Closing the Gap – Corporate E-Commerce Development

There has always been a compromise in IT between the need for speed and the requirement of disciplined development to ensure that systems will integrate and interact with the established infrastructure. In the rush-to-implement era of e-commerce, however, it’s become a dichotomy. And it epitomizes IT/business misalignment—the next generation. E-commerce systems get built by a separate team of speedy gonzalezes, while the IT people who carefully architected the core business technology stand on the sidelines shaking their heads in disapproval. If the CIOs and their e-commerce counterparts can’t bridge this gap, the consequences could be more than a loss of turf and prestige for the IT department. It could mean a technological bifurcation of e-business and core business technologies that would be crippling for any company. We asked an e-commerce leader and an IT chief how they’d keep the rift from spreading. The following dialog features John Rosenfeld, e-commerce leader at GE Aircraft Engines in Cincinnati, and Steve Schuckenbrock, former senior vice president of IT for PepsiCo in Dallas and now managing partner of client services at the Feld Group, a technology and management services company in Dallas. Tom Davenport, CIO columnist and director of the Accenture Institute for Strategic Change and a distinguished scholar at Babson College, moderated the discussion.

Tom Davenport: In some organizations today, e-commerce is a separate function from the CIO’s organization. Is it going to come back to the CIO at some point, or is the cow out of the barn forever?

John Rosenfeld: We’re in the middle of this right now. GE did something very strategic by kicking off e-business as a separate initiative. Six Sigma was another huge initiative that GE kicked off this way. At GE we handpicked some of the best folks we could find, set up an entirely new organization and said, “Go do this. Do whatever you have to do.” The team was provided with just about anything it needed, based on the huge potential returns. It was a lot like a very well funded startup environment.

That initiative approach allowed us to achieve incredible momentum and speed. We launched the Customer Web Center that included more than 6 million pages of content, order capability, customer collaboration tools, warranty management and more. I could go on for an hour about what we launched on January 1, but the thing that’s most amazing is not the breadth of functionality; it’s the fact that the team built the entire thing in 58 days. We picked vendors in nine days instead of several months like we typically do, and that was all because of the [separate] nature of the initiative. But it’s important to recognize that we didn’t sacrifice our policies or processes, we just moved through them at light speed.

Now, GE is making the transition of taking this initiative and pulling it in and making it part of our business. My former boss was the e-business leader, and we also had a CIO. These were two very separate jobs. Recently our CIO left, my boss went to lead e-business at NBC, and we established a vice president position that now runs IT and e-business. If you looked at all the GE businesses, that’s already happened in many of them.

Davenport: Certainly you think if e-commerce and IT don’t get put together, it’s kind of a bad sign for the CIO, right?

Rosenfeld: Right, exactly. I think the big challenge we face is that we’ve got two different types of organizations that are now brought together. E-business in our company is based on the excitement, the energy, the disregard for bureaucracy and the ability to make change incredibly fast. But there is considerable discipline, standardization and analytical thought processes that exist in our IT organization, which most of these [e-commerce] guys aren’t familiar with. How do you take those two and make the best of both worlds? When we built things, the first thing on our minds was speed. And although we wanted to do things smart, standard processes and disciplined procedures were an afterthought compared to speed of delivery. That led to legacy apps that don’t know how to talk the same language.

Steve Schuckenbrock: I think everything you just said is exactly right. I do think e-commerce and IT belong in one group; the gene pool of that group is an IT gene pool. But it also incorporates some of the characteristics that this separate e-commerce organization brings to it, which is speed. Speed is one of the critical issues facing IT today. You originally cut IT out of the process in order to create speed, innovation and thought leadership. Now folding it together into one thing is essential. “E” is not going to mean much unless it and your core IT business are the same thing. So separating your IT organization (and your core business’s technology) from your e-business efforts is basically a formula for separating your “e” and your business, and I think that’s a failure formula for the company.

Davenport: John, what did you put in place to ensure that you would have an easy reorganization?

Rosenfeld: Absolutely nothing. When we launched the e-business organization, we had to draw on the talents and skills of the IT organization because most of the new e-business employees didn’t know how to integrate with our legacy apps. The IT staff were excited to interact with us and be part of the energy, if you will. They loved it because we didn’t impose on them the typical discipline that they were required to follow in their existing jobs. Now our challenge is that our team needs to develop the level of discipline needed to manage and develop systems for the long term. So we’re starting to learn that and to become more of a hardened organization. It’s not easy-it’s a big transition to pull these groups together. It’s almost like pulling parachutists and scuba divers together and telling them, “Go in the same room and come out when you can both do either thing.” It’s challenging the very different thought processes.

Davenport: Was there any one person held accountable in this whole process during those 58 days?

Rosenfeld: I was primarily on the hook as the e-commerce leader along with my boss, Dave Overbeeke, who led all of e-business. And of course, the CEO was being measured as well by Jack Welch. But that was fun. I liked the challenge.

One more thought on the whole discipline and trying to merge these organizations: We called our approach “launch and learn.” For the e-business group it was: Let’s build something, put it out there in front of the customers and have the customers help us figure out how to tweak it to make it what they really want. The typical IT mentality before that was, nothing goes out this door until it’s perfect.

