Data sharing has had a fair amount of negative press with the Facebook/Cambridge Analytica case. Consumers are often sceptical when asked to share their personal or financial data with an organisation.
Data sharing certainly has its challenges but it also has potential advantages for the consumer.
On 18th April, some of the UK’s leading academics in personal finance, data sharing and vulnerability gathered at a roundtable dinner hosted by Flexys Solutions in Cardiff.
Researchers and industry practitioners from a number of sectors gathered to discuss how lending organisations can build trust and transparency with consumers and articulate the value of sharing their data.
Trust and Transparency
Professor Sharon Collard, Director at the Personal Finance Research Centre, University of Bristol, explained that from the PFRC’s research, the most likely point in the process at which firms can identify vulnerability is when the individuals disclose it themselves. Therefore, the challenge is getting people to engage. In terms of where things currently sit, engagement with customers occurs more often towards the end of the debt recovery process.
PFRC researcher, Jamie Evans explained that:
- 44 per cent of respondents in PFRC’s survey have told their bank about their mental health condition
- 38 per cent have told other lenders
- 63 per cent have told a money or debt adviser
Debt collection staff receive on average 15 disclosures per month of a serious physical illness (of the customer or their family), 12 disclosures of a mental health problem and 9 bereavement disclosures
Collaboration and Culture
One difficulty Professor Collard raised with disclosure was the necessity of multiple vulnerability disclosures for many customers. People who are experiencing physical or mental health problems or going through a temporary, difficult period in their lives, find it challenging to repeatedly share information about their circumstances. The emotional cost of this repetition is high.
What is clear is that the infrastructure we have to deal with vulnerability needs to change. A centralised database similar to the Ofgem Priority Services Register, which enables vulnerability data sharing between energy companies, could be one answer.
Ethics and Infrastructure
When building trust with the consumer, the critical part is maintaining an ethical approach, said Bola Karimu, of the University of the West of England. If a centralised industry database were to be founded, there would need to be procedures to ensure companies sign an ethical agreement to agree not to misuse data, which would help build customer trust.
Sharing data around a vulnerable situation can be a starting point for an outcome-focused conversation between the customer and the organisation. The infrastructure needs to be in place for customers to communicate using their preferred channels.
A valid point was raised during the discussion that vulnerability data sharing might not be streamlined enough. Different departments in a bank, for example, may not share vulnerability data effectively – people in the savings department may not know their same customer has declared details of their vulnerability to the current account department.
Technology and Machine Learning
The panel discussed how technology can assist with vulnerability detection. Dr Karimu said that using machine learning to detect vulnerability risked removing the human element which is such a key part of the interaction. Equally, he advised that if we can identify potential vulnerability characteristics earlier in the customer communication process using advanced technology, then advisors can intervene more quickly to extend help.
Potential vulnerability can be better understood by deploying machine learning to detect a variety of signals and indicators, said Flexys CEO, Jon Hickman. When the aggregated data suggests a customer’s vulnerability likelihood has reached a designated threshold, an agent can be automatically prompted to have a conversation with the individual, bringing back the human involvement at the right time. At no point will a machine learning algorithm make a unilateral decision affecting a customer. The technology is there to support human decisioning, not replace it.
The combination of advanced technology, rigorous regulation and a high standard of awareness and training within collections is changing the experience of customers in vulnerable circumstances but the panel felt this varied across sectors and across organisations. Rather than pursuing a prescriptive approach, they agreed that the most productive way forward is through collaboration – learning to engage and collect data in the right way, and using and sharing it within carefully defined boundaries, with customer outcomes firmly at the centre.