The role of behavioural insights within collections

Our thanks to Steven Johnson, Behavioural Insights Advisor with Voicescape, for taking the time to take part in the interview upon which this article is based.

Behavioural insight is the intersection between data science, technology and behavioural science. These three elements work together to design and implement targeted interventions that influence behaviour. This may include interventions that prevent missed payments or encourage the customer to enter into a payment arrangement.

Steven has undertaken a statistical analysis of data and then backed that up with reports of who is more likely to fall into arrears or break the agreement. He then designed interventions based on these to create improvements.

Data science

Data science is used to target and tailor the intervention. It allows you to understand the type of case you are dealing with and the associated risk for that case, using data on past behaviour and demographics. This creates a risk ranking and what the risk is associated with.

Technology

The technology is around automating and personalising the intervention. You can set up the algorithm to analyse a particular type of risk and then determine what resource is best to manage it. For example, if the circumstances are low risk and your team has low resources at that particular moment, the best resource would be to send the customer an automated intervention that could include a voice contact or text message.

In a very high-risk case where the team has medium capacity, the algorithm might recommend a face-to-face visit and other interventions in between. To work effectively, the algorithm should be self-learning, so that if you send a voice message, the AI (artificial intelligence) will decide which of the pre-prepared messages is most appropriate for those specific circumstances.

It is important to avoid using a “one size fits all” approach to technology.

Behavioural science

Behavioural science is around creating steps or using specific measures or working to make the desired behaviour more likely. It is all about influencing the customer’s behaviour.

Results of trials

In 2015, Steven started trials with Capita plus ten local authorities and social housing associations. He designed a series of nudge-based interventions, set up randomised controls and included evidence-based practice. The randomised controls were scientifically based on behavioural science and behavioural insights.

Nudges are small changes to an existing escalation pathway and are used early in the process, such as a message to say, “don’t miss your payment, catch up quickly.”

One of the changes made was to add one sentence to a letter which said “99% of people in your neighbourhood are paying on time.” This is tapping into social norms – the customer has been told most people like them are doing this, so they are more likely to do the same.

This change led to a 9% increase in payments.

Another trial was the sending of thank you messages to customers, such as a text to thank the customer for keeping up with payment arrangements. There was a 19% increase in the number of customers who kept up with their payments, demonstrating the power of gratitude – a “thank you” gets more people to do more.

Communication was another area that was assessed during the trials, with a view to reducing the cognitive friction caused by reading an unwelcome letter. They introduced basic information design principles: understanding how the eye moves around the page, what colours, fonts and information hierarchy are most effective. They considered how to get the envelope opened - a brown window envelope is seldom good news - so they changed the colour of the envelope and used a handwritten font for the address.

They generated an 8% improvement just by using a different envelope. They also saw a 24% increase in customer engagement by redesigning the letter using plain language and improving the visual appearance.

These results were beneficial and showed a real improvement. However, remember that evidence of what doesn’t work is as important as what does.

Doing nothing

When trialling changes, do ask the question “what would happen if we did nothing?”

This trial was run several times and one of the control groups got no early-stage interventions, but there was no impact, negative or positive, on payment. As a result, one organisation removed their early-stage escalation pathway and saved a quarter of a million pounds, although they are still running interventions later in the collections process.

On that note, it is worth bearing in mind that escalation processes can be too ‘trigger-happy’. For example, the customer’s account is £10 in arrears on Monday morning, so the organisation starts the escalation process, without considering that those customers were probably going to pay anyway. This results in a waste of money on the intervention, a poor customer experience and potential damage to the relationship.

Moving from tactical nudges to behavioural science

Trials show that a tactical nudge is fine, but it can be amplified by the more strategic approach of behavioural insights.

This looks at areas including:

• How to build productive relationships with customers
• Perceived fairness and perceived competence
• How the customer feels they’ve been treated, which has a massive impact on their future behaviour

How to implement a behavioural insights strategy

You need to build the service and design the customer experience from the ground up with behavioural science in mind and the tactical stuff sitting underneath, filtering through the process and day-to-day interactions.

You may have evidence about nudges that work now for you but remain open to developing these and please don’t take them off the shelf by copying boilerplates. It’s not just about tactical quick wins.

Then run random control trials to check the impact. Don’t underestimate the complexity of a random control trial. There is a great deal of technical rigour in the design, sampling, randomisation and statistical analysis.

And, of course, you will need to invest in the technology to deliver it.

Finally, when implementing a behavioural insights strategy, design your system to ensure that the relationship is not damaged between you, the local authority, and your customer.

Voicescape is a technology company whose mission is to couple rigorous behavioural science with automation technology to help social landlords and local authorities connect with their communities, developing more effective proactive customer engagement and gaining operational efficiencies.