Why data is not enough

The last couple of decades could easily be described as the Age of Data - the forerunner to the Age of AI. Once people started to move their worlds online, whether their network of friends, their shopping or travel or their banking, the richness of data on human beings grew exponentially.

Data is everywhere. Someone somewhere is collecting data on people’s movements through a store, the products they scrolled past, the locations they visited, the text they typed in for searches. In fact, in some ways we are drowning in data.

Nonetheless, despite all that data individuals still make poor decisions, or have conflicting viewpoints, or make different judgements about the world. Just having data is not enough - no matter how prettily it is packaged or charted.

To give an example. I can open up the bonnet of my car and see all the bits that are there, but if the engine goes wrong I’ll call out a mechanic because when he sees what I see, he will see the fault. We can both see and so have the same data, but the difference is that he knows what patterns to look for, and interprets what he sees better than I can.

Sports fans can see the players on the pitch, and how those players perform but that does not make them coaches. The coaches see the same data, but interpret and use the data differently from the fans.

In business, the same happens with accounts. The financial data shows the business activity, but the skill of the accountant is not in being able to collect more data, but in being able to make sense of the patterns because of their training and experience.

Data does not stand alone. It only makes sense if you have a framework for interpretation. It is a reason why management schools teach using a case study method, where the data is fixed and limited to what is in the text, but you are required to find conclusions, because what they are trying to teach is how to ‘see’.

Experience and learning teach us what patterns to look for in the data, because some data is useful, some is interesting, some is important, and some is irrelevant. And our reading of the data depends on the task at hand and what we need to accomplish.

The emergency exit signs and fire points in the airport as mostly irrelevant when you are traveling. The data you care about are the flight times and gate numbers, or the prices of the duty free, or where to get a coffee. We do not see the safety signs and emergency doors until the point at which there is actually an emergency, when our data priorities suddenly change and all the coffee and chocolate become irrelevant.

It means useful data starts with having a framework or understanding of the patterns that are neede, and to realize that the frameworks and viewpoints that apply to a given situation depend on what you trying to achieve. The data does not change. It is choosing the right way to look.

Like debugging a computer program, the right piece of data at the right time will tell us what is broken and what to fix. But, crucially, we have to know what data we have to look for.

So collecting data can be an end in itself for purely scientific interest. But to have meaning it needs to be attached to a purpose, a pattern and a question you want the answer to.

A long time ago I was involved in conducting a usage and attitude study where the end client wanted to know everything. But because it was all data, it was extremely difficult to present and share the findings. There was no perspective on what was important because the study was not created to answer a question. To build a story needs an objective or motivation that creates a journey and narrative through the data.

Secondly, the value of the data is defined by the knowledge and skill of the person looking. A sales manager is looking for messages that can be amplified, and to be forewarned about potential sales objections. A CFO will translate uptake percentages into units shipped and revenues. What you see is affected by your expertise and how you see a pathway to translate the data into an action or solution.

More data without the vision to know what to do with it can end up as just more straw on the haystack. The needles can get harder to find. Most importantly, data which does not change anything might be interesting, like facts in an encyclopedia, but they are not important unless the help someone solve a problem.