’s copy of a prototype at Xerox Palo Alto Research Center. All the good computer scientists knew that the future was already under development at CMU, Berkeley, Bell Labs and other great centers of UNIX-based innovation: X-Windows, NeWS, TCP/IP, Distributed Computing Environment (DCE), Kerberos, C++, Object Oriented Programming, etc.
The “roadmaps” of the various standards committees simply spelled out how each of these “excellent features” would be integrated into product over the next few years. Microsoft (probably Gates) looked at these and again saw the danger… the challengers were about to blow by Microsoft into the new “networked” world, leaving the boys with their toys in the dust, just the way they had left IBM a decade earlier.
Gates went out and bought the best Operating System architect willing to challenge UNIX dominance and the race to Windows NT was begun. By time Windows NT 3.1 was finally released (some say the “Not a Toy” version of Windows 3.1) the UNIX community had largely self-destructed into a series of battles between vendors who were sure their fellow UNIX vendors were out to get them. The roadmaps that were waved so bravely in 1989 were never followed; USL was sold to Novell and then dismembered. OSF stumbled along and finally joined X-Open and the attempt to “standardize” UNIX faded into history. Though IBM, NCR and DEC did field DCE implementations that spanned UNIX and mainframes, Sun and other UNIX vendors scoffed and followed ARPA funding, forsaking much of the early research and standardization efforts. Meanwhile, Microsoft fielded a cheaper, stripped down DCE in the form of the Windows NT Domain model which is today, with Windows 2000 beginning to meet the goals of the original DCE effort. Lesson: if you're going to publish a roadmap of where you're going, be sure you get there before your competition!
This bit of history makes me sick. In the case of the Apple/Microsoft encounter, Apple was the proprietary ‘fool’ and suffered for it. In this case, Microsoft first tried to join the UNIX community early on (remember XENIX?) but was roundly ostracized from the community. Apparently Gates saw a flaw in the UNIX brotherhood that we all missed. We missed the self-destructive Not-Invented-Here attitude that ultimately doomed the Open Systems revolution. We forced Microsoft into a one-against-many position, not unlike the position Apple chose. We should have won. Instead, we fractured into a mass of incompatible (enough) UNIX variants and proceeded to blame each other for the failure to meet the #1 promise of UNIX: source-level platform independence.
In this case, we committed fratricide while Microsoft took our plans for a Distributed Computing Environment and made it the centerpiece of their entire Enterprise business. I don’t know about the rest of the UNIX community, but living though this history from my position in Bell Labs, I felt we did fall on our faces!
4) AOL loomed as a threat, growing in visibility with a custom "user friendly" GUI that ran on top of DOS and just about anything else at the time. It didn't need Windows and actually ran better without it. As the AOL "desktop" grew, it became a threat to Microsoft’s key platform. It was clear that on-line services would add value to any desktop, MS needed to react with their own on-line service or they’d see AOL become the “platform of choice” for new applications. Windows 95 was escalated from being yet another upgrade to Windows 3.X to becoming the vehicle for MSN, Microsoft's new online service. Following AOL’s technical model, MS contracted with AT&T for a nationwide X.25 access network. This new access network, under the MSN name, was scheduled to be available the day Win95 hit the street.
Looming in the background, however, was the growing importance of the Internet. Both AOL and Microsoft didn’t see this coming until late ’94, early ’95 when it suddenly became obvious that MSN would be an “also-ran” if it tried to compete with AOL in the X.25 access arena. Microsoft turned the release version of MSN into nothing more than a tool to call up over the X.25 network and download the “new” MSN based on the Internet protocols. The August release of Windows 95 included this (already historical) MSN interface.
