Is Microsoft Smart – Or Just Successful?

Microsoft will spend $250 million on basic research this year, but it can’t shake its copycat reputation. Gates & Company say, “Just wait and see.”

Out In front: Research sociologist Marc Smith (left) and Microsoft Research VP Rick Rashid use a 9-foot map to track Internet discussion groups.

Under the bright, hot stage lights of Las Vegas in November 1996, Bill Gates faced a difficult job: How was Microsoft’s then-CEO going to rev up a jaded, tech-savvy Comdex crowd about the launch of Office 97, a relatively uninteresting, and essentially mature, product category? His answer: artificial intelligence capability, in the shape of Clippit, the Office Assistant.

To get there, MSR has assembled about 480 full-time researchers in Redmond, Wash., with another 150 or so in labs in San Francisco, Beijing (a rapidly growing facility) and Cambridge, England, bringing the total of the Microsoft brain trust to about 630. Microsoft plans to spend close to $250 million on research this year, with about $37.5 million going directly to universities to fund projects and fellowships.

The headcount and budget may seem small in a company with 39,000 employees, $23 billion in revenue, and $9.4 billion in profits for fiscal 2000.

Where do you want to go?

For-another MSR product contribution-e-commerce databases that predict customer preferences. Is there any apparent reason-at least from a short-term product perspective-that such seemingly common innovations have actually to come from in-house? Of course, there is, replies Craig Mundie, Microsoft’s senior vice president of advanced strategies. Integration with an in-house lab is always greater than with a licensee. “We frequently have researchers who leave MSR for our product groups to expedite the transfer of technology, and then return to continue developing advanced software,” he says.

The long haul

As for the long term, Mundie and others at Microsoft make the case that far-out research is valuable because it will ultimately provide a competitive advantage. By having a long-term research lab, Mundie says, the company can stay ahead of the major technology leaps that constantly reshape the market and the company’s products. So maybe right now software that predicts customer behavior offers help in Office or retrieves text might seem woefully imperfect, and fields such as natural language processing and artificial intelligence still have a long way to go. “That’s why these are research problems because they are extremely difficult,” says MSR senior program manager Kevin Schofield. “But, boy, you figure it out, and it’s a huge win.” Microsoft wants to be the one to achieve that win.

If that is so, does MSR really have a shot at it? Despite the rancorous public debate about Microsoft’s ingenuity (one critic chides Microsoft as pursuing “search and redevelopment instead of research and development“), MSR has won the respect of leading academics.

In data mining, for example, MSR has an exalted spot at the top, says Richard Snodgrass, a University of Arizona computer science professor and chairman of the Association of Computer Machinery’s data management group. When it comes to publishing peer-reviewed papers, he says, “They are among the most prolific.” MSR is not simply cranking out mediocre work, either. “They are highly regarded researchers worldwide.” Snodgrass praises the lab’s work with metadata and indexing. He also praises MSR’s development of the TerraServer, a 20-terabyte, Web-based database of satellite images that push the limits of database scale and management.

MSR has also impressed leading researchers in computer-human interaction. Dan Olsen Jr., a computer science professor at Brigham Young University and the vice chairman of publications for ACM’s computer-human interactions group, says MSR is second only to Xerox PARC among corporate labs in the field. In particular, Olsen lauds the work of MSR researcher Ken Hinckley. Hinckley won the best paper at ACM’s User Interface Software & Technology conference last year for his work on handheld devices. Hinckley equipped a PDA with accelerometers so it can sense how he holds and moves it. If he tips the gadget toward a friend standing across from him, for example, it automatically flips its display so that it does not appear upside down to the friend. If Hinckley turns the PDA on its side, the display rotates accordingly.

Of course, Hinckley’s is not the ultimate breakthrough in computer-human interaction. But even though such research is a race of small steps, it is notable that MSR is near the front of the pack. Someday-probably decades from now-MSR might be the lab to announce the biggest step of all in software: passing the Turing test. Proposed by computer science pioneer Alan Turing, the test requires that a person be unable to tell whether she is interacting with a computer or another person. If MSR does make that big breakthrough, look for the demonstration onstage at Comdex.

Microsoft’s Licensing Switch

Microsoft wants to revamp the way corporations buy software. Will it score a coup, or aggravate customers?

For the past year, Microsoft (MSFT, info) has been moving toward a subscription-based model for selling software. Now analysts expect Microsoft to announce changes to its licensing policy soon to keep large corporate customers paying for Office and Windows, indefinitely.

Currently, corporations purchase Microsoft software under a three-year agreement, typically on a per-user basis. When the term expires, companies are free to continue using the software for as long as they wish without paying additional fees. This is known as a “perpetual license.”

If Microsoft implements its new policy, some of its larger customers will either have to continue paying fees at the end of the initial three years or stop using the software. Or, better yet for Microsoft, upgrade to the latest version, which would seem more attractive than paying a fee on older software.

While such a move has obvious revenue potential, it could also turn off some important customers.

Analysts believe Microsoft will have trouble convincing smaller companies to go along with subscription plans, which is why the company is starting with enterprise customers.

“I think they’re starting to outsmart,” King says.

“The highest-end enterprise customers are people who are used to dealing with equipment leases. It’s also the customer base focused most intently on Windows 2000 right now. But when they start getting down to the consumer end, it could be more problematic. They may run into some problems with small-business folks who don’t want to upgrade. Eventually, those copies of Windows 95 and 98 will be superseded, and the company will eventually stop supporting them as robustly as before.”

Posted by on May 10, 2001