Guest authors: Chris Fitch, Vulnerability Lead, Money Advice Trust & Research Fellow, Personal Finance Research Centre, University of Bristol and Jamie Evans, Researcher, Personal Finance Research Centre, University of Bristol
Online loans: using data to see difficulty
Do you have to see something to observe it?
I know. You might have to read that a couple of times before it makes any sense.
But it is a question that every online lender needs to consider.
The reason for this is that online lenders have the same responsibility to identify customers with decision-making difficulties, as lenders offering loans in person or over the telephone.
This includes identifying customers who – for health reasons – can struggle with understanding, remembering, evaluating their options, or communicating a decision.
And that introduces a challenge: unlike ‘traditional’ lending settings, online lenders can’t see or hear their customers.
When a customer goes into a bank branch to ask for a loan, or makes a telephone application for credit, staff will look for signs of decision-making limitations.
These can include the customer struggling to understand the information they’ve been given about a loan. Or the customer having difficulty with recalling personal information (like their date of birth).
When difficulties like these are spotted, staff should offer the customer support to overcome them, or they can pause the application, and even decline credit.
This is all straightforward (well, on paper at least), with the Financial Conduct Authority providing guidance on the steps that firms can take.
And such ‘monitoring’ can also help those of us living with mental or physical health conditions that can, at times, impair our ability to make financial decisions.
But online loans are different – they don’t involve face-to-face contact with a customer (some lenders – such as Nationwide – now offer video loan applications for selected products). Nor do they involve telephone contact (some lenders will follow-up an online credit application with a telephone call to verify customer details). What then should we expect?
Currently, a common misconception exists that:
- this lack of personal contact with a customer makes it impossible for online lenders to identify difficulties with decision-making
- online lenders cannot therefore prevent detriment to these customers, even if these customers might have been identified if they had applied in person or on the phone
- online lending is consequently a ‘special case’ and is exempted from FCA regulatory guidance on identifying and supporting customers with decision-making difficulties.
However, none of these misconceptions are correct. Online lending is covered by the FCA’s Consumer Credit Sourcebook, and online lenders can spot decision-making problems.
So, what should online lenders do?
1. Use data differently
Online lenders may not see or hear their customers, but they have another ‘sense’: data.
When applying for an online loan, we will enter information about ourselves (like our date of birth or employment status), which is used to help decide whether we get credit.
However, other data are collected for administrative purposes which are more interesting – critically, these data are not about ‘who we are’, but are instead about the mechanics of how we filled-in, navigated, and completed our online application.
These vary, but can include information on where customers made data entry errors, how many errors were made, and how long was taken over this. Data may also be recorded about on-screen navigation (including repeated use of help pages).
Although collected for internal design teams to fix any difficulties that the overall customer base has with completing an application, these data can be used for another purpose:
to see (and respond to) individual decision-making difficulties.
2. Use data to see difficulty
When making a decision, we use four main abilities – our ability to understand, remember, and weigh-up relevant information, and to then communicate a choice.
Lenders can use this ‘four ability framework’ to consider what the administrative data they already collect might reveal about customer decision-making difficulties. For example:
- understanding: indicators of customer data entry errors – including entry of incorrect information, repeated failures to provide requested information, or inconsistent answers – can all signal a difficulty in understanding ‘what needs to be done’.
- recall: data collected about the number of mistakes a customer makes when trying to recall basic personal information (e.g. date of birth, home address), as well as the time taken to do this (including ‘dwell time’ on individual questions), can indicate potential memory difficulties.
- weighing-up: data highlighting customer difficulty in choosing between available options, or the repeated changing of answers and preferences, may flag a problem.
- communicating: customer difficulties with numerical data calculations or form-filling, can indicate a potential communication problem.
As administrative data will vary across different online lending platforms, lenders should map their data against the four decision-making aspects above.
3. Use multiple data-points
Where organisations have access to data which could indicate a decision-making limitation, they need to interpret this sensibly – and this means not relying on single data indicators.
The reasons for this are practical. Just because a customer cannot recall their home address (they may have recently moved), enters random characters into an application field (they may have leant on the keyboard), or spends longer than average answering questions (they may have been interrupted), does not automatically signal a decision-making limitation.
Instead, online lenders need to identify constellations of indicators (e.g. multiple problems with understanding or other factors in the same application), or outliers (e.g. a customer entering their date of birth incorrectly on 15 occasions).
Doing this will reduce the ‘false positive’ rate, and improve the identification of actual need.
4. Think about support
Identification without support is meaningless. Consequently, lenders should consider what support is provided to overcome a decision-making limitation, and how this is delivered (e.g. within an online application, outside the loan application but still in a digital channel, or via telephony or face-to-face channels).
5. Think real-time interaction
Decision-making limitations need to be spotted in real-time – otherwise intervention and support opportunities will pass.
Fortunately, online lenders have already set precedents here – with the most common example being an ‘inactive application’ triggering a web-chat ‘help box’.
While some lenders may also consider asking customers if they have a decision-making limitation during an online application, this is not straightforward – indeed, people with such difficulties are often not aware of this.
Similarly, although lenders may consider using an interactive test in an online application, this could run counter to FCA guidance, as well as deterring customers from applying.
This blog aims to provoke discussion, but is not a ‘think piece’ – online lenders are already investing in this area, including our own year-long project with a major digital player.
Consequently, while online lenders may not have face or voice contact with their customers, they may soon be using another ‘sense’ – data – to see and hear their needs just as clearly.
Read our guide on online lending, and find out about our work.