On December 7, 1995, Gates made his “Pearl Harbor” speech, redirecting his company to focus on adopting the Internet protocols and standards at every level in the Microsoft product line. The result, after a series of free Internet “updates” to Windows 95, was an MSN that abandoned the extensive AT&T X.25-based access network of the year before and signed with UUNET to provide Internet-based access. (Within a couple of years, AOL adopted a proprietary TCP “tunnel” to allow it to also offer access through ISPs, starting the “bring-your-own-access” form of lower-cost service now available from all of the on-line services.)
Preparing to fend off an attack from AOL, Microsoft very nearly made a fatal mistake that could have constituted the ultimate “falling on their face”: they very nearly missed the Internet revolution!
That one year was critical to the Microsoft “monopoly” we hear so much about. In 1995 they could have lost it all by being the same type of ‘proprietary snobs’ we in the UNIX and Apple communities became. I think we have to admit our guilt in the clarity that comes with hindsight. And yet, the claim is leveled against Microsoft that somehow they knew they were a monopoly then and were set to wield monopoly powers against a little start up called Netscape. To me, it appears to have been another close call.
5) Netscape was founded by other ‘kids with toys’, escaped from graduate studies in Illinois and summer work at the National Center for Supercomputer Applications (NCSA). The ‘toy’ in question here was the Mosaic browser. Netscape took the brainpower trained under government funding to be expert in the new World Wide Web technologies of HTTP, HTML and the concept of a Web Browser, and had them develop a proprietary version which they proceeded to give away until they reached 90% market penetration. The intent was originally to give away the “razor” but make money selling the “blades”, or more precisely, give away the browser but make money selling Web Servers.
This was an excellent opportunity. Mosaic was multi-platform: Windows, Macintosh and UNIX versions had been developed and shown to work well. It was obvious that an entire range of services could be offered under the World Wide Web (WWW) model. If the AOL “desktop” appeared to be a reasonable challenge to Microsoft, imagine what a clean, server-centric (i.e. relatively light “client side”), vendor independent service could do. This is especially true when you consider all the free download hosting that comes with offering a ‘free’ browser! Thousands of sites would give up gigabytes of bandwidth to deliver over the Internet Netscape's ‘free’ browser, each download multiplying the value of Netscape's server side “product”!
Wow! Another great blow-by opportunity!
It is very hard to say Netscape “fell on its face”. Selling out for $4 Billion after only 8 years as a company certainly doesn’t constitute a failure in the mind of most observers! The original business plan was blown by 'free' web servers from Apache, Microsoft and others. The ability to profit from sales of the browser was foreclosed by Microsoft's free Internet Explorer and the fact that Mosaic was still free for non-commercial use and available at a low price from Spyglass for commercial users. In retrospect, Netscape probably made the best decision for the shareholders when it turned down Microsoft’s overture for some type of partnership or buy-out. But, from a technical point of view, it was a grand opportunity missed!
Background: Microsoft already had in place the concept of browsing proprietary and local resources over the LAN/WAN in the form of the Windows Explorer. This came with the first release of Windows 95 and was the key improvement over Windows 3.1. By early 1996, Microsoft had a crude Internet Explorer running side-by-side with the Windows Explorer for browsing the non-Windows resources: HTML via HTTP, GIF files, JPG files, etc. The separate IE 1.0 was foolish duplication that cried out for integration with the Windows Explorer. Further, on the smaller Windows PCs, the overhead of running both Windows Explorer (to support the START menu and file directory listings) and Internet Explorer (to support Internet browsing) made Windows 95 run like a pig when browsing the Internet. Running a copy of Navigator merely increased the number of incompatible DLLs loaded at any given time negating much of the advantage of Netscape’s more efficient and complete implementation of the browser concept.
For a moment, suspend disbelief and consider what could have been:
a) the Netscape Navigator could have replaced Windows Explorer and become the standard shell for future generations of Windows;
b) the integration of Navigator with the shell, directory and file management services of the Windows operating system could have leveraged a similar deal with Apple and the UNIX community (e.g. Sun);
c) the concept Microsoft pioneered of including an HTML “template” in each directory, making each directory effectively a “web server” would be standard across all computer platforms!
