Torture the data enough, and it will confess to anything

I was reading this article on Rethinking rules for a data-driven economy in the World Economic Forum by Angel Melguizo and Pepe Zhang . In it, they discuss the explosive growth of data in the next few years, with nearly 125 billion connected devices expected by the year 2030. The authors highlight the importance of building a data culture and have discussed the need for more data culture in certain industries and geographies worldwide.

Many articles discuss the need to be data-driven, but what is the opposite of data-driven? Avinash Kaushik coined HiPPO (Highest Paid Person's Opinion) in his book Web Analytics: An Hour a day.

Being data-driven isn't just a buzzword; it's a philosophy that champions letting data guide our actions and strategies. It's about being open to what the data reveals, even if it challenges our initial beliefs. Over the past two decades, we've witnessed the rise of data-driven companies, which have emerged as leaders in their respective industries. These companies recognize data's value as their primary or most significant asset.

I have worked with several companies that claim to be data-driven, but in reality, they are pretenders driven by the HiPPO principle. The most ironic was the CEO of a company that used the HiPPO slide in their presentations. They would often force the team to manipulate the data so that it validated their opinion in critical business decision-making.

I often refer to this as "Torture the data enough, and it will confess to anything." This situation is dangerous when data isn't used to inform decisions but to justify preconceived notions. This data manipulation can lead to misguided conclusions and poor outcomes.

This attitude towards d?a?t?a? ?t?o?r?t?u?r?e? data manipulation to achieve desired results is not just limited to executive decision-making. In this post, benchmarks don't lie, but liars use benchmarks; I discuss how benchmark analysts twist the results of a benchmark.

Steps towards building a data culture

Here are some basic steps to help in building a data culture:

1. Secure leadership buy-in and commitment to prioritize data-driven decision-making. No more HiPPO.
2. Provide comprehensive training and education on data literacy for all employees.
3. Foster transparent communication about the role of data in driving business outcomes.
4. Implement robust data governance practices to ensure data quality, security, and compliance.
5. Integrate data into everyday workflows and decision-making processes.
6. Recognize and celebrate successes resulting from data-driven initiatives.
7. Embrace a mindset of continuous improvement and adaptability in building a data culture.

Conclusion

In Silicon Valley, there is a joke that there will be only three types of companies in the future:

1. Companies that are data-driven
2. Companies that will be data-driven
3. Companies that will be bankrupt

The journey towards becoming data-driven is about more than just collecting vast amounts of data or employing sophisticated analytics tools. It's about creating a culture that values data-driven insights, embraces change, and remains agile in uncertainty.

What tips can you share on building a data-driven culture?

If you like, please subscribe to the FAQ on Data newsletter and/or follow Fawad Qureshi on LinkedIn.