Can Microsoft Innovate In A Web-Centric World?

At next week's Professional Developers Conference, Microsoft will unveil both the next version of Windows and its long-awaited cloud computing strategy. But it's not just Microsoft's products that need to change to keep up with the times.

Despite continuing to rake in revenue and produce reams of research -- 1% of all U.S. patents last year -- Microsoft has been largely an afterthought in some of the hottest recent tech trends. Amazon has captured the cloud computing thunder. Google is a search and advertising juggernaut. Salesforce.com brought software-as-a-service to the fore. VMware got the early drop on Microsoft in virtualization. And Apple has all the cool factor in the world.

In the past, Microsoft's often been a fast follower, rather than being the initial agenda-setter, but competitors have quaked at the mere thought Microsoft would be entering their market, and the public didn't flinch at using any Microsoft product handed its way. That was then. Now, more than ever, Microsoft needs to find a way to prove to the public that it is indeed an innovative company. The public is looking for results, and much depends on Microsoft changing the way it innovates.

So in addition to the hundreds of millions of dollars it's pouring into advertising, Microsoft is making better use of the Web for early product testing, opening a series of applied research labs to more quickly turn research into products, and even toying with more grassroots innovation. By putting itself in the public eye more often, Microsoft also has the chance to shed more light on areas where Microsoft could, in fact, be seen as innovative -- like Microsoft Research, Office OneNote, and Windows Server.

"The perception of the company has suffered a bit, particularly given the emergence of other glamorous companies," admits Microsoft chief research officer Craig Mundie. Turning that around won't be easy. The company faces numerous dilemmas, from important legacy platforms making big bucks to growing bureaucracy. But Microsoft should never be counted out, and Mundie maintains Microsoft does a better job than "any other company in history" of delivering products based on its research. Now, to prove it.

Dealing With Legacy

Microsoft's biggest challenges to paradigm-shifting innovation could be its two greatest successes, Windows and Office.

"They involve huge amounts of legacy code, and that's unique to Microsoft," said Michael Cusumano, an MIT professor who's written several books on innovation in the technology industry. "Microsoft has to test anything with a huge code base and lots of other legacy apps built by Microsoft and the community. It's a tremendous asset, but it's also a disability."

Microsoft is slowly adjusting, updating Windows incrementally via Windows Update and Windows Live add-ons, recently introducing a suite of SaaS offerings in Office Online and is poised to unveil its cloud computing strategy, but it's got a long way to go.

Part of the answer might be to test and release features in new ways. Traditionally, new features were just included in the next major release of Office and often went unnoticed and sometimes went without being tested as a standalone feature. Now, Office Labs is doing small pilots inside and outside the company and releasing some new features online.

"We're more likely to try ideas because the cost of trying each one is much cheaper than building the whole thing," said Chris Pratley, Office Labs general manager. "That means we can go faster because it’s a couple of months to put a test out instead of a couple of years. There's value in getting stuff out in beta and seeing what clicks as long as you set people's expectations properly."

Bureaucracy is another great challenge for Microsoft. The company eclipsed 90,000 employees this summer, up 20,000 in two years. Product groups often compete with one another to get similar products to market first. Layer after layer of managers obscure the sight lines from top executives to working engineers. As Microsoft grows, quality control gets deeper and product testing more extensive.

"All the things which made Microsoft more professional and a company enterprises trust more have made it clumsy and slower," Cusumano said.

Though Microsoft has had some of its own grassroots efforts, it has long been more about top-down innovation than a company like Google, where the 20% of time employees spend on pet projects has brought about successes like AdSense and Google News. "I have focused in the 15 years I've been here on what I'll call more strategic top-down incubation," Mundie said. "It really is where a small number of people -- in the past, many times it was me and Bill Gates -- would sit around and talk about an area we thought might be a strategic opportunity."

Inside the company, there are a few efforts to flip that on its head, so far with mixed results. Ad Astra, a group founded by Jon Pincus, former general manager of strategy development in Microsoft's online services group, aimed to set up a grassroots approach to innovation at the company, something that could cut through what Pincus calls the "old boy network."

Ad Astra set up a blog as an open-ended discussion forum, and a wiki to spread ideas. It hosted "Mash-Up Days," live events aimed at creating new ideas like a new search technology that Pincus said "blew Bill Gates away" but has yet to be released. It pushed college interns to come up with new ideas for the company like a way to tie Zune and Virtual Earth together to read driving directions aloud.

Unfortunately, out of the gate, the group was in disarray. "Literally" as he gave his presentation, Pincus' supervisor, then-Microsoft VP Martin Taylor, was pushed out. Pincus struggled to find new executive backing, a budget, and willingness to shift on Microsoft's insistence on patenting everything, and amicably parted ways with Microsoft in November 2007.

Once Pincus left, much of his work transitioned into what is now called IdeAgency, a group that sits under Ray Ozzie. IdeAgency, led by Randy Granovetter, acts similar to an internal venture capital organization. Microsoft employees submit new ideas to a Web site, where they are evaluated and voted on and decisions are made to fund them. Once projects get their initial funding, employees who came up with the ideas build prototypes, and executives bid on them as part of a decision whether to turn the ideas into real products. Any employee can contribute new ideas and products that went through the IdeAgency wringer -- sources won't say which products -- will be out soon.

One of Microsoft's applied labs, Office Labs, has its own grassroots efforts under way, where employees submit ideas, a voting process takes place, and Office Labs prototypes the best. Still, Office Labs general manager Pratley said, opening this up to the entire company might not be the best idea.

