Monday, February 11, 2019

2019-02-12 Fintech Lending & collections

Loan business - Fintech- is very complex process and cosinsist of following steps:
A) Market knowledge
B) Usage of the technology
C) Social media - ways to talk to customer
D) Attracting of customer
E) Getting customer on your web
F) Customer start to fill-in application
G) Customer finish the application
H) Data verficiation
I) Knock-off criteria
J) Fraud detection
K) Underwriting
L) Subscription
L) Loan disbursement
M) Repayment
N) Receivables management - provisions
O) Collection - provisions
P) Writing off - sale
R) Back to model - model adjustment

Therefore the company can not focus on collection only, as the whole process us much longer and the first priority is to attract as many customers to apply for the loan. The big amount of resources is spend there. A reasonable ratio of DPD60 is expected, as DPD 30 is the most profitable loans company gets. DPD 90 or DPD 180 is a cutt off, as the money is reasonably not going to be collected and therefore it is a pure loss for company.

Collection itself will also not work one way meaning from underwriting to late payment over management of receivables to collection, but also other way from collection to underwriting - to adjust the model and even sooner from collection to data verification and application setup - as information obtained shall be used to improve health of the company.

A cooperation between all the deparments is critical, as the pairs of departments with different values shall be created.

F.e.
Risk department - financial department
Underwriting department - collection department
Client acquiring - Underwriting
Market knowledge - Social media
Collection - Finance
Finance - Investor
Data verification - Market
Audit - Risk
and so on...

The main objective of Fintech is to underwrite as many reasonable risky loans with good enough cash flow.

It is the complexity of initial steps, importance of matching the data with other sources in beggining and later in collection where the company can improve fastest. As the underwriting models could not be shared and if they could it will be limited as the products change, the market data, knock offs and collection could be done on universal basis that are no company specific.