If Only BP Knew Now What it Knew Then

Posted on Harvard Business Review: August 2, 2010 9:35 AM

The new BP CEO, Bob Dudley, told ABC that the Gulf oil spill "has come out of nowhere" for the company. Maybe not. Perhaps it was coming for quite a while.

In the mid-nineties BP was all the rage in management circles, in part because it was one of the very first firms to take the notion of "knowledge management" seriously. John Browne, its then-new managing director, was a very smart and experienced executive who seemed to "get" knowledge. We had interviewed him for a project and found him both enthusiastic and alert about what changes might allow his people to make better decisions based on insight and experience BP had already gained, and to keep building those stores. The company was also exploring multiple new technologies (including videoconferencing) to give managers more access to resident expertise. Two of its knowledge managers wrote a book to share BP's success story and help others "learn to fly."

However, by early 2000 most of these well-publicized efforts were in tatters. The firm was focusing more and more on controlling costs and boosting its share price (it had a severe case of "Exxon Envy"), and its prized "peer assist" program—the jewel of its knowledge program—was barely functioning. By the time of the Texas City refinery explosion in 2005, the Toledo refinery explosion in 2006, and the Alaskan pipeline rupture in 2006, knowledge management was a relic. The cultural byword had become "full steam ahead" (by 2009, it was "every dollar counts, every seat counts") and the emphasis was off organizational learning.

A new "operating management system" [pdf] was implemented in 2007, supposedly to prevent such disasters from happening. Its rigid structure specified 8 key elements and 47 sub-elements. In setting new judgment parameters for managers, this system stressed compliance. But BP also created higher incentives to find shortcuts and take risks. Record profits were posted soon after, perhaps confirming BP's new direction.

We are not saying (we'd like to, but we can't prove it) that there is a causal relationship between the dismantling of a knowledge management program and the subsequent missteps that culminated in the Gulf disaster. However, it seems likely—in at least three ways—that good judgment was distorted and threatened in the decision-making climate BP subsequently developed.

By taking apart a successful program and not replacing it—or by replacing it in part by a system—a message was sent about human judgment versus rote processes, checkboxes, and rules.
A series of lesser disasters with no systemic and widely understood learning taking place enabled more and greater disasters. The lanes were widened for what risks were permissible without punishment; rewards were primarily for cost-cutting and not withdrawn for risky behaviors.
A "climate of opinion" was established, invisible but as strong as iron, that was deleterious to debate, discussion, and collaborative decision-making. The parameters of decision-making were set up to reflect the emphasis on short-term costs and benefits.

Dudley claims that BP will learn a lot from the Gulf disaster. We hope that's true, but there will have to be drastic changes in culture for that to happen. Maybe it's time to bring back knowledge.

Tom Davenport, Larry Prusak, and Brook Manville are at work on a book on judgment and how to cultivate it as an organizational, not just individual, strength. Their posts in this blog will test-drive ideas and invite input as the research progresses.

Before it's here, it's on the Bloomberg Terminal. LEARN MORE