This week we unveiled our latest update to Atlas (our manual underwriting platform): Risk Insights View, designed to bring everything you need to assess risk into one easy-to-use screen.
The feature presents key financial behaviour indicators such as time spent in unarranged overdraft, predicted gambling spend, returned direct debits and use of different types of credit, using a simple traffic light system to highlight the behaviours that need investigation.
You can click on each indicator to deep dive into trends in the behaviour over the last year and all transactions relevant to the behaviour. You can easily filter transactions by month to focus on key periods such as a spike or turning point, and move between these and the balance graph. This allows you to understand key indicators in context and quickly get to relevant transactions.
The Affordable Credit Challenge
The concept behind the Risk Insights View was born in November, when we teamed up with our long-time partner Serve and Protect Credit Union (SPCU) for Nesta’s Affordable Credit Challenge. The task was to build an Open Banking product that would help community lenders to compete against the bigger players and increase access to fair, affordable credit.
Together we identified a key opportunity. SPCU’s expert team makes sophisticated, highly nuanced underwriting decisions. Analysis of decisions and loan outcomes, paired with Open Banking data, could help to formalise, and then streamline, this process - by distilling it into key financial behaviour indicators.
This could enable financial behaviour to be monitored in real-time during the loan term using our Refresh Reports functionality. This in turn could support SPCU’s goal of a ‘rebate’ loan, in which customers could get a refund on their original interest rate if they enacted positive financial behaviour change.
The Affordable Credit Challenge was an iterative process; the task of phase one was to distil and align the underwriters’ processes within Atlas, our platform for easy manual underwriting. We worked with SPCU to identify the key financial behaviours to assess in an underwriting decision; bring them into one View and add pass/fail thresholds for each in accordance with SPCU’s policies.
An early prototype
We found that the thresholds were an excellent way for underwriters to get a quick ‘feel’ for a borrower’s behaviour. However, after various design iterations were tested over calls and in-person sessions, we came to the conclusion that a decision required more detail about each behaviour, and an understanding of how these behaviours interacted and their context.
For example, your applicant might have a predicted monthly gambling spend of £250, 10% of their predicted income. To understand what this behaviour means for their creditworthiness, you’ll want to dig deeper: are they gambling in their unarranged overdraft or when they have money to spare? Are they on a downward spiral? Do they gamble many times in a single day?
We honed in on the concept of custom indicators and thresholds combined with a deep dive into month-by-month historical trends and relevant transactions - employing the principle of progressive disclosure to support efficient underwriting. This became the Behavioural Insights View, with which - we’re proud to say - we won one of the three spots in the challenge!
Adapting Behavioural Insights for our lenders
The core of the Affordable Credit Challenge was to create solutions that could benefit the broader Credit Union and community lending market. With that in mind, we tested variations of the Behavioural Insights View with a number of underwriters from our client base. We settled on combining some of the indicators developed with SPCU with others that already existed in Atlas:
Unarranged Overdraft Usage
Returned Direct Debits
High-cost Short-term Lender Relationships
Debt Collection Agency Relationships
DMP and IVA Relationships
Negative Net Disposable Income
This research reinforced that for a given indicator there is no simple ‘good’ or ‘bad’ amount (since behavioural context is vital to interpret behaviour), but that visual aids such as colour coding and icons were hugely useful to help pick out the indicators that need further investigation. In place of SPCU’s pass/fail thresholds, we implemented a simple traffic light system with easy-to-remember thresholds. As well as highlighting what needs a closer look, this helps lenders to get an immediate ‘feel’ for the applicant: some lenders might look for all ‘greens’; others might look for all ‘oranges’.
We also found that underwriters frequently used the balance graph in risk decisioning, as well as transactions from a particular time period like a turning point or period in unarranged overdraft. We built these into the View so that underwriters didn’t need to go to other parts of the report for this information.
After BETA testing with a group of clients, we’re now live with the Risk Insights View for all of our clients. The process doesn’t end here: we’ll be collecting feedback, integrating our scores, working on new versions, and adding new features and indicators as our offering expands.
If you want to find out more about Credit Kudos and our latest product development, get in touch with one of the team here.