2008年9月15日 星期一

Interview: Microsoft CIO Tony Scott

September 15, 2008 4:00 AM PDT

Interview: Microsoft CIO Tony Scott

SAN FRANCISCO--Last week, Microsoft Chief Information Officer Tony Scott was in town for an environmental conference and he took some time out to chat with me about his still relatively new role.

Tony Scott is CIO of Microsoft.

(Credit: Microsoft)

Since I hadn't had a chance to talk with him since he moved from Disney to take Microsoft's top IT job earlier this year, I was excited to get some time to get his thoughts on how he likes being a guinea pig for every new Microsoft product that comes down the line.

One of Scott's main missions--and that of his predecessors--is a process called "dogfooding" in which Microsoft becomes a major customer for every new product that comes down the pipeline.

Scott said he doesn't see that mission changing, but he did say he wants to shift that role so that Microsoft's experience, at least some of it, better reflects what its customers go through. Here's an edited transcript of our conversation:

Q: You've been in the job for a few months now. I'm curious to get your thoughts on what it's been like.
Tony Scott: It's been an interesting transition. I was CIO at Disney for three years before coming to Microsoft. I thought I would be there forever. So this all came about kind of suddenly.

The CIO at Microsoft really has three roles. One is to do all the classic IT stuff that probably every CIO at every company does. There's the role of working with our product groups where we do the "dogfooding." And then there's working with our customers because everyone who comes to Microsoft wants to know 'How does Microsoft do it?'

What I am trying to do is improve our world in all three areas. On the dogfood side, I think this is where maybe I bring some value as an outsider. I've been going to Microsoft for years...What I was always disappointed in was the relative degree to which Microsoft could talk to us as external CIOs about what the upgrade experience was like.

It turns out the reason why most of the former CIOs couldn't talk about that is that, internally, Microsoft used a very different process than what customers would use.

We never historically went from production bits to production bits in terms of the upgrade process. We went through a series of betas.

One of the changes I am trying to bring is--we'll still do all the dogfooding*; we'll still do all the betas--but we are going to take some segments of the company and use them to experience what customers experience and go through the normal upgrade process. I think by doing that we can be more relevant to the ultimate consumers of Microsoft's products.

*[Microsoft, Netscape] Interim software used internally for testing. “To eat one's own dogfood” (from which the slang noun derives) means to use the software one is developing, as part of one's everyday development environment (the phrase is used outside Microsoft and Netscape). The practice is normal in the Linux community and elsewhere, but the term ‘dogfood’ is seldom used as open-source betas tend to be quite tasty and nourishing. The idea is that developers who are using their own software will quickly learn what's missing or broken. Dogfood is typically not even of beta quality.


Everybody thinks Microsoft runs all on Microsoft and there is no other stuff there. That turns out just not to be true. We're predominately Microsoft but not exclusively. I want to take some of the learnings we have in those other environments...and make those experiences relevant to our customers. I'd say what I am trying to do is small changes like that. This isn't a sharp right turn in terms of IT strategy but it is trying to make what we do inside Microsoft more relevant to our customer base.

What was the biggest surprise when you actually got into the job versus what you expected?
Scott: I think the complexity of the company was startling. Disney was a very big company. I thought of Microsoft as a big software company and I expected it to be more like General Motors was or maybe even Sun when I used to work there. I didn't fully appreciate all of the complexity of the company--the breadth of products, the different ways in which the company markets itself, distributes, supports, consults...There was just a lot more complexity there than what I realized as an external customer.

Some of your predecessors also had to run Microsoft Online, selling services. You don't.
Scott: It's an interesting model. We run all of Microsoft Mail on the services we sell to customers. So, I have very few people in the IT organization that directly work with the mail, although it's a service we obviously provide to all Microsoft employees. Right now, the mail that we use is kind of an outsourced-but-insourced model. We're going to also launch soon the mail-in-the-cloud service which will be even more mail as a service versus mail as a hosted sort of thing. We'll put substantial numbers of Microsoft employees on that.

When will that happen?
Scott: We're doing it internally first. I think we have some limited customers on it now, even. But we have lots of billing things to create and lots of other systems to create before we roll it out as a big scalable service to end customers.

When you were at Disney or any of your past IT roles, what was your biggest pet peeve about Microsoft?
Scott: We hear from customers all the time, the complexity of our licensing models. It remains a big issue for, I think, most customers. There's just a lot of products there. It's just constantly a topic of discussion and I think that's something we have to just go tackle.

What are some of the big projects you are working on?
Scott: The big ones right now are a lot of investment in enterprise data warehouse strategy, also in what we think of as our core CRM (customer relationship management) underpinnings. Those two are probably the big bets right now. Emerging as a platform is PLM (product lifecycle management) or product management capability, which would include licensing simplification and some of those things.

Are things like mobile devices big? What are the kinds of things the average Microsoftie is clamoring for?
Scott: The pressure is to make things more mobile, more portable. I think you see it even surface in our products, Outlook Anywhere and Outlook Web Access--the degree of fidelity that you now have in terms of that mobile experience is just one indicator of the pressure that is on.

Also, I think when you think about where the next billion customers come from to Microsoft. What is it that they are going to buy? Increasingly it's likely to be gaming platform, phone platform...devices that look different maybe than what today's customers' (devices) look for.

What do you think is going to be taking up the bulk of your time when you look out?
Scott: It's probably three things. One is the continued simplification of our infrastructure and application portfolio. We have the legacy of richness of too many applications that are too fragmented across the company on a global basis. Simplification and consolidation onto globally scalable platforms--I think I will be doing for years and years--it's a big deal. The second is developing a set of apps that Microsoft will need two, three, four, five years out to engage in the businesses we want to engage in. We're in the middle of thinking through all of that right now. The third is just developing people.

You are here for an environmental conference, EcoForum. I'm curious what is going on in that area? I know power is a big issue
Scott: Most CIOs have come to recognize that both their employees and the customers of the company want to know that the company that they are either working for or buying products from is acting in an ecologically responsible way and that you take these issues seriously. From a Microsoft standpoint, we have some great products on virtualization. We're also here talking about that and here learning what other companies are doing.

In our own space we've gone from 8 percent to 25 percent virtualization in our data centers in just a year. Next year we think we are going to hit 50 percent. That's as dramatic a progress as I've seen, any company anywhere.

One of the things I am convinced of is that the entire technology community is going to have to come together to solve some of these issues. I came out of automotive. There was a day when if you wanted to know car gas mileage you had to write down the mileage, then drive and write down the mileage again. Then you went to the gas station and did long division to figure out what your gas mileage was. Eventually as the world got interested in this a chip got built in every car. Most cars have a chip built in to tell you what your miles per gallon is.

We don't have the functional equivalent to that in the IT world. As a CIO, you really want to know, what is this app costing me, all up? It's the people resources and the energy costs. The tools to do it are emerging but we are not there yet.

It shouldn't be that hard. If the technology community works together and develops the right standards and interfaces, one day you will be able to say here's my compute factor or my miles per gallon in terms of the technologies we use. With that we should be able to do a better job of managing our resources. I'm hopeful we could get that done.

沒有留言: