Affordability Insights: Unlocking the secrets of lender decisioning
Overview
“Affordability is a measurement used by lenders to assess whether you can afford to repay the amount you’re seeking to borrow, in addition to your current financial commitments.”
A feature designed to improve our customers’ chances of being accepted for credit by using their open banking acquired income & transaction data to produce a list of clear & actionable recommendations - highlighting the main ‘good’ and ‘bad’ factors from a lender perspective.
Problem statement
When deciding whether to accept someone for a credit product, lenders decide on several factors, from past behaviours, and current situations to projections on future behaviours.
A customer credit report can provide context to a customer's past financial behaviours and situations, however, the credit report is only one part of the puzzle and increasingly lenders are using more and more of the customer's financial and situation data to make those important decisions.
There are two parts to these decisions; Creditworthiness and Affordability.
Affordability is a forward-looking view of whether you can afford to make repayments going forwards.
Creditworthiness is more backward-looking; reviewing how you have used credit in the past. This is used to make a judgement call on how likely you are to make credit repayments in future.
Creditworthiness & affordability are both key elements that contribute to overall eligibility.
We had been working hard to make customers understand where they stand on the creditworthiness side and now wanted to unlock the affordability side.
We wanted to show customers that;
Lenders look at all your finances and how you use your money
They will decide if they want to give that loan or card based on what they find out
We will tell you what they look for so you can do something about it
Users and audience
In general, the audience base of TotallyMoney tends to skew to the more near-prime and sub-prime ends of the markets and it is these types of customers who, with fewer credit options available, need to be armed with the right information before they make important application decisions.
From the current customers’ open banking data, we had available we were able to see that;
56% of our customers have used BNPL in 2023
64% of our customers do not save anything each month
20% of our customers gamble each month
Of those who gamble, customers spend an average of 17% of their income on gambling
Also
The current eligibility rate of those customers was 42%, meaning they are rejected more often than accepted
Our creation of our Improve area of the experience had shown great sustained usage (50%+ of App MAU) so we knew there was an appetite for this type of content
Roles and responsibilities
Responsible for the design and strategic direction of the product and team roadmap alongside the Product Manager. As well as our 2 front-end and 1 back-end engineers, we also closely collaborated with Copy, Product Marketing, Commercial and Compliance teams.
Scope and constraints
The feature would sit within the recently evolved Improve area of the experience and would require customers to connect their bank accounts through open banking. The open banking connection was key to us being able to construct these insights.
The feature would also be app only, due to the open banking aspect, but we would use the existence of the feature to hopefully onboard customers from web to app.
Our process
Following our initial investigation and analysis of the open banking data of our customers (there were two live open banking features already within the app experience) and the data from credit reports, we understood there was an opportunity to pursue here.
Through conversations with our open banking provider Bud, we understood the data we had access to and aligned this to the outputs of conversations with lenders (through the commercial team) about what they look for when considering applications for credit products. This allowed us to shortlist several factors to investigate for an initial release.
Affordability factor possibilities
Armed with all this information we needed to find out if this was a problem our customers knew about, cared about and we could build them something useful that would have sustained use and value.
We bucketed our assumptions into three steps that would help guide the development of this project.
This is a problem our customers care about
We can effectively articulate the value to customers
This is a helpful solution for our customers
Assumption 1: Is this a problem our customers care about?
The only way we could figure this out was to talk to our customers. Over the past couple of years, we had built a 10,00 strong panel of our customers through Great Question which has proved invaluable when you need to talk directly to the people who use your product.
With the recent evolution of the Improve area of the app, we had a pattern set up for an ‘insight’ screen through the existing Credit Report Insights feature. This allowed us to create a quick rough prototype that we could use in our initial conversations with customers.
Early affordability exploration
Through these conversations with 5 of our customers and this rough prototype, we were able to learn several things.
Affordability as a theme is understood or can be guessed
Participants mostly understood there were 2 parts to decisions lenders make, historical credit use and something around disposable income and how you use your money
Participants have had little or no explanation as to why they have been rejected for credit.
The structure of the plan and the detail pages were received well.
Whilst not screening for anything to do with gambling issues / addiction, we had a couple of participants who had direct or indirect experience with gambling addiction. Both appreciated the content and the tone of the information displayed to them.
Participants understood through the prototype that bank accounts were used and analysed to provide insight.
