In the News: Data from Social Media has a Place in Credit Decisions

June 10, 2015

Author: Paul Thomas

Abstract

Traditional sources relied up by lenders to assess the risk involved in approving a would-be borrower’s loan application can be supplemented with data gathered from social media sources. There is no question that the data available from such sources is wide and varied. It also offers a different insight into an individual’s behaviour and lifestyle than can be gained from both information offered up by the individual in their application and from credit history sources that can be used to produce a credit rating. How this data can be used and if it should be accessed in the first place are other questions entirely, and proof that the data can make a substantial difference to the credit assessment outcome is yet to be gathered. In this article, the author examines what social media offers to the risk assessment process and the benefits and challenges of using it.

Download the pdf 

Introduction

The potential social media has to be a source of information for financial services companies making money-lending decisions has been debated in the industry and the media for some years. Undoubtedly the public is sceptical; concerned even. There is a general lack of understanding over what data could be used in this way, how it’s accessed and, in fact if it should be. The likes of Facebook, Twitter and LinkedIn as platforms for information sharing are inherently public; they are internet-based and multi-directional. Yet passions run high that the data they contain is personally owned by individuals and that they are in control of it.

Elsewhere, we see that businesses do make use of social media. For marketing purposes, companies monitor social media and use information from it to proactively engage with customers, aiming to improve customer service, address customer complaints and ultimately to retain business.

There are benefits to lenders and borrowers alike of bolstering information provided in loan applications with social media data. It can provide genuine demonstrations of behaviour that point to an individual’s capability and likelihood of meeting repayment terms. But the data generated by social media is vast and unstructured. The challenges around making it usable are many and significant.

Then again, for the next generation of young adults starting out in life and just beginning to create their ‘credit footprint,’ online information channels may be the most sensible source of data. They are the channels financial services companies’ newest customers engage with most readily; increasingly to the exclusion of others.

Business drivers

There are definite business drivers for lenders to use social media as a supplemental source of data for making credit decisions. Where little credit history exists for an individual, social media can provide useful information. Being able to get at this information could speed up the credit decision process, and this is something lender and borrower alike want to see. Lenders are striving to know their customers much better. In recent times we have seen a shift in emphasis of responsibility when it comes to the two-way relationship of money lending. It doesn’t rest solely with the borrower. We are in an era of responsible lending where lenders are sensitive to their part in offering the right product to suit the customer’s needs. And this is likely to increase in importance.

Lenders want all the tools they can get to help them reach a decision that sees them lending an appropriate amount with appropriate repayment terms.

Technology is playing an ever more central role in supplying those tools and sophisticated analytics are part of the kit-bag. Social media data is the next logical step.

Digital footprint

For consumers, it’s likely that we will see a gradual shift away from total resistance to the information they share through social channels being used by companies. People understand now that they leave a digital footprint. There is an element of expectation that companies selling to them will ‘know’ them. In fact, customers don’t want to receive blanket marketing communications. They have more of a sense of themselves as an individual than any generation ever has – they are the selfie generation – and expect companies to tailor their offerings accordingly. This level of marketing tailoring requires high – and wide – data input.

Importantly, people don’t tend to conform to a lifecycle anymore that could previously have been considered ‘standard’ – mortgage, car loan, pension investment and so forth. They buy homes later – if at all; they may use car pools rather than own a car that depreciates; they may invest in property for their pension. All of which adds up to a lack of credit history, a smaller credit footprint in the financial data sources lenders are used to tapping and being shut out of the financial eco-system. The latest generation have so many options when it comes to sourcing financial services; they are spoilt for choice; it’s a situation traditional lenders didn’t have to contend with before.

This generation exists increasingly online – managing their lives and their interactions; making payments; managing their money. The proactive use of social media data has the potential to alleviate the emphasis on the consumer to ‘go find’ themselves a credit history in order for the computer to ‘say yes.’

Consumer interactions with social media are already used with a high degree of success in customer relationship management. Information sweeps of online sharing channels reveal where customers are complaining about or praising companies or brands, and automation tools can generate standard responses to complainants or advocates. This is a useful demonstration of the power of the medium in giving companies a holistic omni-channel approach to their interactions with existing and potential customers.

