The CTO In A GTO
Don't like your car? In the future, says General Motors's Tony Scott, you'll just download a new one.
By John Battelle; Tony Scott

(Business 2.0) – Tony Scott has a simple message for people who make hardware and software: Listen to your customers or risk losing them. As both carrot and stick, General Motors's affable chief technology officer wields an annual IT budget of nearly $3 billion. It's that kind of spending power that led many to credit Scott with prompting the recent détente between Sun and Microsoft.

Some in Silicon Valley fear that the area is turning into the next Detroit. (Sun Microsystems CEO Scott McNealy and Microsoft CEO Steve Ballmer grew up there and saw the decline firsthand.) But turning into Detroit could be a good thing, Scott argues: Remember that the auto industry pioneered interoperability, industry standards, and--egad--warranties. The two industries are a lot closer to merging than you might think. Software and electronics already account for a third of the cost of the average car, and given GM's billion-dollar investment in the Hy-wire, a hydrogen-powered concept car (see "GM's Race to the Future," October), high-tech and hot rods are only going to become more intimately entwined. How so? Read on.

I hear you're entirely to blame for making peace between McNealy and Ballmer.

[Laughter] It's really interesting the way urban legend evolves. We never said, "Sit down and go smoke a peace pipe." But GM has spoken with a very loud voice to Microsoft and Sun and others about interoperability. We've been diligent about making our pain known in that space--it's very costly to make incompatible technologies work well together.

Can you give me a specific example?

We use Microsoft Active Directory to let employees sign on to their PCs. That has to interface with LDAP [an Internet directory protocol], which we use for applications in every other environment. In the past, that's required a lot of custom integration. And every time there's a new release--of the Microsoft environment or LDAP or our own applications--we have to go solve that problem again. It's the same for everyone else: Ford has to go solve that problem, Procter & Gamble has to go solve that problem. It's been the full employment act for contractors and consultants. It's silly. These things should just work together.

Why don't they? Is it that each installation is unique and requires customization and ongoing service?

I think a lot of those customized requirement claims are hogwash. I don't believe it. Now, I don't have a problem with having a service relationship. We have a very profitable service business in the automobile industry. But it's very different from where the software industry is today. Software has no standards for quality and no standards for interoperability. The industry hasn't had to do it--the customers didn't demand it and the market conditions were such that there was no reason for them to do so.

And that's changing?

Yes. I participated in a trip to our major IT vendors earlier this year. And for the first time, instead of pitching us about some new capability or feature, every single one of them said to us, "Here are the things that we have done as a company to make our products cheaper to install, cheaper to operate, and easier to support."

Is this a sign that the IT industry is maturing?

It's maturation, standardization, and commoditization. The industry is going through a transformation--one the auto industry and railroads and other big industries have already been through. I'll make a car analogy. Imagine if GM shipped you a car that you had to assemble yourself, and mechanics had to come to your house just to put it together, and it had no warranty. I think somebody would come along pretty quickly and say, "We've got a better idea on how to make and sell cars."

When you go to a gas station now, you can stick the nozzle into the gas tank and it works. But in the early days of the auto industry, there were 2,000 car companies, nothing was standardized, and demand far exceeded supply. In that early era, you could do whatever you wanted. The tech industry has by and large been in that same mode. But that's now changing.

People in the Valley are terrified of the idea that Silicon Valley might turn into Detroit. In that scenario, innovation and growth slow, and margins look more like--well, GM's margins. The whole culture here has been built on startups and innovation.

Well, let me correct a few misperceptions. Yes, Silicon Valley probably will become more like Detroit, but it's not the grim picture that many would paint. Automobile innovation is very profitable, but insulated--there's a tremendous amount of testing and quality that gets built in before it ends up facing the customer. It shows up in a high-quality, highly integrated, supported manner. I think that's the model for what's going to happen in the tech industry.

Can you name a major IT project that has dramatically improved GM's business?

Yes--vehicle design. In 1996 our vehicle development process took in excess of 48 months. IT played a huge role in reducing that. We're down to about 18 months now--and when you're in that range, you can not only follow fashion but actually lead. That's huge.

You have a strategy that calls for completely outsourcing your IT work. Why?

It enables speed. If we want to go do something new and innovative in the IT space, all we have to do is put out an RFP. Then we hire the best minds to do it. We don't have to worry about an internal staff of programmers who may not have the right skill sets, or who have other demands and can't respond to what we need. Very few internal organizations can muster the sort of resources and global deployment capability that's needed to do what we need to do. Especially in cases where you need those resources for short periods of times. You can't build an internal staff in Argentina, say, for a six-week project.

So your IT organization is basically mostly management?

Yes. It's 1,700 or so people who manage the contracts with the outsource vendors all around the world.

And if you were to do it in-house?

Oh, it'd easily be times 100 or more.

You've streamlined IT at headquarters. What's happening in the cars themselves?

Today in the typical car, and it doesn't matter whether it's a GM or a Honda, electronics and software represent roughly about a third of the cost of the car, more than the labor or the steel. They're the single biggest cost in a car today--and rising. The power train, for example, is very highly computerized. There's a set of very finely tuned algorithms to meet emission control and mileage standards.

With all this technology, are we losing the ability to "get under the hood" and customize our cars?

I think there are some interesting possibilities. Let's look to the future. We've put more than $1 billion into our fuel-cell program, and we think by 2010 there'll be fuel-cell vehicles that you can buy. These cars are pretty exciting from a user tweakability perspective.

How so?

Their inherent design opens up possibilities. In these fuel-cell vehicles, you don't have a traditional power train, a big engine and transmission up front. The chassis is very thin, maybe 8 inches--think of it as a magic carpet. The things you can do with the interior of a vehicle, the layout and display, change the whole user experience, and that experience can be customized.

So is the gasoline car destined to be the vinyl record of the auto industry--ultimately abandoned to the hobbyist market?

We had some of the fuel-cell vehicles out on our test track, and I walked up to a guy who had just gotten out of one. I asked how he liked it. He said, "Oh, it really was fantastic." So I asked him if he'd buy one. He scowled and said, "I'm never driving a car that sounds like that." What he missed was that throaty exhaust sound. His mental model of what a car should be was that it should make a certain noise. When you step on the gas, you should hear the engine, you should feel the vibrations ...

Well, you have room for a great sound system, so why not the auto equivalent of a laugh track?

Exactly. People are going to want feedback that's more analogous to what they're used to. Maybe in one of these future vehicles, you dial in a 1965 GTO look, feel, and sound. In the future, if I ask, "What do I want my car to sound like?"--well, I could download it.