"The larger the population, the lower the signal to noise ratio," he said.

One longstanding quasi-grassroots effort at the company has been Gates' Think Weeks, where employees submit ideas twice a year and Gates read through stacks of them, deciding where Microsoft should invest. With Gates gone from full-time work, Mundie will continue the tradition with himself and chief software architect Ray Ozzie as the judges. Think Week also has found a twist inside Live Labs, which hosts occasional Out of the Box Weeks, where employees can show off prototypes they've built. This year will see the first products released that originated with Out of the Box Week ideas.

It's too early to tell how IdeAgency and similar efforts will pan out, but Pincus worries IdeAgency doesn't have the budget or autonomy from top execs to be effective. Mundie said he's hesitant to go further down the path of creating new grassroots innovation efforts until Microsoft witnesses some real success from them. Still, he adds, "you want to bring forward as many reasonable mechanisms as you can to allow innovations to be identified and supported."

The Long Arm Of Microsoft Research

Microsoft Research helps to put Microsoft among top contributors to major academic computing conferences and top patent-creating machines worldwide. Some recent big-name products have come directly from MSR, like the Surface multitouch computer, the RoundTable IP camera, and Windows Mobile Live Search's speech interface. But MSR's wares are rarely shown off, and it's something Microsoft has to change.

"We need to do better at getting in the public eye," said Rico Malvar, managing director of Microsoft Research Redmond. "What you see here doesn't always match with the perception that people have of Microsoft in innovation. Researchers often think the thing that is most innovative is what gets the most awards at the conference. That's important, but what gets to the user and makes the user's life better?"

Microsoft will soon re-launch the MSR Website, making it easier to navigate and find downloadable software developed by Microsoft researchers. Microsoft Research also has been slowly opening up its annual TechFest event more and more to the public and to reporters, showing off its latest developments. TechFest is what Mundie calls an "internal trade show" where product teams look for ideas to put into their products.

Plenty of MSR inventions make it into products eventually -- Mundie said "people would be blown away by the number" -- but many never do. Following somewhat in the footsteps of companies like IBM, Microsoft's five-year-old IP licensing team has begun to mine the company's research for technologies unlikely to be used and puts them in an online catalog. "The goal is to get Microsoft inventions into the market even though we may not be using them," said Tony Bailey, Microsoft's senior product manager for technical licensing. Datacastle uses Microsoft technology to compress encrypted content, and China's Talkweb uses technology that turns photographs into cartoons.

Microsoft Research also is working closer with product teams. Senior Microsoft managers match research projects with product groups, and MSR loans researchers to write code for products. That's both a blessing and a curse, because it means that, as Malvar said, "a good chunk of those innovations go into improving the performance or to a few new features of a big product like Office or SQL Server."

One of Microsoft's newest organizational structures, what it calls "Labs," focuses on building prototypes and early-stage products, typically as part of larger product groups. Among the approximately 20 labs, there's Live Labs, Office Labs, adCenter Labs, and BizTalk Labs. These groups look for new opportunities, using mixed teams of MSR and product group employees to build prototypes and release them to the Web if possible, even before integrating them with larger products.

"Microsoft has got to have the capability to get products in the hands of consumers more quickly," said Alex Daly, Live Labs' group product manager. "And as a leading indicator, I think we're having some success." Live Labs has incubated a number of ideas that have translated into products as well as several others still under incubation. Most notable is PhotoSynth, a Web app that meshes together a set of photographs into a navigable, three-dimensional picture.

The Web has the potential to drastically transform how Microsoft presents its innovations to its customers: new features can be released to the Web as add-ons before they ever hit shelves as integrated features, and new products can hit the Web early during testing.

"I think you find the ability to add a feature, compose it in, and test drive it is, again, much, much simpler than it is when you're delivering a large, integrated piece of software that's sort of bought and installed in a monolithic way," Mundie said, adding that people also should expect more beta software to find its way to the Web than in the past.

Other companies have been doing this for some time. Google, for example, releases almost everything as a "beta" at first, sometimes soon after the germination of the idea. Microsoft hasn't done this as much, but recently there's been a quick succession coming out of Office Labs, such as a new search bar for Office commands and an open source multitouch application that uses a Webcam to sense touch.

Another new way Microsoft is trying to take advantage of the Web is with the Microsoft Experimentation Platform, a recently created group that evaluates ideas for new products and features controlled experiments with actual customers. For example, Microsoft ran an image test on Windows Marketplace to choose between two images that would link to a gaming download section. Surprisingly, a solitaire icon had a 61% better click-through rate than a poker icon, potentially pushing customers to buy more games through the site. There are admittedly few results to measure by, but so far the Experimentation Platform looks promising.

Experimentation Platform general manager Ronny Kohavi brought this idea from Amazon, where a search feature tested in its experimentation platform resulted in a 3% increase in revenue for the company. "The culture at Microsoft was old school, and there wasn't much possibility to interact with customers when I got here," Kohavi said. His customers now include a number of product groups and new tests are taking place weekly. Kohavi runs classes for Microsoft employees on how to use the platform, and on a recent day had 250 people sit through a full-day course.

There are signs of life in Microsoft, but the company needs more than a public relations campaign to regain the allure that once left competitors shaking in their boots when Microsoft entered a new market. The road back to there is potentially filled with potholes, but some structural fixes to the company can help point in the right direction.