Loan Amount Sensitivity Analysis and Interest Rate Sensitivity Analysis
a. See the effect of one factor at a time to each relevance factors (removing effect of other factors).
b. Form a strategy by combining all factors
c. Aim will be to increase the portfolio amount and thus increase revenue, keeping acquisition cost and risk low.
Bureau Data Use Optimization
In the current scenario very high acquisition cost is expected as they inquire many sources.
Business Rules
- Based on segmentation
- Based on Business Intuition
Predictive Models
FPD Model, Revenue Model, Conversion Model, Attrition Model, and Risk Model (like by behavior up to cycle 6)
a. Leads approved but not converted will affect the performance as it is resource intensive (like have to call to verify, do verid checks etc).
b. Leads which generate income less than Acquisition plus Operations cost are actually the loss making leads (hence can be called Attritors).
c. Good Profit comes from people
i. who actually go bad in the first pay but pay later with the late fee
ii. who renews cycle over cycle and stays long
d. Risk based pricing to decrease withdrawals
i. Lower interest rate for reactivations.
ii. Higher loan amount and low income for reactivations.
iii. Low interest rate for high score candidates (who score high in all scores – Revenue, FPD and 6 Cycle Risk).
Multiple Objective Decision System
a. Strategy Table and Joint Odds Approach
i. This technique is used if the number of dimensions is less. Up to four relevant factors (like Generic Score, Revenue Score, FPD Score, and Conversion Score) can be combined using this approach.
ii. Tackling more than four factors becomes extremely difficult.
b. Joint Score Approach
i. All Risk Related Scores are combined to form a single score (Joint Score).
ii. This approach can handle any number of scores.
iii. All decisions are based on this one score only
iv. There are ways to give more weight to one Relevance Factor over other (like giving more weight to Revenue than FPD).
Other General Analysis
a. Fraud Check (few examples below)
i. People who write name in improper capitalization are more likely to be fraud.
ii. When Home Phone and Work Phone is the same.
b. Miscellaneous Analysis (examples below)
i. Behavior in Military Population
ii. Absolute Income Behavior
iii. Loan to Income Ratio Sensitivity Analysis
iv. Lead Aggregator Performance Analysis
v. Lead Position Performance Analysis
2 comments:
I am interested in your analytics discussion of shortterm loans made via the internet, can u meesage me more info on this, such as software that can be used, how you measure performance, important factors to consider when analyzing the profittability of a ad campaign, and how to analyze the risks associated with applications.
Can you reply with more info regarding analytics for payday leads and business processes. I am interested in how you analyze the profitability of lead generation and risk analysis. gungor122@gmail.com
Post a Comment