Data - JuicyScore Blog

Bringing New Clients – Increasing Approval Rates!

09 Oct 2017 I 2 min read

As many of you know, big data tools can add considerable value to identification of “good” and “bad” borrowers’ clusters. This research covers the use case of the seamless web identification technology in identification of “good” clients or in “saving clients from a ruthless credit conveyor.”

Many companies often encounter an absence of user income data; however, this should not be a reason to reject the new client. Alternative data can be of considerable help in this situation, and as part of our research, we identified a few patterns of “good” clients. These examples can help to increase approval rates by 2-5%, depending on the product and circumstances of alternative data collection:

  • Background of credit behaviour by device ID, previous approvals if any, and timely repayment of loans. This example is explained by the fact that apart from the main borrower or user there may be additional users in the family whose direct income is insignificant or non-existent. At the same time, net family income is sufficient to allow an increase in the credit line for the additional family members.

  • Proxy to devices’ price based on brand and hardware performance features of the screen, video card, RAM and hard drive of the device and other components. Analysis showed a 75%+ correlation between the disposable income of users and the price of their device.

  • Online employment check. This example includes data from the commercial web domain, and time of the day and IP aging can help to verify the user’s employment.

  • Permanent contact details. As part of the analysis, we noticed that if the user does not change his/her main contact details for six months or more, then the write-off rate decreases by 30-45%, and vice versa–if the user changes his/her contact details all the time, then the write-off rate grows exponentially.

  • User behaviour on the website. A number of behavioural markers significantly influence the quality of the borrower credit behaviour. For example, the amount of detailed reading they conduct on the website data entry speed, whether they adjust the data entered, etc.

We hope you find this review useful. We plan to conduct similar researches and provide presentations to market participants.

Have a question for us about something that is relevant to the development of your business? Send your queries to and we will do our best to cover them in our publications and studies.