Global CIO: A Global Project, With No Big Project PlanGlobal CIO: A Global Project, With No Big Project Plan

As I talked with Eric Choi about a project he led for Experian, one of the big three companies in credit reports, the words "leap of faith" kept coming up.

Mind you, that's a lot different than a "wing and a prayer." Choi, the company's e-commerce director, thinks an agile, relationship-focused style is the least risky way to do a global Web development project. But it took faith that this approach--with minimalist executive dashboards, ever-shifting priorities, and iterative learn-as-you-go development--would deliver.

The project: Get more than 100 Web sites, run by 40 different Experian offices worldwide, onto one Interwoven content management system. The deadline: April 2010, one year from the initial planning.

The top reason for doing the project was to get Experian to show up higher on Google searches. Rather than looking like a global online enterprise, it looked like dozens of smaller operations, since each country had its own site. Each country also put its own brand variations on their sites and duplicated the costs of maintaining a site.

When Choi communicated with top executives about the project status, he didn't show a lot of data on interim milestones. He showed them a simple heat map, with a black box representing each country that was done, a gray box for those in progress, and a white one for those not started. There was no step-by-step waterfall plan, of this country by X date and the next by Y date. Those things change once teams start work and learn the real problems, Choi says.

"Project plans and a waterfall approach scare everybody, but everyone expects it," Choi says. "One of the risks is the project plan manages you."

Experian wanted a central Web platform but still wanted local managers to create the content, taking advantage of their understanding of the language and market. In the changeover, the local teams--aided by Choi's project team--had to load content into that central system.

The project team lived with a lot of ambiguity. As the team started migrating content in Japan, for example, it became clear that country would be one of the more difficult ones. So the team switched to focus on other countries, while keeping just a small team to crack Japan's problems.

Choi calls it "leading with our successes." It let the team knock off easy countries--and color some boxes black on that heat map the executives saw. It finished 25 countries in the first four months.

But how did Choi know he wasn't papering over a deadline-busting problem, just to look good on a heat map? "As we iterated with the smaller countries, we were learning what to do and not do," Choi says. Plus, success can help bulldoze some problems that seem intractable in the early project stages. Everyone sees that heat map, and no one wants to be the last country still in white. It created some useful internal competition.

Choi's team needed to built relationships in every country--with the country managers, and also with the IT teams, since every one was run differently. The teams had to convince country managers "we're going to take away their Web sites, only to give it back to you to control. That's a tough message," he says. Each country had clear milestones, such as loading the content and training people on creating new content. Choi and his team relied heavily on face-to-face meetings--over teleconferences--to build those ties and work through problems.