Exploring Alternative History

Tablet circa the year 2000

Bloomberg has a nice piece revisiting the broad vision Microsoft laid out in June 2000 at the grandiosely named Forum 2000. They even have an edited cut of some of the scenario videos developed for the event to bring the vision to life.

The event was originally scheduled to open the new millennium in January with a bang and more importantly to lay out Microsoft’s vision for the 21st century. The company felt intense pressure to reassert its thought leadership on at least three fronts. Microsoft, by virtue of being a business with actual profits, probably had an even greater relevance problem during the height of the dot com mania than it does today. Second, the unrelenting antirust assault on the company was building to its climax and the company desperately wanted an alternative narrative. And, largely forgotten today, the company was very concerned about demonstrating it would not lose a step in the transition to its brand new CEO, one Steven A. Ballmer, who had just taken the reins from Bill Gates.

The event was pushed from January to March and then ultimately June because the dates kept overlapping with antitrust rulings. That event kicked off a train of events that lead us directly to Microsoft’s current situation and the search for a new leader. It definitely provides a glimpse of a different trajectory for Microsoft over the last decade.

The original process to define Microsoft’s forward vision was through a set of technical committees made up of Microsoft’s best and brightest technologists. Despite high hopes and leaks to the press that such a process was under way, this approach was pretty much a complete bust. As that became apparent, Steve asked Paul Maritz to oversee an effort to pull together a compelling vision for the company. Paul described this exercise as “creating a new parade”.  A small and fairly eclectic group was formed of people from across the company but largely outside the engineering ranks. We sketched out the broad attributes of what things going forward would look like and then five sub-teams were tasked with making this brave new world come alive for consumers, small business, enterprises, knowledge workers and an end-to-end healthcare scenario that illustrated what developers would be able to build. Each group scripted and then shot professional quality videos from the user perspective.

I don’t think people appreciate how close Microsoft came to completely imploding in 2000. Employees woke up every day to relentlessly negative headlines from the DoJ case. It was not yet evident that the surreal world of the dot com bubble had ended, and even if you weren’t being wooed daily to join revenue-less startups with ridiculous valuations, you felt obligated to explore options.

Forum 2000 changed all that internally. It provided a sense of purpose and showed how the whole could be greater than the sum of the parts. Soon after Forum 2000, Mark Lucovsky, one of the key contributors to Windows NT, the inventor of Win32 (he also likes to brag he invented “DLL Hell”) and one of the company’s distinguished engineers, showed up with an architecture to make the Forum 2000 videos real (the ability to actually implement had not been the foremost consideration in painting the vision…). This lead to .NET My Services (aka hailStorm) which was announced in the spring of 2001.

.NET My Services was a cloud-centric model using web services to deliver consistent, personalized experiences across a wide variety of devices (including non-Microsoft devices – the very first device demonstrated was a Palm). It was also an API-first model, which meant any endpoint could access the system (this is an approach the Facebooks and Twitters claim, but regularly violate in the name of “protecting the user experience” which really means protecting ad revenue). It was a major departure from Microsoft’s traditional PC-centric platform and also introduced a subscription business model as opposed to the traditional license model. It was an open platform accessible from any device or service through open XML-based protocols but could be bootstrapped using Microsoft’s vast footprint of devices, applications and services. It represented a fundamental shift to a service-centric world, both technically and in terms of business model.

By 2001, it was clearer that the dot com mania had been a giant bubble and wasn’t coming back, plus the appeals court ruled that Microsoft would survive the antitrust proceedings intact. The confidence stemming from these positive external factors, however, ultimately undermined Microsoft’s desire to invest and realize the new vision.

Competitors like Sun Microsystems, no doubt horrified that their years of lobbying had failed to hobble the company, now faced a revitalized Microsoft with a vision for the future that was compelling to customers and driving the industry discussion. When they probably should have been figuring out how to save a business shackled to the dot com ship, Sun embarked upon a very effective campaign of demagoguery around Microsoft leading the shift to a user-centric model. As a result, My Services ushered in the first big industry debate around identity and privacy in the cloud. In retrospect, Microsoft’s personal computing heritage and fundamentally user-centric approach to give users full control over how their data would be shared looks vastly superior to today’s world administered by the almost interchangeable Big Brothers of Facebook, Google and NSA.

With the competitive and existential threats from dot coms and the DoJ having abated by this time, Microsoft chickened out on seeing it through. It was easy to back down from the industry debate over identity, shirk the challenges of figuring out the subscription service model and revert to the comfort and familiarity of good old Windows and Office. Harvard Business School later did an interesting case on the tensions between the old and the new camps inside the company and how it played out.

I don’t think it is widely appreciated that WinFS was born of a desire to realize the My Services scenarios, but to do it in a Windows-centric way. There was broad agreement on the importance of the scenarios, but strategic nostalgia for Windows resulted in the company trying to rethread them through the franchise and revisit the eternal dream of integrated storage. Instead of the truth being in the cloud, the truth would reside on Windows and everything would have to sync with your PC (just don’t turn your laptop off…). This decision triggered a sequence of events that directly brings us to the present day where the erosion of the Windows franchise played no small role in Steve Ballmer’s departure,

My Services was shut down, with CTO and WinFS cheerleader David Vaskevitch dismissively telling the team the company “didn’t need another $500 million business”. The fixation on WinFS technology brought down Longhorn, the release of Windows scheduled to follow Windows XP. Simply put, WinFS was too ambitious technically. After much internal debate, Longhorn came to an end with the “Longhorn Reset” whereupon the company embarked on the far less ambitious Windows Vista (and WinFS was never to be seen again, although I have flashbacks when I listen to the Hadoop guys talk today). However, given it had been over five years since Windows XP had shipped, the company felt pressure to rush the product out the door to meet obligations to customers who had paid for a new version of Windows as part of their enterprise agreements. Hence the “Vistaster” of shipping a half-baked version of Windows.