Davenport: What was fun? You mentioned that it gave the IT staff a chance to work in ways that they weren’t allowed to work in the regular organization? What do you mean by that?

Rosenfeld: They could take more risks. They didn’t have to go through four weeks of disruptive testing on every program, so they could try something and see the results immediately. The cost associated with this testing and integration with legacy apps is huge. Let’s say, for example, the company requires that they work on a Unix platform. We tell them, “To get it done in three weeks, tell me what you have to do.” They come back saying, “We’re going to have to do it on NT. But we don’t have an NT infrastructure.” I say, “I don’t care, we’ll go get one. You go and build it on NT, and if it works, we’ll install an NT infrastructure.” That was an opportunity for them to do something they wanted to do without being limited by our existing disciplines, if you will.

Schuckenbrock: Let me throw the wet blanket on that approach. The issue with that is it’s very much like the departmental computing of the past. Data fragmentation, losing control of who your customers are, losing perspective on what ways and from what points a customer has access to the Web. Speed is wonderful, especially in a green field. And it is wonderful in terms of testing and innovating and driving what you in fact delivered quite successfully. In the end though, you’re trying to run the holistic business in the most efficient, highest customer service way, but you’ve got all those fragmentation points—you’ve got your customer definition over here on an NT SQL server, and over there on a Unix Oracle server, and over here on a Sybase server. Which ones are right? Which are the most current? These are real problems that come out of this process, and they have to be solved because the ultimate objective that Jack Welch has is not for GE to be an e-business, it’s to have the highest customer service, lowest cost business that he can possibly run.

Rosenfeld: Those are some of the new challenges we have recently attacked—kind of cleaning up the mess left behind our wake.

Davenport: Is it the same group of people who bring it together, do you think? Can you take this team that has gotten used to basically gunning it and going and doing all this fun stuff and then all of a sudden say, “Great! You’ve done all this wonderful development work. Good job, guys! Now we’ve got to clean it all up and turn it into a real IT organization.” Or, do they all quit and then you have to bring in some other people?

Schuckenbrock: Whether people quit or not I don’t think is the point. The point is we’ve got to run our business, and there’s a certain way we need to behave in order to run our business effectively. Now, I wouldn’t attack the people who did it because they did exactly what they were asked to do, and they did it wonderfully. Instead, I’d go to this new vice president of e-business and IT and have him task the group with delivering integrated quality and not sacrificing speed. He will come back with a very difficult set of conversations around the trade-offs that need to be made in order to achieve those two counter-balancing objectives. And then the senior management team will engage and make good business decisions. You’ll still need speedy programmers, but complimented by people who’ll decide that this is where you’ll get your customer data; you should quit trying to invent that in three other places just because that’s faster for you.

Rosenfeld: I think Steve’s dead on, and even when we were launching this 58-day project, we did some things slowly because we made an assessment: “Should we take this shortcut here that’s going to take us three days to complete? Or, should we do it the ideal way in three months?” We did make some decisions to do it in the ideal way and in other instances we said, “We’re going to do it, take a shortcut here and fix it later.”

So I think to Steve’s point, the people are going to still take risks and shortcuts to accomplish certain business goals, but the leadership of the new regime is more tech-savvy, and so they’re going to be a little better informed and a little more knowledgeable about what those risks and shortcuts are and what their impact will be.

Davenport: I agree—a lot of leaders coming up now are more tech-savvy, but they’re not tech-savvy about things that are the order of the day. They’re not tech-savvy, necessarily, about data integration or standardization or back-office transactions; they’re tech-savvy about webpages and glitzy front-ends. Are we going to have a mismatch there in what they know and appreciate versus what’s needed?

Schuckenbrock: Let’s not swing so far to tech-savvy that we require them to care about Java and integration and databases and all that other junk. What they care about is the business. What they’ve learned to appreciate is how business is so hard-wired into the technology. They don’t have to be real tech-savvy to understand that, say, we have a horrible view of our customers’ experience because the view is hard-coded into four different databases that have four different definitions of the customer as opposed to one definition that feeds four different systems. The savvy ones say, “How the hell did we do that?” And you say, “Well, don’t you remember when we built it and cut a few corners? Those corners turned out to have big business implications. You just didn’t know it. Now it’s time to start paying attention.”

Davenport: Are they going to be happy though, paying multiple hundreds of millions of dollars for problems like that and the five years to integrate it all? I think we’re very impatient now with our technology desires.

Schuckenbrock: The answer is no. Nobody likes to pay hundreds of millions of dollars, and nobody likes to wait five years. There’s no reason to wait five years for some of these things. I do think there is a very difficult set of conversations at the senior management level where you pick priorities. If you are an airline, your priority can be your baggage handling system or your ground control systems or it could be your customer experience. There’s not a good or bad decision in that, but there is a senior management decision needed that says, “We have a major reengineering piece of work to do, and we’re going to pick the priorities for how we do it.”

I see too many companies fail by trying to do five or six things that are massive transformations simultaneously. What happens is they wake up one day, realize the problem with their current systems and make the gut decision to put a couple hundred million dollars up against it and then launch everything at the same time. These knee-jerk attempts at a silver bullet rarely succeed.