What is survivorship bias and why does it matter?

Survivorship bias refers to the tendency to focus on those who have survived a particular event or process, while ignoring those who have not. This can lead to an inaccurate understanding of what actually happened, as well as to suboptimal decision-making.

A famous example

Abraham Wald was a Hungarian mathematician who worked in statistical analysis in the 20th century. He helped improve bomber aircraft protection during WWII.

When bombers returned from missions, they’d often come home riddled with bullet holes. These bullet holes appeared to be concentrated on the wings and fuselage, almost twice as often as other spots on the plane. The military brass concluded that additional armour plating should be added in these areas. It seems obvious, right? These spots were clearly taking the most fire.

Wald disagreed.

Bullets holes weren’t found on other areas of the planes … because the aircraft that had been shot in those locations hadn’t made it back to base. Wald surmised that bullets were actually hitting every area of the planes. Which makes sense because it’s not easy to fire accurately at a moving plane. The massive damage on bomber fuselages and wings was evidence that these areas did not need reinforcing, as they had taken a beating and the plane had made it home.

Therefore, counterintuitively, Wald concluded that the planes should be reinforced in the areas that had received the least damage.

Business applications

In business, we see survivorship bias in the way media outlets often cover stories about successful companies more than unsuccessful ones. In our family life, people tend to remember examples that confirm their belief systems more easily than those which don’t. Think of your kid saying something like, “You always make me do the dishes, never my brother!”

This can lead to bad decision making. For example, the media makes a big deal out of ‘maverick’ business people, the ones who claim to have dropped out of school, or who didn’t have a business plan, or who made a seemingly gutsy move that bet the fortunes of the company on one big thing and paid off handsomely.

What the media doesn’t show you that often is all of those companies that blew up spectacularly when they did the same thing. This type of thinking can lead people to ignore important things such as risk management or budgeting in favour of trying to replicate what they see as a winning formula. 

For a more mundane example, consider a business called FastThink that is trying to decide whether to invest in a new software system. We’re always looking for the silver bullet, right? FastThink may look at other businesses that have already invested in similar products and see that many of them have been successful. They read the testimonials for the software: wow! Awesome results. Success is assured.

What they may not realize is that there are also many businesses that have tried and failed with this same product; their failures just aren’t as visible. Or that those companies that did succeed with the software had different parameters and circumstances; that is, what works for one company might not work for another for a variety of reasons. As a result, FastThink may make the decision to invest, only to find out later that the product is an absolute flop, and it doesn’t do what they need it to do.

The reality is that businesses fail all the time, and even the most successful businesses have had their share of close calls and near-misses. If we’re not careful, survivor bias can lead us to taking unnecessary risks that we would be better off avoiding.

Survivor bias can be difficult to avoid, but it’s important to be aware of its existence and try to take it into account when making decisions. Entrepreneurs must look at both successful and unsuccessful businesses AND their processes before making any major decisions. This way, they will have a better understanding of what is required for success and won’t overestimate their chances of succeeding with a new venture.

Image credit: Martin Grandjean, CC BY-SA 4.0

Leave a Reply