2009 : 05 : Powering Up Innovation at Your OrganizationManaging Innovation : An Interview with Todd Warrenby Dian Schaffhauser

Todd Warren joined Microsoft right out of college 21 years ago, when he became the project manager in a three-person team building Macro Assembler. Since then he spent a career at the company involved in managing research and development of products and services that have included Microsoft Exchange, systems management for Windows Server, an email client that eventually became Outlook, Project, Small Business Server, Windows CE, and Windows Mobile, among others. Recently Warren retired from Microsoft to devote himself to other endeavors, among them, philanthropy and teaching software engineering at Northwestern University in Evansville, IL. Warren serves on the Board of Directors of Pcubed. In this interview, he shares his perspective on how innovation works within Microsoft, how to infect an organization with innovation, and why an innovation office shouldn't be the only source for finding the good ideas within your organization.

You've seen how a lot of different groups operate within Microsoft.

Todd Warren: I've had the opportunity to work with a number of great leaders of innovation and see a number of different styles of innovation.

The first project I worked on was tiny compared to the last one I worked on. The first project I worked on was Macro Assembler. It was really pretty simple to manage. It was just three of us - myself as a project manager and two developers. One of the big innovations we had there was the introduction of a separate testing group to test software independent of development. Up until then we still had to test our own software. In that kind of environment, you can run things pretty informally. Everybody can get in a small room and talk to one another.

Contrast that with the last role I had as the leader of Windows Mobile Products Development, where we had over a thousand engineers working on developing the new Windows mobile operating system as well as servicing the mobile operating system for launch on a lot of different mobile devices from a lot of different mobile operators. There, we had to put in place a much more structured methodology about how we worked.

When I took responsibility of the Windows Mobile group about four years ago, one of the first things we did was implement Portfolio Management. At that time we were developing [four] different products at once. So we had just completed Windows Mobile 5. We were beginning Windows Mobile 6 and also a future version of Windows Mobile that's yet to be released. We were also doing a new version of the CE operating system.

There's quite a lot of resource contention among the different projects. The first thing we needed to do was get a good handle on how people were spending their time and how engineers were allocating their resources across the products. Implementing portfolio management was a key part of that. I was the GM of the project, so it was easy for me to put that in place by fiat.

We then began a process of tracking the resource commitments across those different projects. Our ultimate goal was to streamline development overall. In some sense the key innovation was collapsing the number of projects down to just a few. Some of that occurred naturally. We completed the Windows CE release, for example. But others we had to be a little bit more "planful." One of the things we ended up doing was delaying some of the future work, so we could really concentrate on getting Windows Mobile 6 done...

The core thing at the beginning was getting the priorities right, and really understanding where our resources were allocated, then as a manager putting in place a set of processes so we could review and then change the priorities for where those resources were used - and do it in a way that people felt comfortable with and understood the overall direction we were going.

So innovation can happen in the process of production or design or the process of how you operate in an organization, not just in the development of new products.

[With new products] the issues are a little bit different. I'll give you three different examples from my career at Microsoft.

The first example has been on the market now for about 10 years : Small Business Server. Steve Ballmer, who was the president at that time, had come to the Server group and said, "Hey, there are a lot of small businesses underserved by our technology. These are businesses that don't have IT. We'd like to provide an easy way for them to get access to our server technology." Coincidental with that, for the larger IT systems, we had been doing a lot of user research into how system administrators worked. We used a technique that I'm using in my class currently, called "contextual design," to surface user needs. It was a technique developed at DEC in the '80s by an anthropologist, Karen Holtzblatt, and a gentleman named Hugh Beyer, to really [apply] some of the techniques of anthropology to information systems.

As a result of looking at system administrators, we really got a good idea of what things we could collapse and automate for a small business that didn't have IT or had part-time IT or a [value added reseller (VAR)] or system integrator, to help them get things done. That product was quite successful... The particular way we went about it, we really looked at user data both in small businesses and outside of small businesses to put that together.

The second example that I'll give you was the introduction of [Microsoft Office] Project Server when I was the GM of the Project group. In some sense that was more a line extension than a new product. But there our approach was a little bit different. We certainly looked at the user data in a similar way with what we did with Small Business Server, but there we knew there were some other competitors and complements to Project in the market that were doing more advanced portfolio management.

One of our discoveries was that people were getting away from doing just critical path project management to also looking at resource management across their projects. We wanted to add some better resource management into the system itself. In that case we actually ended up acquiring one of the complementary products to Project, but we had to give it a twist that made it extend from Project itself. The results of that was Project Server, which is on the market today. I would say the competitive environment and looking at the industry adjacencies were the keys in planning the innovation - as well as taking advantage of other things that were emerging at Microsoft at the time, such as integration with SharePoint...

The third example was really a new kind of product, which we did actually shortly after the introduction of the iPod. When I was running the Embedded Device group, we looked at what new kinds of devices would be emerging. There we looked at the competitive advantages we felt we had at Microsoft. One of those at the time was some new technology that had been developed around video playback and a related technology around digital rights management (DRM), and in particular this idea of supporting subscription content. We set about developing software called Portable Media Center, that took advantage of the video playback and DRM and some new things coming online with the Windows Media Player, and collaborated between three different groups within Microsoft. That player software actually evolved into the current Zune player that's on the market today.

Now I'll tell you an example that didn't turn out quite as well.