Affordability research conclusion matrix
Assumption 2: We can effectively articulate the value to customers
The insights we learned from this research study allowed us to begin crafting how we could communicate this feature to customers. Working closely with product Marketing and Copy and armed with 2 (compliance approved 😬) messages we began testing these in promotional entry points in the Improve section of the app, these entry points opened a modal with further information about the new feature to build anticipation amongst customers.
Message1: ‘Improve how lenders see you’
Message 2: ‘What’s impacting your affordability?’
Affordability messaging options
Of the 2 messages tested ‘What’s impacting your affordability?’ was the clear winner with 9.3% click rate against a 6.3 click rate for message 1. This clear winner meant we could be hopefully much more effective at launch with how we communicated to customers.
Affordability messaging test results
Assumption 3: This is a helpful solution for our customers
As product teams, we make sure we are learning the right stuff quickly, cheaply and early in the process, making sure we can put the right decision points in and get enough information to the ‘next step’. I’m a big believer that the real truth comes from a live feature in your product with real customers using it in a real context. ( The Truth Curve from Giff Constable gives a great overview of this).
Using the learnings gathered so far, designs progressed into a state where we were moving towards build. There were a couple of key decisions we still needed to make. We narrowed down our shortlist of factors affecting affordability, through conversations with our back-end team, to an initial list of six;
Income
Employment
Savings
Buy Now Pay Later usage
Regular spending
Gambling
The gambling factor has been a topic of internal conversation due to the sensitive nature of the issues and addictions that can have a huge impact on customers’ lives. We wanted to make sure we were being honest with customers, and that the amount they gambled was having either a positive or negative impact with lenders but allowed an opt-in option to see more detail beyond that. Some of the learnings we gathered from our initial research study had given us confidence in this approach.
Another key decision as we moved towards build was to make these new Affordability Insights an opt-in experience for those customers who were already connected via open banking (via one of our other open banking powered features), through conversations with compliance we felt that using those customer's data to create these insights without explicit consent could be considered a ‘bit creepy’, but something we could revisit after some maturity of the feature.
First release Affordability Insights designs
First release Affordability Insights flow
Following a round of internal testing, we were ready to launch the feature but not really much closer to answering our assumption. After a week of the feature being live in the app we added a feedback module. This is something within our team that we have found really valuable when launching new features, being able to get quick feedback from customers using the feature in real-world contexts.
Our initial success metrics alongside open banking consents were tied to this feedback module. Customers were asked to rate the feature on a scale of 1-5 and were asked how the feature could be improved. This anecdotal feedback would be vital in shaping future updates and releases.
Initial results from the feedback were good, 73% of customers rating the feature 4 or 5 out of 5. This succeeded the 70% threshold we had set and stacked up very favourably against other features we had launched. The anecdotal feedback also proved invaluable, a theme emerged where customers wanted more detail or to understand how these calculations or insights are made.
Part of our vision we had for this feature was to make sure we added the right amount of depth, adding transactional data for each factor was something we had descoped for the first release but now we had validation for how we can progress the experience for the customers. As an initial first step, we added overlays for each factor explaining how each factor is calculated and where the data comes from.
Factor description overlays
Post-release of these overlays, the 4-5 star rating increased from 73% to 80%. It's a great example of establishing a baseline, making quick changes based on customer feedback and then being able to see positive change quickly.
We have continued to evolve the feature adding entry points across the experience, alerts and notifications when a factor moves from positive to negative (and vice-versa) and are in the process of launching the factor detail pages (BNPL and savings first, based on the feedback module) which include transactional data and further details about that factor.
Savings and BNPL factor pages
Outcomes and lessons
This was a really interesting project to be part of, trying to unlock something that has been a bit of a black box for customers when applying for credit products. How and why are lenders making a decision about me and my suitability for their credit product and its great to try and demystify those for the TotallyMoney customer base. Any feature that uses open banking data to power and populate always comes with the knowledge that you have to work hard to prove the value to the customer with reconsent windows lasting 90 days and using open banking will always (at the present time) impact the scale you can grow the product to. As open banking continues to grow across financial services, we hope that the adoption rate will continue to grow
We had a good process on this project and being able to craft our process around key assumptions worked out well. Having those key inflection points allowed us to make key decisions at the right time. If we had to do this again, I would have pushed back harder against the opt-in state for those already connected customers, as I think growing the early adoption rate outweighed any perceived concerns about the use of data.
Total open banking consents attributed to Affordability Insights: 19,443
Affordability Insight customers: 31,610
Daily Affordability Insights page views: 50,540
Feedback module: 584 responses, 80% scored 4 or 5 out of 5 against target of 70%
Graph of open banking consents attributable to Affordability Insights
