The portfolio average hides the vintage that is turning
Blended arrears look stable while a recent cohort deteriorates, and by the time the average moves the cohort is most of a year old.
Consumer finance
Loan performance by vintage, not by portfolio average
Originations, servicing and collections joined at the level of the cohort, so credit performance is visible early enough to change what you are writing.
A consumer lending book reported in aggregate is a book you cannot steer. Vintage curves, roll rates and collections effectiveness are the operative measures, and they usually require a manual extract and a fortnight. We build them properly, so the performance of what you originated last quarter is visible now, and so pricing, credit policy and collections strategy are argued from the same numbers.
Discuss your operationWhat we usually find
The problems that come up in almost every one of these businesses.
Collections effectiveness is measured by activity
Contact and payment rates get reported without being tied back to the balance actually recovered per pound of collections cost.
Unit economics stop at the interest margin
Acquisition cost, servicing cost, funding cost and loss are held by different teams, so nobody can say what a loan is worth over its life.
Where we start
What the first engagement usually covers.
Scope depends on the state of your systems. These are the pieces that recur in this sector.
Vintage and cohort analysis
Roll rates and loss curves by origination cohort, channel and credit band, refreshed on a timetable that allows a response.
Collections analytics
Recovery per pound of collections effort, by strategy and by segment, so treatment decisions are evidence-led.
Loan-level unit economics
Acquisition, servicing, funding and expected loss joined to revenue, so contribution is known by product and channel.
Regulatory and IFRS 9 reporting
Staging, provisions and disclosures built on the same joined data as the management reporting, so the two never disagree.
What changes
What is different afterwards.
You own all of it: the code, the written definitions and documentation aimed at whoever maintains this after us.
- Vintage curves refreshed without a manual extract
- Collections measured on recovery, not activity
- Contribution per loan by product and channel
- Provisioning built from the same data as management reporting
The builds behind it
Sector knowledge decides the order. The builds are the same four.
Single source of truth
Your finance, customer and operational records joined into one dataset, with each figure defined once and traceable back to the system it came from.
Analytics layer
Reporting built on top of the joined data, aimed at the few revenue and cost drivers that actually change the result.
Workflows
One defined task, automated inside a process that already exists, measured against whatever it replaced.
Financial modelling
Driver-based models built on the same definitions as your reporting, so the forecast and the actuals stop disagreeing.
Worth asking
Questions this tends to answer.
- Which vintages are turning, and how early can we see it?
- Which collections strategies actually recover money?
- What is a loan worth over its life, by channel?
- Where is credit policy costing more than it saves?