One of the first device projects I worked on when I took responsibility for the Embedded Device group was a thing called Smart Display, which was probably not the smartest thing I've ever worked on. We had this idea that portable components were becoming cheaper and more readily available. Wireless networking was really catching on. We had this great capability in Windows XP called Remote Desktop, where you could remote from one desktop to another and use it. We thought, well, what if you built a lightweight client that was essentially a tablet-style computer - a touchscreen computer that connected back to your PC and could be used in your home as second way to access PC functionality? It turned out to be a very difficult concept to explain to the end user, that this thing was a computer, but it was really the same computer as their desktop computer, and it was remote. [Editor's note : Smart Display was announced in early 2002, released in early 2003 and cancelled in December 2003.]

We learned a couple of things there. One was, it was really much more of a technology-driven project. We kind of did it because we could. I don't think we really understood how it fit into the different ways people would use it. That was an interesting learning experience. It certainly highlighted to me the need to really have a good grasp on user needs and to have a good simple concept for a new product.

These products bubbled up from many different places. You didn't have an innovation officer saying, "Hey, let's do this. It's cool and it's new."

I would say the innovations that I've seen that are most successful are the ones that are aligned with the overall strategic direction of a business or a company. On top of that, one of the important things to do is to layer a process that can really harness the power of all the power of the people on the team to bring it forward.

For example, in the portable media center case, we were doing an embedded operating system using lots of different kinds of devices. We knew that new device types were emerging, so we did a couple of simple things. First, that team was about 300 engineers. Any engineer that wanted to have prototype hardware to try out their device idea, we made sure within reason they could get the hardware to experiment on those projects.

Then we had a way that the engineers could bring those ideas forward. Sometimes we did them. Sometimes we didn't. I would say it wasn't an overly formal process, but we had some time set aside every month to look at what ideas people had, to see whether there was some promise. Things started as moonlight kinds of projects for people. If we reached a stage gate where we felt it was a promising enough idea, then we augmented it with folks from marketing, and [user interface] design, and started to do the business analysis around how to bring it to market.

The Portable Media Center started that way. There were two very talented engineers who had seen the iPod and were using Windows CE to prototype a portable audio player. We brought that forward and said, "We don't want to do something that's exactly the same as what Apple's done". So we want to figure out how we can couple some Microsoft competitive advantages. That's when we started linking it with some of the work that was coming out in DRM and video. That was a more bottoms-up way.

Small Business Server was very much top-down. The CEO - at the time the president - identified a space that he felt we needed to go to, and we got that very simple mission. And then we looked at the user research and shaped the product. But the direction wasn't more specific than figure out a way for small businesses to better utilize our server technology.

Would an innovation office - staffed with folks from, say, marketing --have helped to establish a need there?

Yeah, I think so. There's a great book called Scale and Scope : The Dynamics of Industrial Capitalism, which covers a lot of topics. One of the interesting chapters in that book for me was the discussion of Dow Chemical. Dow in a lot of ways is the model for how most industries think about research and development and innovation. One of their key insights was, it didn't make sense to just have a centralized Dow research group, although they had that doing more fundamental research. But associated with each business, they'd allocate some percentage of their resource to looking at new products and adjacencies. And then the role of the governance structure is how to lace those together.

So my involvement was more at the business unit level, having some work set aside for innovation. Then we'd work with the other business units to see where the competitive advantages are. I think any organization would be well off to put in place a process where they can both get the bubbling of ideas - so a very low overhead process - for the talented people in the organization if they have an idea to put it together and get it reviewed, and then of slowly staging the expansion of the ideas that show promise in the system - thinking through the different stage gates. The first one could be small - a one page proposal. Then maybe a small allocation for budget for a group of engineers to build a prototype and then, based on the promising prototypes, getting some senior staff from across disciplines involved to build the business case around this.

Does it make sense to impose meetings and deadlines on those people?

I think having a timeframe associated with each of those stage gates makes a lot of sense. There's a whole concept of failing fast in innovation.

I think one of the things for instance I learned from that failed project, smart displays, was we would have been much better off to have taken a stage gate process where we developed and tested a prototype, seen how customers reacted, and really used that as guidance for how we wanted to bring it forward. And setting it in a time-bounded way, so we wouldn't extend a lot of resources on something that didn't work. In that case, we extended resources to bring the whole product to market.

So innovation can be approached incrementally?

I think it should be approached incrementally in terms of figuring out how to scale it up, especially in terms of new product development. Having some staged way of looking at it and scaling it up at the appropriate time in its lifecycle makes a lot of sense.

Sometimes what happens is that projects scale too quickly, and it's very hard for some of those projects to absorb all of the resources thrown at them. A company realizes that it needs to respond to a particular competitive threat and allocates a lot of resources. But the resources can end up bumping into each other without a very thoughtful scale up.

How do you infect an organization with innovation? Give me three steps...

The first step is to let the organization know that you value innovation and new ideas - in all aspects of the business. Process innovation for getting these things worked on faster. Whether it's a new product idea. Whether it's line extensions to the current set of products to help serve customers better and more profitably. Setting a tone that management's open to those ideas. And then second is putting in place a process that gives it teeth. At the smaller organization, it could just be an informal policy that people can bring their idea forward. In a larger organization it can be putting in place a more structured review process for those innovations. Third, be somewhat directive about the kinds of areas where you think there's possibilities for innovation so that people in the organization have some guidance for the areas in which the overall business is looking to evolve.

Todd Warren is a board member with Pcubed and Ashesi University. He worked at Microsoft from 1987 to 2009 in a variety of roles, his most recent as corporate vice president for Mobile Product Development. Todd is now a professor of software engineering at Northwestern University. Contact Todd at twarren@pcubed.com.