DATA DRIVEN, OR FOOLISH?

Data doesn’t lie. But does it tell the truth?

‍

The current fixation on data analytics suggests that data is the only basis for sound decision-making. To succeed, we should rid ourselves of human bias and depend on cold hard numbers. This is foolish.

‍

Data is a tool, not an oracle. It can help us make better decisions, but it cannot make those decisions for us. Misusing data can be more dangerous than not using it in the first place.

‍

A British Prime Minister once famously declared in Parliament that there are three kinds of lies: lies, damned lies, and statistics. His words carry an important warning. If you cherry-pick the data you want, you can tell practically any story and claim it’s data-backed.

‍

Even without cherry-picking, over-reliance on data has many dangers. Without experience and intuition, decision-makers can allow naive assumptions to go unchallenged. And data is always backward looking. We can study what has happened before, but the world is full of discontinuities, like innovation, that don’t show up in past data.

‍

The financial crisis of 2008-2009 was caused, in part, by a naive assumption on Wall Street that we would not see mass mortgage defaults across the entire United States at the same time. It hadn’t happened before, so that meant it couldn’t happen. Wrong. Before the iPhone, Blackberry dominated the smartphone industry, and their market data showed that people only wanted phones with physical keyboards. Wrong again.

‍

The internet now makes it incredibly easy for us to access massive datasets that were not available before. Before investing, we can find historical transactions, tax records, demographic data down to the neighborhood level, and a wide variety of other information that was once difficult to collect.

‍

Should we just feed all this data into a model? If so, then we still have to decide what variables really matter, how to weight them, and what time frame is relevant. Questions like these call for human judgement.

‍

To use data well, it’s important to first have a clear understanding of your objectives and strategies. From this starting point, you can ask smart questions and answer them with data. I think of this as akin to the scientific method: have an idea first, then test it with an open mind.

‍

In the right hands, data is a powerful tool for understanding the world and making better decisions. But data is only as good as the judgement of the person who wields it.

‍