Let’s take an example.

Consider a retail bank offering savings products to its customers. It collects all kinds of data – the customers, the transactions they do, the amounts of money they keep in their accounts. Let’s take a random piece of data.

Customer x001: ATM user: Yes Customer x002: ATM user: No

Now what?

By itself this piece of data means nothing beyond what it is. To make anything from it, one needs something else – a purpose. I must want to do something for which this piece of data can do something.

Purpose is context dependent. If I am in charge of ATM usage at the bank, my purpose is to increase the number of people using ATMs. For this I need to make the ATM services more customer friendly. But what is “customer-friendly”? And how will this piece of data help me?

The point being made here is that generating the data report is not the end of the process. It is not even the beginning. It is a small input into a larger context. In the above example, the ATM Manager will, most probably, follow a chain of thinking that goes a bit like this:

Who are the customers who are using the ATM? What do they find attractive about it? Who are the customers who are not using the ATM? What is holding them back?

From the random piece of data, the ATM manager needs to extract some relevant information and from that perceive a critical insight that can take him closer to a solution. For example:

Profile of ATM non-user: 77% are senior citizens 68% are rural customers 44% are women 23% are High Networth Individuals (average account balance > 500,000)

The above is information gleaned from churning the data. But what useful insight can one draw from it? For this the ATM Manager needs to go beyond the data and study the human beings. An example:

Women are scared of entering ATMs alone. They look at the ATM kiosk as a place where they may get robbed, attacked, etc. They prefer to go to a branch where there is safety in crowds.

Now, the ATM manager has something concrete that he can act upon:

Make ATMs safer for women – relook at lighting; location; behavior of security guards, etc.

Thus, we have a journey from a random piece of data to concrete action points. This journey is what fulfills the promise of data. And this journey happens not in the computer but in the mind of the user. There is a thinking process that works backwards from the purpose and asks the right questions, supplements the data with appropriate additional information and designs hypotheses for action.

How do we enable the user to make this journey each time, every time?

The Fifth Elephant 2012