Microsoft's Missed Opportunities: Memo From 1997

New submitter gthuang88 (3752041) writes In the 1990s, Microsoft was in position to own the software and devices market. Here is Nathan Myhrvold's previously unpublished 1997 memo on expanding Microsoft Research to tackle problems in software testing, operating systems, artificial intelligence, and applications. Those fields would become crucial in the company's competition with Google, Apple, Amazon, and Oracle. But research didn't do enough to make the company broaden its businesses. While Microsoft Research was originally founded to ensure the company's future, the organization only mapped out some possible futures. And now Microsoft is undergoing the biggest restructuring in its history. At least F# and LINQ saw the light of day.

Comments

Hindsight's twenty-twenty

[rmdingler]

It is so difficult to stay on top in any field, let alone atop a technology that changes virtually overnight, that even Microsoft's relatively short run as apex predator was commendable.

You can make a hundred correct predictions in a row as to where the market is heading, and then whiff on two, and an apple or a google gain a foothold.

It's not rocket science... it's way harder than that.

Re:Hindsight's twenty-twenty

[plover]

Microsoft has done some really brilliant things as of late. They've wholeheartedly adopted automated testing for everything. I don't know if they have any product teams that aren't Agile, or aren't doing test driven development. I recently asked a product manager about his product's defect backlog, and he shot me with a cold stare: "We don't have any known defects in our product. As soon as a bug report arrives, the entire team drops what they're doing, and within 15 minutes a developer is working on repro'ing it, and it's fixed within a day. These are very rare occurrences." This was for a million line shrink-wrapped product.

Although it's taking them a long time to turn their teams around, Microsoft finally knows how to engineer code right, and they are quite willing to share with anyone willing to listen. But too many of their clients don't listen, too many of their vendors and suppliers don't listen (driver bugs, etc), too many of their own internal teams are still dragging legacy code bases forward, and they still have a long history of bugs that we all remember. Another problem they have is economic: their primary competition is their old products, like Office 2007, which are good enough for most businesses and students. They really want to get everyone on their Azure cloud, using Office365, live, OneCloud, and to rent computing resources from them, and that's driving a lot of their products in an unnatural direction for their consumers.

Their marketing people haven't helped. Windows RT? Really, they had to emulate Apple's walled garden? The closed iOS ecosystem is about the worst thing Apple ever did to their customers, The Apple tax sucks 30% from every dollar spent on the platform, and there's virtually no escape. And because we all know it sucks, we won't willingly jump into it again - so Microsoft loses even more.

Their forays into other platforms have been abysmal: Ford's SYNC is a crime against drivers. They bought a failing phone company for their hardware, turned out walled garden phones, and nobody showed up. Their previous attempts at embedded systems make people WinCE. And because they start everything out as closed source, and try to contain their own stuff, they see every product as a battle entering competition to the death, instead of an opportunity to cooperate. That got them a long way, and made them a lot of money, but now there are good alternatives, and nobody gives a damn anymore. The stuff they're producing now will all be too much, but way too late.

Microsoft "At Home" lab is a bust

[derinax]

In the late nineties and into the last decade Microsoft just dumped too much time and money on their vision of a hyper-connected home. They dumped so much research money into building out test spaces and building out test devices, they failed to realize that people don't want an intelligent dryer and an intelligent toaster and an intelligent melon baller. The reality is whatever fancy device you own that has any kind of transistor in it, much less a CPU-- a phone, a tablet, a TV-- you're having to fuss with it. Constantly. And the same is/was always true for their "Microsoft At Home" vision. And yes, these things were connected-- but only to each other.

That, and the fact that Microsoft has always misread the Internet, from coming to TCP/IP late, to ignoring the vital interoperability that cloud services demand. It's always been about the toys with them. Toys that run Windows. Ugh.

Gratefully, only a few of these monstrous things ever saw the light of day beyond the lab.

