Have you studied cognitive biases in hope of refining your own decision-making and thinking. I have. But, it’s been quite useless for improving my own thinking. That’s disappointing, even if it is now easier for me to see bias in other people, organizations, politicians, and random jerks who annoy me.
It’s not a side of myself I really want to indulge!
Firestone
In the 1990s, Firestone tires were standard on Ford Explorers. But in 1996 the State of Arizona reported that tires were separating during high temperature use. The CEO, John Lampe, sent engineers to Arizona.
The engineers’ report blamed the separations on “off-road use and maximal loads”. As a gesture of good faith, Firestone provided heavier duty replacement tires.
Later that year, a Firestone engineer voiced concern about similar separation issues. No action was taken. In 1998, someone contacted a State Farm Insurance employee named Samuel Boyden about multiple crashes caused by tire separations on Explorers. Over the subsequent 18 months Boyden found 51 similar crashes. He sent his data to NHTSA (the National Highway Traffic Safety Administration).
There were also reports from Saudi Arabia, Kuwait and Qatar. And in 2000, a Houston television station ran an expose about the issue. Suddenly, Firestone and NHTSA were flooded with calls.
4 years and 238 deaths after the first spate of crashes, Firestone acknowledged the problem. In August of 2000, they recalled 6.5 million tires, costing Bridgestone (its owner) half its value.
Why?
Firestone executives exhibited a trifecta of cognitive bias. Their years of experience led them to over-confidence. Their quest to find other potential explanations was riddled with confirmation bias. And finally, the entire executive team reinforced each other’s Groupthink, allowing no dissent from the party line. They simply could not see their own poor judgment.
Wait. I Hear Hooves!
We’re wired to be blind to our own bias. Millennia ago, action paid greater survival dividends than nuance.
If a Neanderthal leader suspected a stampeding herd of mammoths ahead, he stopped the expedition—despite protests from his clan. If he was wrong (and his fellow clansman right) then they lost out on slaying an antelope. But if he was right, the whole clan lived to hunt again. We inherited that instinct to double down on our gut feeling.
Of course, today, when we dig in our heels it’s not a mammoth’s tusks we avoid. The stakes have changed, but the over-confidence remains.
We are all—regardless of intelligence or knowledge– subject to these misjudgments. In fact, the smarter you are, the better you may be at defending your biased conclusions.
Bad News Bad News?
Most business leaders feel like they could never be as wrongheaded as Firestone, Blockbuster, or Kodak’s failed teams. We are more rational. More data driven. But the steady reports of new failures contradicts that.
You’re unlikely to see bias in yourself. That is also true for your employees.
This must seem pessimistic. Is there any good news — any way to reduce bias in our enterprises?
There is.
But better organizational decisions won’t emerge from studying bias or experiencing anti-bias training. Even the greatest behavioral economists claim they have not been able to reduce their own bias –or catch themselves in the act.
Instead, we need structures and processes to mitigate bias. That entails multi-faceted, multi-subject approaches to how we work.
At last count, there were 180 described cognitive biases. There is no way to structure your way out of all of them. But you can reduce the most frequent ones. Anthropic’s Claude helped me whittle down the field to the top three.
Optimism Bias, Confirmation Bias (which include over-confidence) and Groupthink. Given the constraints of a single essay, I’m only going to touch on two of those.
That Timeline is Nuts!
Most errors aren’t fatal. But they slow progress, reduce reliability, and create stress: projects running over time and budget, product decisions made in the absence of market data, untested strategy, ignored warnings.
Let’s talk first about optimism bias because it is the most insidious.
I work with lots of clients on learning how to use time blocking to achieve their goals. It’s a simple concept: schedule everything you intend to do –with real durations. When they try it, 100% of them report back that it was awesome—BUT, that everything took longer than they scheduled. Optimism Bias. We underestimate how long things take—as well as how complex they will be.
Knowing that doesn’t help. The bias always wins. But a simple heuristic will make a difference.
Apply a multiplier to every phase of every project.
Start with 100%.
That may sound crazy but trust me—it will be closer to that than you imagine.
You can refine this with machine learning if you use or build it in your enterprise. Using past data, an AI can create algorithms that will make more accurate estimates than humans.
Be forewarned: You will be sure the estimates are way, way, off. Suppress the urge to meddle.
Multiplying every duration by 2 will singlehandedly increase reliability, employee wellbeing and quality by orders of magnitude.
Groupthink
Everyone thinks they are immune to Groupthink. But, multiple studies show we are not. It isn’t that we consciously edit ourselves. But we are constantly calibrating our perspectives with whatever input we get. Much of that input comes from our peers and leaders, and so we are subtly persuaded through a combination of arguments and social pressure.
All of that—plus a survival instinct that may curtail disagreeing with the boss– creates the phenomenon of Groupthink. It is ubiquitous. Whether you’re discussing strategy, the roadmap, or where to eat lunch, dissenting views will quickly evaporate into magical unanimity.
There are some very easy (and fun) ways to reduce Groupthink and get a greater diversity of feedback. The basic tenet is: Don’t solicit ideas in a group. Here’s what that means in specific scenarios:
Brainstorming: In a group, have everyone write down all their ideas individually for the first 15 minutes. Or, before the meeting via email. Collect all of them and vet them as a group (anonymously).
Problem-Solving: Send emails in advance explaining the problem set and requesting solutions. Ask people to think about it on their own without conferring, and to come to the meeting with at least 3 ideas. Start with those ideas and build.
Strategy Meetings: Red Team the idea. In other words, have a specific group of team members whose job is to argue against the strategy. They can use every tool available: data, market analysis, scenarios, history, etc. Then, appoint a new group to be the Red Team and have them do the same thing –argue against the idea. If the idea survives the barrage of opposition, it passes the test. This process uproots hidden Groupthink by gamifying the dissent.
Many brains are better than one. But only if you can keep them separated from each other’s influence. Otherwise, you will receive fewer and poorer quality ideas.
There’s a lot to tackle to address Confirmation Bias or Over-Confidence. I promise I will do some of that soon. In the meantime, this is a good start.
The bottom line is that structures work to reduce bias. Learning and training doesn’t. So, if you love investigating cognitive biases and find it fascinating, carry on! But do so knowing that it won’t help you think better.