Small businesses

While the focus on company use of social media tends to be on the consumer market, there is much potential benefit to be gained in the SME sector. Fledgling businesses often hit a black spot in the processes, systems and procedures of banks. They struggle to serve them. They don’t treat a loan application as coming from an individual yet the start-up business has no financial history to plead its case. Banks stand to lose out if they can’t find a way through this conundrum as a start-up loan is likely to lead to a current account, a credit card and a banking relationship that could be potentially lucrative and long-term. If lenders use the same tools they would use in assessing an individual loan application to assess that of a small business, they stand to make a more accurate, quicker decision.

Players

So who is in the game when it comes to exploiting these new data channels? Established banks aren’t yet in a position to make use of this information source. With processes and systems tightly linked into an infrastructure ecosystem, it is hard for them to try out something new – to test the water and see if it’s warm. This leaves them open to competition from alternative lenders that potential customers turn to because banks, unable to identify a risk level, cannot assign capital to them. So mainstream lenders face some serious competition.

For all lenders it’s a question of risk. To understand it, be able to assess it and to minimise it. For the consumer and small business borrowing smaller amounts it is frustrating that a traditional institution cannot adapt or evolve its processes to consider an individual case. Challenger banks and bespoke lenders are more able to try out additional sources to input to credit scoring, and more quickly update their systems and procedures to implement anything that works.

The type of knowledge lenders can gain from social media sources is an insight into behaviour; into how an applicant spends their time, what their lifestyle is, their habits, their career history, their aspirations, the company they keep. As a holistic, fully-rounded view of a potential customer it is powerful stuff.

Vast data

The way social media data is tapped into varies and mechanisms to get at these sources are still in the infancy of development. There are examples of social media users giving permission for their data to be accessed and used. Some organizations use surveys delivered across social networks. In this situation, users willingly share information that is then passed on to credit companies. Results from these surveys can be interpreted and fed into wider decision making processes.

Beyond these quite focused examples, the prospect of mining social media data generally is a daunting one. The world’s data is growing rapidly. 500 million tweets sent per day(1); Over two million pieces of content shared every minute by Facebook users(2) . If you are sourcing information specific to an individual, how do you know you are looking at the right information?

Data is essential for all aspects of the financial services industry. Analysis helps by quickly differentiating between relevant and irrelevant data; good and bad. Therefore the focus has to be on getting at just the core data that is going to be useful to the decision making process. For this reason, the automated collection, collation and analysis of social media data for credit scoring purposes has so far largely focused on population level information, to provide trends insight.

Benefits

For lenders, social media can provide valuable data where traditional sources yield little. Recourse to a predictive credit score generated from analysis of social media data could mean an application that would have been rejected, is approved. It is important to lenders to reduce the time it takes to grind through the decision mill to spit out an ‘approve,’ ‘defer’ or ‘reject’ outcome. Applications sitting in the pipeline are of no benefit to lenders – they are a drain on resources – or frustrated borrowers. The better lenders know their customers, the more targeted their marketing can be and the higher their customer retention rate is likely to be.

As consumers become more savvy about the power of the information they share through their social network channels, so they may be more inclined to grant permission for it to be used. If it gets them what they want – an approved loan application – it can be a mutually beneficial information share. Effectively, their social media channels become their opportunity to sell themselves to would-be investors.

Challenges

Integrating social media data analysis into risk assessment tools is not a simple process. Online platforms evolve and change their algorithms regularly. Credit scoring technology would need to keep up.

Lenders want to tap into information sources easily without needing to change legacy systems – which is expensive and time consuming – so their integration needs to be simple and backward-compatible. Adding more information sources means more data – and as we’ve seen the data out there is vast – so robust, automated data analytics are needed to turn that data into usable customer information that can be acted upon.

Then there are issues of security, regulation and data protection that need to be understood. As practicalities and questions of ethics are worked through over time, the issue of customer trust and faith needs to also be addressed. Customers need to see the benefit of the process but also be comfortable with the journey – for some, the end may justify the means; for some it may not.

Money lending has a very long history and the tools used today for the assessment of creditworthiness are well established and heavily relied upon. Social media is the newcomer to the group, but data is king so it has the potential to firmly establish its place.

Its usefulness will need to be proven over time. Assessing predictions underpinned by social media data against a borrower’s actual proven ability to pay will help lenders understand if the additional data does in fact help them make the right decisions, quickly.

References

  1. https://about.twitter.com/company
  2. https://www.domo.com/learn/data-never-sleeps-2
[if lte IE 8]
[if lte IE 8]
[if lte IE 8]
[if lte IE 8]