Note: the “html template” idea enables Windows 98 and later Microsoft systems to present to the user the entire breadth of graphical, scripting, and formatting power (fully standardized by the W3C) upon displaying the contents of a directory . This could have been available through Netscape licensed software on ALL three platforms!
(Search your Windows 98 or later system for *.HTT files. These hidden files control the rendering of a directory when viewed "as a Web page". Click on "My Computer", make sure the "View | view as Web page" option is selected, then highlight the C: drive. You should see a pie chart reporting the free space on your drive. Highlight the other objects to see explanation. The use your imagination to turn any directory into a customized web page appropriate to the contents stored there. Warning: the lack of user identity and file security in default Windows 98 makes this much scripting power a serious danger on shared computers. Either keep your PC personal or go for Windows 2000.)
Now, many will say that Microsoft would never have made a deal with Netscape that would have been that favorable. It would have been tantamount to the deal IBM made with Microsoft giving birth to the “little giant” that is now a monopoly. Remember, however, that IBM made its deal with Microsoft because it was ridiculously cheap in the short run. I believe there is the possibility that we would be talking about the Netscape “monopoly” now if they had negotiated with Microsoft a bit more.
Note: I’m not condoning any “conspiracy” that Microsoft may have proposed to Netscape. I don’t know the facts and I doubt even the participants in the discussions can be sure when the talk around the table was serious and when it was merely fleshing out the opportunities available. This is as irrelevant to my argument as any analysis of the bundling discussions between Oracle and Sun. These two companies, according to the original ZDnet article, claim to be bundling the two dominant components of e-World success: Sun’s HW/OS platform and Oracle’s DBMS platform. Imagine if Netscape and Microsoft had struck a similar deal for "integrating" the Windows desktop and Navigator!
Summarizing the Netscape / Microsoft interaction, both seem to have done what was best in the interests of their stock holders, with the possible exception of Microsoft because of the potential consequences of being accused of monopoly power. It was a difficult call to make: partner with our challenger or build a clone of your challenger's product. I will not pretend to know which action was the “more moral” one. Certainly Microsoft saw how integration of WWW browsing with Windows Explorer functionality would reduce overhead while improving functionality (e.g. HTML directory templates, HELP in HTML format, sharing the same rendering engine with the browser, etc.) It was also clear that Netscape feared that if each file system directory was effectively an HTML server, the unique value of their Enterprise Server products could be diminished. Regardless, the result was that Netscape chose not to integrate the browser with any operating system, effectively giving the field of the integrated browser to Microsoft.
Note: It is certainly possible for an operating system to be structured such that all “user level” functionality is performed through vendor independent interfaces: e.g. “start” menu services, directory viewing, application loading and management, backups, etc. Doing so would certainly assure a great deal more competition in the design of user interfaces. It would also engender a great deal of new costs for training and support as the number of possible “user environments” explodes. If Microsoft argues they cannot remove IE from the OS, they are telling a lie; if however they say they cannot remove it and still have the interface they market under the trade name Windows, they are right.
You choose which path actually reduces overall cost to the users of computers: having a common GUI and Web API across all platforms, or having multiple GUIs and varied Web APIs. I’m only willing to say that I’m sure, in retrospect, we’ll come to see that one side or the other here “fell on its face” and missed a glorious opportunity for a technically superior solution. In a sense, I think both companies fell. Microsoft is obviously moving toward greater and greater OS independence by supporting W3C standards, becoming the lead application supplier to Apple customers, investing in cross platform middleware, etc. Netscape obviously cashed in very successfully. However I believe Microsoft would have moved further and swifter toward supporting other OS platforms, broadening its market even faster, if it had been able to harness Netscape as a partner; and Netscape would have been a long term contributor to the the state of the art if they had taken better advantage of the opportunities to partner rather than to 'win'.
