Data – the way to overcome your cognitive biases
In classic economics, humans are often described as ‘rational actors’ who always analyze risks and costs, and make the optimal decisions. But as a sports bettor, I have come to recognize that nothing could be further from the truth. Following sports, especially football in my case, can be emotional and involves many ups and downs in the course of a game, let alone an entire season. When we combine this with betting, which is nothing more than an investment of money, weighing up risks and rewards, we can discover a number of human cognitive biases. And these biases are one reason for the typical non-rational, emotional human behaviour – not only in betting, but also in business.
The confirmation bias
One of the most common cognitive biases is the confirmation bias. This bias plays a significant role in sports betting. People often have a pre-conceived opinion about a specific match they want to bet on. So they look for information to support their hypothesis. Information that may disprove the hypothesis is just ignored or not given enough consideration.
Let’s take a football example. The season 2018/19 of the German Bundesliga was one of most thrilling for years. Halfway through the season, Borussia Dortmund sat at the top of the table with a six-point lead. It looked as if Bayern Munich’s dominance could be – finally – broken.
Arguments and counterarguments came thick and fast at this point, making the case for and against Bayern Munich defending the title:
- They won the championship seven times in a row
- They suffered many injuries this season
- Their coach was criticized for the second year in a row because of the team underperforming from time to time
- Their strongest opponent Borussia Dortmund’s new coach was renowned for getting teams to perform better than expected
So how might different people have interpreted this information? Bayern supporters would surely have ignored the last three points. In their opinion, Bayern Munich is the strongest team and nothing will change their mind.
Supporters of Borussia Dortmund, of course, would have a different view. They might have overvalued the information that indicated Bayern Munich’s title-winning years could be over. But who was proven right in the end? The Bayern supporters. And do I need to mention that they told everybody that they had predicted this right from the start of the season? That’s the so-called hindsight bias – but that’s a story for another blog.
You need to investigate every argument for and against to reach an objective opinion – regardless of which team you support. Sports bettors, in particular, need to remove emotion from their decision and become a true ‘rational actor’. That’s what I realized over and over again and what can only be achieved by trusting the numbers. I use models to predict the outcome of a football match. They’re based on the same information for both teams so don’t favour a side. If the model is accurate enough, I’ve got a method to estimate the outcome of a football match without any subjective influences. I can then place my bets based on this information. My decision is data-driven.
Data-driven decision-making
Decision-making in business is also influenced by confirmation biases. Investors who hear rumours about a bankruptcy overvalue any news supporting this opinion – even if the company developed a new product that may help their business recover. A product manager knows about the ‘next big thing’ and uses market research to prove his point, but not to disprove it. People mainly seek confirmation of what they already believe.
To avoid this trap, Warren Buffett follows two simple rules:
- Be aware of the danger of confirmation bias, and acknowledge that our judgment can be clouded by it.
- Aggressively seek out and understand information that disagrees with your existing belief.
Rule one requires personal awareness of each decision maker. Be it the CEO of a company, who defines the new company strategy, or a salesperson, who has to decide which opportunity to tackle next. They have to understand that each decision can be biased.
The second point can be achieved in the same way I make my decisions in sports betting – data driven. Business technology has advanced a lot in recent years. With technology supporting almost every aspect of a business, the data provided by each system can be collected, processed and provided to each decision maker. Data helps you understand what’s going on in your organization.
Follow these tips to achieve data-driven decision-making in your company:
- Implement a data culture
Everybody in your company should be encouraged to use data. Employees shouldn’t be scared to use business intelligence (BI) tools, check the latest KPIs of their business process daily or even look for new insights to improve processes. While this can’t be achieved in two or three days, look to introduce little things first such as starting a meeting with some visualizations about the next topic or some performance KPIs. It should just become natural to use, see and talk about data.
- Provide a self-service strategy
More and more people will request, use and analyze data once you grow your data culture. It’s no longer possible to have a BI department solely and fully responsible for all available data and reports. Otherwise half of all your employees would have to work for the BI department. The best way to handle all the upcoming requirements is to implement a self-service strategy. You should enable people in all departments to work with your data – provide tools to enter the data and sandboxes as playgrounds, and, most importantly, to share their knowledge about the data and how to use it.
- Model your data business-driven
Using a business-driven approach, when working with your data, is, in my opinion, the most important point – even when listed last. Collect, process and analyze data only if there is a business case. A unified business model ensures that everybody shares the same understanding of the business and provides the opportunity to validate the business processes, rules and definitions. Therefore, domain experts and data experts have to refine and clarify their requirements collaboratively via conceptual and logical data modelling. Both sides benefit from such a collaborative process and business problems can be solved faster.
But as good this idea may sound, you should always remember – even the best data-driven approach shouldn’t stop you thinking for yourself. Data is there to support your decision-making process, not to replace it.