The company then spent three more years cleaning up the quality, performance, haphazard user experience and packaging of Vista, resulting in the very solid Windows 7, but failing to move the PC industry forward in any material way during that time. Windows XP remained the most popular version of Windows and Microsoft was forced multiple times to extend the end of life of Windows XP by a customer base that was just not compelled by multiple subsequent releases.

Meanwhile, iPhone and iPad (and imitators like Android) were in market and Microsoft’s Post-PC nightmare was looking very real. The company decided to focus Windows 8 almost exclusively on tablets, hoping to pull the tablet category back into the PC universe. Except that didn’t happen. Windows 8 on tablets received mixed reviews. Surface was a costly mistake, both financially and in its impact on critical OEM relationships. And it screwed up the desktop experience for billions of PC users (one senior Microsoft executive told me Windows 8 was only for tablets, but didn’t answer my question of why they neglected to mention that in the advertising). Which brings us to the present and the search for a new CEO.

In retrospect, this sequence of events is crystal clear in a way it never is in the fog of the present. Even with greater commitment, there are a million other ways the Forum 2000 vision could have gone wrong. Parts of the vision were dead on, others such as assuming tablets would rely on a stylus were big misses. It was still predicated on bootstrapping from the current Microsoft installed base, which would force a myriad of tradeoffs between old and new every day. There were major business challenges to overcome in building a successful subscription business, particularly as the Google advertising revenue volcano was just beginning to erupt. Microsoft subsequently spent years drooling over the prospect of hundreds of billions of dollars of advertising moving online, without fully internalizing that capturing those revenues would require behavior that was in many ways antithetical to the personal computing ethos at the core of the company. The good news is the rise of the tablet kept Microsoft from turning Windows into an ad-funded desktop billboard monstrosity.

The Price of Success

And of course, Microsoft faced the innovator’s dilemma in spades. The last decade of Microsoft’s history is a classic and very public case study on how a very successful company deals with disruption (disruption it knows is coming). The dissipation of the Forum 2000 vision was very much the result of a battle between seemingly reactionary forces exercising their fiduciary responsibilities and the hazy dreamers of a less distinct and unproven future. The counter-revolution obviously prevailed, at least temporarily.

Some have even argued that Microsoft did the right thing by maximizing Windows profits for as long as it did (and is still doing even though Windows profits dipped slightly below $10 billion in the last fiscal year). Horace Dediu’s recent podcast “The Limits of Executive Power” has an interesting take on this (and smart commentary recently on Microsoft is hard to find):

We begin with a defense of Ballmer for preserving great things, continue by condemning him for not having destroyed those very same things and end by asking whether anyone could have done the right thing.

The Innovator’s Dilemma came out in 1997. We all read it at Microsoft and were looking for disruption behind every tree. Discussions about the need to cannibalize Windows before someone else did have been going on for at least 15 years. And in the meantime, the company has banked profits from Windows alone in excess of $100 billion (I have not done the actual math but the number is of this order of magnitude). Clearly undermining that profit stream proactively 15 years ago was the wrong thing to do, but how should the company have avoided its current situation? The company invested in pretty much every kind of non-PC device including smartphones (Microsoft was the leader in this space as recently as 2006), tablets (albeit with styluses) and a bunch of goofier form factors.

I believe the fundamental problem was the unwillingness and/or inability to transcend the single device (UPDATE: I should make explicit that this is a reference to Tim O’Reilly’s popularization of Dave Stutz’s farewell from Microsoft missive which was written after the My Services experience). Microsoft had the vision and means to both lead the industry and bridge its existing businesses to a cloud-centric, multi-device world, but failed to follow through. Now it finds itself belatedly embracing this model but from a disadvantaged position. It is yet another technology industry example of innovations conceived in one place being successfully commercialized elsewhere that lacks the baggage of the conceiver.

Near the end of my tenure at Microsoft, I was in a meeting with a cast of thousands. One of the presenters said “we wouldn’t want to do another hailStorm”, expecting all the heads in the room to nod in unison. The SVP in the room turned to me and said “we should have done hailStorm” to which my answer was “damn straight”. The rest of the room was aghast. I suspect this viewpoint is not as contentious now, even at the board level. Success can be a bitch.

4 thoughts on “Exploring Alternative History

  1. Pingback: Microsoft flashbacks show what could have been — Tech News and Analysis

  2. Pingback: Microsoft flashbacks show what could have been | 8ballbilliard

  3. Pingback: Microsoft flashbacks show what could have been | Earthgrid

  4. Peyman

    Good stuff.

    HailStorm was a nice idea, but the implementation was a mess — over-engineered, unnecessarily complicated, and burdened by Microsoft’s strategy tax. Authenticating via Passport — the first step in building any app — was just awful. Five years later, when we integrated Passport into Max, I remember a good engineer needing a couple months to do the work b/c there were all these wacky edge cases the Passport APIs made you handle.

    Contrast that with what developers have to do to authenticate with Apple’s iCloud services, say to store a file or geolocate an address: nothing. The OS & APIs do it behind the scenes. iCloud has its own problems and we’ve all learned a lot about web services in the last decade, but even so, Windows could have managed the user’s Passport credentials, system wide, so that services like HailStorm could use it, so that 3rd party developers wouldn’t have to.

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>