Is Data the New Oil?

Something I frequently hear is the phrase “Data is the new oil”. This is supposed to have us picture data as the great commodity of the 21st century, or even more, ourselves as the next Rockefeller if we can just pick it up off the ground and simply refine it. The problem with this is that data is not a commodity. We cannot take one piece of information and treat it as if it is just as good as another. More importantly, once it is refined, it is not equally valuable to everyone. Once someone has data, its worth can only be defined by what that person actually does with it.

In my day to day job, I administer a lot of data. A lot of it comes in and a lot of it goes out. A lot of it gets processed further and is reported on, and a lot of it gets archived for possible future use. Some people may be interested in the sheer amount of it, but the actual amount is only useful when discussing hardware and bandwidth requirements. Instead, we ought to be discussing the usefulness of it.

In a study done by Bastardi and Shafir (1998), they consider the gathering and use of information in decision making. In one experiment, they would ask a group of participants to pretend they were a loan officer at a bank and needed to approve or deny an applicant’s loan request. They were told that the applicant already had an unpaid loan, but there was conflicting paperwork as to whether the amount was $5,000 or $25,000. The participants could choose to approve, deny, or wait until the unpaid amount was confirmed. After the participants answered, they informed the people who chose to wait that the amount was confirmed to be $5,000. At that point they also needed to decide to approve or deny. Here are the results to that experiment.

Time of choice
Choice Immediately After waiting Total
Approve 2% 54% 56%
Reject 23% 21% 44%

In the experiment, 75% of the participants chose to wait to find out the true amount of the unpaid loan. You may be thinking that most people are planning to deny the loan if they receive bad news (the $25,000 amount is confirmed) and approve it if they receive good news. However, we can see that this is not the case. 21% of participants chose to deny the application even though they received the good news. In this situation, for these participants, the confirmed loan amount was worthless data.

What are the costs of pursuing additional data and how will receiving it affect a decision? In the above example, the cost was minimal. They were given the confirmed loan amount immediately after telling the researcher they would wait. However, in the real world the cost of waiting is not minimal. It is difficult to know ahead of time how information can affect our analysis about a decision. Thinking a little bit about the possible outcomes and what we might plan to do in any case can not only help identify worthless data, but also eliminate a potentially costly waiting period.


Jason Schalz

Director of Technology

Leave a Comment

Your email address will not be published. Required fields are marked *