Decision Making – Is everything data driven?

From the most complex decisions like waging a war to the most mundane one like selecting a dish for lunch, what intrigues me immensely is our ability to take decisions. Although some decisions appear to be random or, at best, based on intuition; they are all data driven.

Interestingly, decisions are not good or bad but are more like actions. What can be good or bad are the consequences which are associated with the decisions. Consider a situation where we are building a dam. Residents in the nearby areas would benefit (or consequence) significantly from the consistent water supply for drinking as well as irrigation (and innumerable other benefits). However, as we think about consequences (or benefits), we have to get another party in the mix i.e. the impacted party. With the dam project, many would go on to lose their piece of inherited land. Now we would be able to see how the same decision has different consequences for the actor and the recipient. Hence it becomes increasingly difficult to ascertain the quality of decisions and subsequent consequences in isolation.

Have you heard of the reports which talks about by vision 2030 - “80% of population would have smart phones”, “Artificial Intelligence would impact employment”, etc. All these forecasts, at best, are based on gut feeling, stated with a little bit of data and lot of confidence (read Confessions of an Economic Hitman). In many cases, it is the gut feeling, something we cannot pin point to or explain which provides us with that data to be used in our decision making.

Assertion 1: All decisions are data based leveraging conscious or sub-conscious data points.

It may appear on the surface that decisions are not necessarily data driven but even intuition is data driven. It can be explained by our experiences or things we would have heard and/or felt during our upbringing. Classic example would be the stereotypes we have built. We have built them as they helps us to take faster decisions, may not be the best ones though.

Let us take another example. From pre-historic times, key to our survival has been our ability to take decisions. Hope and fear are two integral drivers - whether it was the discovery of fire to protect ourselves or the invention of wheel to travel to far off lands in hope of better life. Now how did we start using fire to protect ourselves! Early men would have noticed that animals ran away when there were accidental fires in the jungle (the first data point). On further investigation, they would have realized that the green part of the forest did not catch fire but the dry part did (the second data point). Hence they got together and decided to give it a try by gathering some dry leaves. But the process of ignition was still unknown. That is where the data point which may have seemed immaterial at the time of observation helped. It may have remained as a sub-conscious memory that two pieces which struck together had generated the first fire (third data point, probably sub-conscious). Now when the time came, they had pulled a part of their memory and the dry leaves. Eureka! They had the fire. The point here is not the accuracy of the story or if that is exactly how the early men got fire; but if we assume ourselves to be rationale beings, this could be a possible approach.

Assertion 2: Data Value = Extrinsic value + Intrinsic value

We have always been fascinated with collecting data with the idea being 'more the better'. Increasingly, the focus is shifting towards utilization and monetization of data. Key step towards monetization is to attribute a value to the data.

Each data element has an extrinsic and intrinsic value.

• Extrinsic value is a function of time difference between the generation of data and the consumption of data as well as the usage of the information.
• Intrinsic value is the intrinsic value of a data attribute which would not change with the passage of time. Each data attribute would have a different intrinsic value.

Let us consider an example of contact details. Phone number would have a constant intrinsic value by the virtue of it being a detail to reach out to the customer. However, as the time elapsed increases from the last successful contact, its extrinsic value decreases. Consider a scenario where bank has to collect dues (like EMI on loan). In such a case, the Operations team would try to reach out to the customer on all the phone numbers provided, if there is no response of the latest one. Here, the latest phone number would have high extrinsic and high intrinsic value (and hence would be used first). On the other hand, older phone numbers would have low extrinsic value and high intrinsic value.

The next challenge thus is how extrinsic value of data plays a role in a bank’s service offering today.