My Financial Times Column 10/20/10
When to turn a blind eye to the facts
By Philip Delves Broughton
The good news from the Gulf of Mexico is that the worst predictions of environmental disaster after the Deepwater Horizon oil rig exploded are unlikely to be fulfilled. It’s still awful, but not as awful as first feared. But when you think back to the worst days, as BP struggled to cap the well against a cacophony of abuse, there were two kinds of decision making on display.
For the men and women fighting to stop the leak, there were the messy, practical challenges to be met, and decisions to be made based on the evidence. And then there were those who made the decision early to pound away at this latest example of “Big Oil” gone wild.
President Barack Obama was caught between the two, trying to make decisions based on the complex set of facts, while others claimed to know exactly what decisions should be made, and compiled the evidence accordingly.
Businesses are often caught in the same trap. Ideally, you want to base your decisions on sound evidence. But often, managers make a decision then rustle up the evidence to support it. As Peter Tingling and Michael Brydon of Simon Fraser University wrote recently in the MIT Sloan Management Review, it is the difference between evidence-based decision making and its ugly sibling, decision-based evidence making.
British schoolchildren are taught the story of Lord Nelson at the Battle of Copenhagen. When told that his commander was signalling for him to retreat, Nelson raised a telescope to his blind eye and said: “I really do not see the signal.”
Within a few hours, the Danish fleet was defeated. He had made the decision to fight and moulded the evidence to fit. “If you get the outcome you want, then everything is fine,” says Prof Tingling.
But in a world that venerates data-driven decision making, whether in financial services or sports management, managers are often loath to admit that they still take the Nelson approach to make the decision first and find the evidence later.
Profs Tingling and Brydon found that evidence is used by managers in three different ways: to make; inform; or support a decision. If it is used to make a decision, it means the decision arises directly from the evidence. If it is used to inform a decision, evidence is mixed in with intuition or bargaining to lead to a decision. If it is used to support a decision, it means the evidence is simply a means to justify a decision already made. They also found that evidence is often shaped by subordinates to meet what they perceive to be the expectations of their bosses.
There are two dangers to letting decisions trump evidence. The first is when decision making is simply ill-informed. Ideally, a decision that contradicts the evidence is an inspired hunch, formed by experience, like Nelson’s. In the worst case, it is the product of ignorant bias.
The second danger is that once your employees know that you, as a manager, are more interested in finding evidence to fit your conclusions rather than seeking out truth, it infects a company with demoralising and destructive cynicism.
When Herman Miller designed the Aeron chair, consumer focus groups were hostile. But the company ignored them and went ahead with production anyway. The chair was a huge hit. The company was capable of taking evidence and successfully ignoring it.
In other companies, however, the cult of data-driven decision making leaves so little room for personal beliefs that people just tailor evidence to fit pre-made decisions.
At Ford in the 1950s, Robert McNamara, then chief executive, demanded data on everything. Interns would cut up newspapers and paste them into binders so executives could point to the voluminous research that went into each decision.
Vince Kaminski, who led Enron’s research division, has spoken of the frustrations of having 50 highly skilled mathematicians picking apart Enron’s risky deals only to be ignored by executives who prized volume above all else.
So what is a manager to do? How do you encourage the use of data, while leaving room for the occasional inspired decision? One solution is to be more flexible in how you categorise decisions. Not all will require the same degree of evidence.
Another is to weigh the costs of gathering evidence. Is it always worth it? If not, don’t fudge it for appearance’s sake. Admit that you are trusting your well-honed instincts. This is especially true for those within your company. Sometimes you will have to come up with tendentious evidence for an external audience, which demands at least a charade of evidence. But don’t ever pretend for those inside. They will know better and punish the slightest deceit either now or well into the future.