6) Internet Appliances: I could go on, but I’ll cap this list with one final example of “falling on your face”, and make this one a double (triple?) header: Sun and Oracle both claimed roughly 3 years ago that the PC was irrelevant. Their view was (as ultimately is mine) that the vast majority of computer users don’t need to own a computer, they need to rent computer services, hosted on the network, through what is effectively an “Internet appliance”. It matters little if we call this appliance a “Javastation” or an “Network Computer” or any of the other names Sun, Oracle, (and most recently AOL) give it. I call the device a “Screen” and argue for this as the ultimate consumer PC platform in my Ultimate Internet Appliances article. I have no argument with the validity of the concept and fully believe this is where we’ll be in as little as five and certainly no more than ten years into the future.
The “falling on their face” aspect of this situation is that both Oracle and Sun claimed they would be releasing product within a year of their announcement and implied it would seriously erode Microsoft’s lead in the market! I, apparently like Microsoft, believed they could do it! I foolishly thought Microsoft’s dominance was so fragile that either of these companies (or more recently, AOL) had the chance to blow by the king and take over leadership in a stunning victory of hardware and software integration over the massive hoard of obsolete interfaces that Microsoft has to carry forward.
So what happened? First, Microsoft bought WebTV, staking that end of the spectrum of Internet appliance space. Second, Microsoft partnered with Intel, Compaq and others to produce “legacy free” PCs that allow much of Microsoft’s historical baggage to be dropped in favor of cheaper and simpler devices that compete well price/performance wise against proposed competing products. Third, Microsoft licensed Citrix technology and integrated it into Windows NT and then Windows 2000, making serving-up “virtual desktops” a business for a who new generation of ASPs and encouraging the X-Terminal makers to add support for these protocols to become Windows Terminals vendors.
In the 3 or 4 years since the Internet appliance idea became threatening to Microsoft, the company has addressed the problem on 3 fronts and made a business out of selling server platforms tuned for their flavor of “internet appliance”. This is computer business equivalent of the type of overwhelming response that the U.S. military has practiced a half-dozen times in the last decade (and I know you can think of many other incidents in the government outside of the military where overwhelming force applied swiftly was key to success.) It is equivalent to the all-out responses from Microsoft when they thought they were about to lose to competitors because of the GUI and Internet revolutions.
Where is the Sun Java Station? Where is Oracle’s entry in the race? When will we see the AOL entry in this race? There is no doubt that Microsoft had the advantage of their Windows NT platform and the huge cash reserves their prior successes enabled. But it is very clear that Sun and Oracle made promises they could not keep, effectively falling on their face and ceding the market (for now) to Microsoft. (Don't accept my characterization of progress on this front! Search for articles on Javastations and Network Computers. Go back as far as 1986 to see original claims.)
There is reason to wonder if Microsoft would have applied overwhelming force if it weren’t for the overblown claims of their competitors. Should Microsoft have assumed their competitors would fail? Should Microsoft have acted like the fat/happy monopoly it is suppose to be and discounted the impact of a new idea on the market? Should they have acted like IBM did in example 1 above? I believe the shotgun blast approach was wasteful and misguided, but it had to be done in the face of the competitors’ claims that this technology spelled the end of Microsoft's run.
Conclusion: I don’t think Microsoft “has it right” now, because the technology is always changing and the information technology marketplace will be quite different in a very few years. Nor has Microsoft always dealt fairly with their consumers or competitors. They may well be deserving of a split-up on ethical grounds and it may well be true that, in the long run, it will be better for the industry. But I hope my journey down memory lane has convinced you that “knowing you are a monopoly” is difficult to do in the face of active competition, and demonstrated how monopoly status can come about because competitors fail to execute on their own plans or promises.
I leave it to the appeals court and history to decide if Microsoft is better broken up; I leave it to God to decide if Microsoft was more or less ‘moral’ than their competitors. I leave it to you to consider the history and make your own decisions for the future.
Comments always welcome!