CaaS works by automatically analysing and categorising a customer’s bank or credit card transactions to identify patterns in their financial behaviour. Data may be shared in real-time by a customer, with their consent using Open Banking or undertaken by a bank using their own customer’s data to proactively manage their relations with customers.
The results have been striking. For instance:
- Poor financial resilience: Our work with one lender helped them identify the 10% of customers who made no savings contributions in a 12-month period, suggesting a lack funds to support a down-turn in their financial well-being.
- Over indebtedness: CaaS identified the 1% of customers in a credit-card portfolio who were financially stretched using cards to pay their mortgage or a personal loan.
- High-cost indebtedness: CaaS revealed for a provider of current accounts the 10% of customers with a high-cost short-term loan – and the 2% with more than one.
Why low-resilience matters
Even before Covid, one adult in five in the UK (20%) showed signs of low financial resilience. For these consumers, a £50 reduction in monthly income or a £300 unexpected expense could be too much to withstand.
Financial firms have ethical and FCA obligations to support these vulnerable consumers. The regulator sees them as “especially susceptible to detriment, particularly when a firm is not acting with appropriate levels of care”.
Finding – and helping – the invisibles
Many are not just vulnerable, but also invisible as credit bureaus don’t always capture information on missed regular payments.
Finding these strugglers is clearly important for credit risk and financial education teams. It lets risk experts manage their portfolios better and helps ensure they aren’t overloaded with additional debt they can’t afford.
And it lets customer management and education teams go beyond basic FCA expectations in how they support people at risk.
How CaaS works
CaaS is driven by a series of machine-learning algorithms. These are trained to automatically interpret bank transactions including spend, frequency and transaction descriptions to identify types of income and expenditure appearing on a customer’s current or credit card account. The tool places consumer transactions into one of 180 different income and expenditure categories. What emerges is a granular, real-time picture of where money is coming from, where it is going, and any significant financial changes. Data can be shared by customers with their consent using Open Banking, or organisations can use CaaS to analyse their customers own transaction data in their back office.
Find out more about Categorisation as a Service and how it can support your businessGet in touch
Helping credit risk and collections teams
CaaS can be applied throughout the customer lifecycle. It can provide a variety of different markers of financial behaviour for a single consumer, providing lenders with greater insight into an applicant’s circumstances, an indication of the health of a particular portfolio, as well as an array of early-warning indicators so that lenders can support consumers before they start missing payments.
By using CaaS, credit risk and collections teams have achieved excellent results.
- Fewer credit decisions going bad: For acquisitions, CaaS makes affordability calculations significantly more accurate. One client has cut acceptances that go into debt by 40%.
- Extending more credit without increasing risk appetite: Using CaaS, clients have been able to accept more applications, offer more products to existing customers, or extend their limits – without increasing risk appetite.
- More accurate credit risk management: Work with clients has shown the CaaS makes decision models 10-15% more accurate.
For existing customers, CaaS can analyse 12-months of current account and credit card transactions in milliseconds. For new customers, it can also be used to analyse any Open Banking data that applicants consent to provide.
Either way, having the full picture lets lenders do the right thing for consumers at the right time.