Is your e-commerce brand working against “survivorship bias”?

Whether you’ve heard of the term “survivorship bias” or not, the chances are your e-commerce brand is working against it. Most brands are. 

Survivorship bias is a form of selection bias. It can lead us to make assumptions about our entire audience based only on the data of those that ‘survive’, or pass a certain selection process. In the case of e-commerce, our ‘survivors’ are those who have made a purchase. 

The origins of the term “survivorship bias”

The term “survivorship bias” was coined by Abraham Wald. He was brought in to advise on a project to reduce the number of World War 2 plans that were fatally shot down. 

The idea was to reinforce the most vulnerable parts of the plane to reduce the number that were lost. But because this reinforcement would make the plane heavier, they needed to be selective with the areas they chose. 

Until Wald stepped in to advise, the army were operating with survivorship bias; they were reinforcing the parts of the plane that had, on average, been shot the most across all the returning planes. Instead, Wald recommended reinforcing the areas that hadn’t been hit. He deduced that the planes that had been fatally hit were probably shot in these areas, and therefore these were the points most worth protecting. 

His input is thought to have saved hundreds of lives.

How e-commerce brands are impacted by survivorship bias

E-commerce may not be a matter of life or death, but that doesn’t mean we’re not susceptible to survivorship bias.

Once a customer has made a purchase, we have so much more data available on them. 

At a minimum, we can see: 

  • Where they live
  • How they prefer to pay
  • Their name 
  • Their email domain

And that’s just through Shopify’s own data; it’s not difficult to access more with additional tools and platforms. 

But before they make a purchase? It’s a different story. 

You can see the number of people who reach each conversion milestone:

  • Session
  • Added to cart
  • Reached checkout

But not who drops off at each stage, or why. 

And this can lead us to make decisions that further optimise our site for our existing customers, and ignore the vast majority that never convert – which is likely where the real pain points are hiding. With a 3% conversion rate considered ‘good’, that’s 97% of your potential customers unserved.

How to avoid working with survivorship bias

Avoiding survivorship bias ultimately starts with having better data available. The issue for most smaller e-commerce businesses is that the tools providing this data are often prohibitively expensive. 

FERO was designed to help plug the gap for smaller brands. We combine over 430 data points to give you an accurate picture of your full audience, and to suggest optimisations you can make for each shopper. 

As our founder, Craig Savage points out: 

“Imagine if 7 out of every 10 people who came to the checkout in a physical shop just turned around, put their items down, and left. The shop owners would be falling over themselves to find out what was wrong, and put it right.” 

In our experience, this is low hanging fruit in almost every e-commerce brand. Try FERO free for 30 days.