Credit card companies make great effort to win a new customer. One of the challenges is to minimize acquisition cost, as this value is highly reflected in a portfolio’s payback. Often the sale is completed and the card is issued, but the customer never unlocks it, failing to generate revenue. Some customers need to be prompted, while others will unlock the card without further stimuli.
In order to act in a preventive manner, we implemented a model that attributes to each sold card its chances of voluntary unblocking. The recurrent implementation of this predictive model mitigated expenses and increased revenues. It allowed us to effectively act on each sold card that had higher chances of “not unlocking” – in addition, it identified a large portion of customers that spontaneously unlock the card within an expected time, without the need for stimuli.
An additional outcome was the creation of bases with process management indicators. The company started to monitor sales channels more closely, monitoring uncompleted sales, ensuring better management of the different sales channels as a whole.