The 5 Dos of Mobile Credit Scoring
As many lenders enter the mobile lending market, competition has demanded that they come up with lending models that allow them to meet fast turn-around times without compromising credit quality.
In Kenya, lenders find that they have to use either mobile phone platforms or use mobile apps to reach customers. In addition, customer transactions can be collected through automated mobile carrier payments, automated bank collections, or through customer-initiated mobile transactions.
Each of these environments will dictate the lending strategy used by individual lenders, whether banks, mobile carriers, or third parties. In all cases, the size and volume of transactions required at least some level of automation, and lenders have warmed up to mobile-based credit scoring, with mixed success stories.
Here are quick dos that every mobile lender should consider when building their credit scoring strategy:
- Make it about limit allocation – many retail scorecards are designed to give lend-no lend decisions, failing to recognize that a customer can fail to qualify for one amount, but qualify for a smaller amount. For any mobile scorecard to succeed, it must return a lending amount decision as its final decision without requiring further evaluations.
- Customer behaviour – irrespective of the limit allocation, scorecards must reflect past customer behaviour. Returning customers who have successfully completed past transactions should be rewarded with higher limits.
- Segmentation is king – consider the modelling effect of existing to lender vs new to lender, first time vs returning, secured vs unsecured, etc. some segments demand different models while others can share models with the segments as model factors.
- A learning model – the mobile market is highly dynamic, which makes models outdated in just a few months. Further, models can be gamed by customers. Stay ahead of the game with predictive models.
- Customer profitability – lending is all about profitability at individual transaction level. Lenders must maintain suitable profitability models, including bad rates both as a loss and as non-revenue portfolio.
To learn more about credit scoring and mobile lending, please email eric.thimba@thimba.co.ke or visit http://thimba.co.ke/.
Managing Director, Enterprise Solutions & Consultancy Ltd
8yWell done Eric & thanks for the article. If given 2 choices 1: Automated scoring & 2 Intuitive Intervention. What approach would you take? Please convince me as a lender why i should opt for automated decision making in current dynamic lending trends as opposed to intuitive intervention.
Risk Appetite & Risk Strategy Thought Leader
9yDavid, very true. But I should note that customers often don't Know that the are being scored. Most times, gaming is done by staff, especially sales staff if they discover the scorecard specs.
Senior Business Intelligence Developer
9yEric,Customer behavior analytic s is important in ensuring the aspect of 'gaming' gives the model an edge.
Experienced Commercial Banking, AVP | WASH Financing Expert | Innovative Financing | Sustainability |
9yNice read.
CRO | COO | Digital Finance & Risk Advisory Expert | Financial Inclusion Advocate | Policy Strategist | Board Member | Most Influential Female Fintech Leader 2024
9yThis is good from a high level approach, but end to end effective management of such dynamic portfolios is also very key among other additional strategies