The case in hand is customer retargeting issue for a travel agency company. They have a big list of customers whom they have called once to sell their product and another list of customers who have purchased the offer from the travel agency. In the second list, some customers bought the offer without any effort from the company. Some other customers in the list bought it because of the company’s marketing effort i.e. direct phone calls. Our goal is to match the both datasets and find out the common customers.
Theoretically that’s not very difficult. Match both tables based on some kind of identification column and voila! But real-world data is not that simple unfortunately. For example, in this case both tables or datasets were recorded by different group of people. Also, information about these phone calls are often missing because customers are not always willing to provide or verify information like address, email etc. Even for phone numbers, which seemingly should be same for a person, differs a lot. Maybe at the time of the first call they used home phone. But at the time of buying they used cell phone. In such case matching by phone number won’t be fruitful.
(more…)