Re:Maybe MSFT was trying to learn from Xerox

[Solandri]

Maybe MSFT was trying to learn from Xerox, Kodak, and other companies that pioneered technologies and then failed to follow through.

While Xerox deserves full blame for missing opportunities (the mouse, GUI, ethernet, and laser printer were all invented there), Kodak does not. They were always on the forefront of digital imaging. They built the first digital camera in the 1970s, and had a line of digital SLRs in the early 1990s. They knew exactly where the industry was heading, and in fact did most of the early R&D to get us there. The only reason they managed to hang around as long as they did was because they owned most of the patents on digital imaging and were collecting massive royalties.

What led to Kodak's downfall is obvious if you look at the pictures in that wikipedia link. Those are Nikon (and later Canon) bodies with Kodak digital sensors. Kodak was a film company, not a camera company. They weren't in the business of making cameras (aside from some cheap consumer models and disposables). When the industry shifted from film to digital, the companies which ended up on top were companies skilled at making cameras/lenses (Canon, Nikon, Olympus, Zeiss, and their arch-rival Fuji which had been busy making decent point and shoots prior to the switch to digital), and companies skilled at making electronics/silicon (Sony, Panasonic, Casio, etc). Kodak thought they could carve a piece of the digital sensor pie for themselves, but rapidly found themselves unable to keep up with companies with decades of expertise manufacturing microprocessors who simply shifted that expertise into manufacturing sensors. In other words, the best business model for making camera sensors turned out not to be knowing how to make camera sensors. It turned out to be knowing how to make microchips.

MS Research was meant to mop up talent, that's all

[echtertyp]

Waaay back I remember someone pointing out that Microsoft was spending enormous sums to hire researchers, especially promising ones in academia. The idea, apparently, was for MS Research to be a sort of "intellectual roach motel" (love that phrase) were IQ would check in, and nothing checked out. This made a certain amount of sense. As a monopolist you don't -want- any innovation. One way to do that is hire hitmen to kill potential innovators. But the risks there are huge. A much easier way if you have the money is to hire promising minds and then keep them neutralized. That's just what Microsoft did.

Re:Too long

[Anonymous Coward]

In order to justify a budget increase of 300%+, the head of Microsoft Research had to write a really long essay beginning with business buzzwords (like embark, unprecedented, and endeavor) and ending with some justifications for his recommendations.

Yep, Myhrvold's memos were always substantial, they often defined the future of the company. This is from a New Yorker article in 1997.

Reading the memos chronologically, one can look at some of the business decisions that Microsoft faced during the years it grew to a nearly nine-billion-dollar giant that in 1996 earned two billion one hundred and ninety-five million dollars. It’s easier to understand the company’s path to success: a rare marriage of technical and business prowess.

Myhrvold's role was essentially to be the futurist at Microsoft. He was their forward thinker and gave them the geeky excitement that allowed them to make many of the right choices throughout the '80s and '90s. Ignoring him and concentrating instead of the business and litigation-driven path resulted in the gradual slide to the barely relevant, spiteful and fading dinosaur, shedding workers and market share we're saddled with today.

Imagine instead if they'd listened to him and worked towards this vision:

Myhrvold then turned to what he called “the truly personal computer—something which has the size and weight appropriate to be carried with you at all times.” This wireless “digital wallet,” as he called it, would allow anyone to communicate, untethered to a wire, by voice, video, fax, E-mail, or pager. The device would be a clock, an alarm, a schedule manager, a notepad, an archive of phone numbers and records, and a library of music and books. The digital signature produced by this wallet would have a personal I.D. for security, and could replace cash, credit cards, checks, and keys. He believed that the obstacles were economic and human, not technological. “The cost will not be very high—it is pretty easy to imagine a total cost of manufacture in the range of $100 to $250 on introduction, which means $400 to $1000 retail price,” he wrote. He guessed that keyboards would be superseded by devices capable of recognizing handwriting.

http://www.newyorker.com/archi...

OP is saying 22 pages is too long a memo to bet the company on, and gets modded insightful? Why?