Big data also exacerbates a very old problem: relying on the numbers when they are far more fallible than we think.
— Cukier and Mayer-Schönberger (2013) The Dictatorship of Data

The article is well written and comes at a moment in history when we need to highlight these topics. It’s also very relevant because data here isn’t really the subject. People are. And how people interpret the data.

“They were grossly exaggerated by many units primarily because of the incredible interest shown by people like McNamara,” said a third.

This is a comment about the effort in gathering the data for data’s sake. This isn’t a problem with data, this is an issue of objectives. In other term, this is a common management fallacy of mistaking the incentive for the final objective. When the incentive is at the same time simpler and more directly beneficial to the person, it takes over the final objective.

There are a lot of examples about this, and you might be able to find a lot yourself. The pattern is simple: someone sees an interesting objective, and defines some kind of metric to measure the progress to that objective. The measurement is, by definition, something simple to understand and simple to grasp. It seems working. What happens however after some time is that people start focusing on the metric itself, and forget the data. The metric in this context is a form of incentive because shows you in a simple way your next target, your +1, your record to beat.

An example are sales people motivated by a percentage over the number of items sold. That is a simple metric and a simple incentive. The issue however is that in the long term people will start caring more about the incentive and forget about the final objective: delivering a product to the people that need it. So they oversell, push too much, lower prices to get the sale, and so on, in the end harming the brand of the company.

Another example is the school system and standardized scores. At its foundation, school should prepare for life and work, and standardized scores seem a good way to point in the right direction: they measure how good the students are in learning. However, what happens is that many students start responding to the system, the scores and the results and forget the end goal: learn to learn. Because life doesn’t set the objectives for you, you have to set them. No school system, no metric, can define that because it’s not a zero-sum game. In life if you don’t like the rules you can change: market, profession, work, place, and yourself as well.

Big data is poised to transform society […] measuring and optimizing everything possible.

One interesting aspect of data again is that it’s often used to optimize, instead of using it to get insights. Optimization and insights can be seen as similar but in reality the first means taking what exists and making it more efficient, while the second can trigger an entirely new approach, smarter, groundbreaking.

Data is data. Information is interpreted data in context. And what you want is information, not data. This means that there’s no truth in the data by itself, but it’s helpful because it helps providing a bigger picture to figure out the information you need.

Data should be used to validate questions, to spot new things, to find improvements to a pre-existing task. When you conduct a scientific experiment you don’t gather all the possible data and then try to see if it validates your theory. You start with a hypothesis and then you check the existing data and new data to see if it’s true. And you’re open to throw your hypothesis away.

But in the end, I can quote Damasio again here: we are all humans. Without emotions, decisions can’t be made and that’s already known.

The key question here is: do you use data to validate decisions you already made, or do you let the data change your opinion?

This is a very human question. Data is just a tool.