An insurance company started a new insurance policy this month. The company has launched a direct marketing campaign to expand its policy holder base. To this end the company sends one of its staff members to personally interact with potential customers who are likely to buy the insurance in an effort to convert her/him to a new policyholder.
On an average, interacting with (that is, probing) each person costs the company Rs. 1,000 including expenses toward transportation, brochure, salary of the staff member etc irrespective of whether the person probed would eventually buy the policy or not. The average profit made by the company per policy holder per year is Rs. 10,000. So, if the person becomes a policy holder after probing, the net profit to the company is Rs. 9,000 (=Rs. 10,000 - Rs.1,000). However, if she doesn’t become a policyholder the net loss to the company is Rs. 1,000 (as the probing cost of Rs. 1,000 gets wasted!). Similarly, if a person who would really have turned into a policy holder is not probed then the company loses a potential policyholder and the net loss to the company is Rs. 9,000 (the opportunity cost of missing a policyholder). Needless to mention if a person who wouldn’t have purchased the policy is not probed there is neither any profit nor any cost to the company as no probing cost is incurred and there is no opportunity cost either.
The company wants to use predictive analytics to predict whether a person would buy its insurance product. If the prediction is “Yes” the person will be probed and if the prediction is “No” the person will not be probed. It may be assumed that the awareness level of the prospective customers about the product is such that no person would buy the product unless probed. Needless to mention, a person may or may not buy the product even after probing.