BANKING & FINANCE
Credit appetite prediction: reach the customers who want a credit
Credit appetite prediction with TIMi estimates, for each customer, the likelihood of taking up a credit product. It is a special case of propensity to buy, with the profile of the customers most likely to say yes.
Appetite score
Rank customers by likelihood to take up a credit.
Customer profile
See which characteristics define the likely takers.
Alongside risk
Combine appetite with credit scoring before any offer.
HOW IT WORKS
How credit appetite prediction works.
Three steps, in one engine, on a standard server with no cluster.
1. Prepare the data
Anatella gathers products held, transactions, past offers and responses, then structures them into variables.
2. Build the model
Modeler learns from customers who took a credit and those who did not, and keeps the variables that matter.
3. Score and target
Each customer receives a score. Offers go first to those most likely to be interested, and the profile explains why.
Anatella
ModelerGETTING IT RIGHT
Choices that shape a credit appetite model.
Wanting a credit and being eligible for it are two different questions.
Appetite is not eligibility
A customer likely to want a loan is not necessarily one to lend to. Appetite scores are combined with credit scoring and default risk, never used in place of them.
Validate the regulatory frame
Targeting for credit products falls under consumer credit and data protection rules, including GDPR. The variables used and the use of the scores must be validated with your compliance team.
Do not use segmentation for a clear target
When the target is clear, such as taking a credit, a predictive model is the right tool. Segmentation is better suited to adapting messages by group.
EXPLAINABILITY
A score you can justify internally.
In a regulated sector, a targeting score has to be explainable. A white-box model shows variable by variable what pushes a customer up or down the ranking, which supports compliance review and discussion with commercial teams.
For commercial teams
See the profile of the customers most likely to take a credit.
For compliance
The variables used are visible, so their relevance can be reviewed.
For model owners
Modeler’s built-in Drift Report shows when a model loses accuracy, so you update it when needed instead of on a fixed calendar.
DATA
What data an appetite model uses.
Anatella connects to more than 80 sources, so this data can be gathered without prior custom development.
Products held
Accounts, cards and existing credits.
Transactions
Account activity and spending patterns.
Past offers
Offers made and responses obtained.
Customer profile
Demographic and relationship data.
PROOF
Credit targeting in production.
These two references are published lead-scoring stories on credit products. Credit appetite is a special case of the same approach.
Sales performance multiplied by 4
Cofidis profiled the prospects most likely to convert and contacted them proactively: 6.2% conversion against 1.5%, and €3.5 million of additional gross revenue.
26% of the top 15% contacted bought a credit
ING targeted the 15% of business clients most likely to need funding and raised conversion by 20%, with limited commercial resources.
FAQ
Common questions.
Is this different from lead scoring?
It is a special case. Credit appetite applies the same propensity-to-buy approach to a credit product.
Does a high appetite score mean we should lend?
No. Appetite measures interest, not creditworthiness. It is combined with credit scoring and default risk before any offer.
Can we use it on existing customers only?
It works best on customers whose history you hold. For prospects, the data available limits what the model can learn.
Who validates the use of the scores?
Your compliance and legal teams. Rules on consumer credit and data protection vary by country and are outside what TIMi decides.
RELATED USE CASES
Keep exploring.
SEE FOR YOURSELF
Test it on your own data.
Run TIMi on your own data with a free trial, or talk to us about your case.
