I was privileged to join Bruce Curry on a series of two digital roundtables recently, where 13 experienced minds of the industry got together to discuss the current and future situation around collections and recoveries during the current debt tsunami.
The attendees had a combined experience of over 300 years across analytics, credit risk and operations across Europe, UK, the Middle East and Southern Africa. Most were from financial services, with portfolios ranging from small loan products to very large mortgage and vehicle asset finance across the consumer and business (SME) sector.
This crisis has been unique, due to the dynamics of virtual operating models and in the spending patterns for those affected financially by the crisis and those not. The impact of this crisis is expected to be even bigger than the 2008 crisis, with NPLs rising 6 to 8 times higher than normal in some areas.
Here’s what participants said about managing debt collection in the pandemic this year — and what they see for the future.
Topic 1: What blind spots do you have in your customer risk and treatment profiling?
- Across all markets, regulation moved quickly around payment holidays and moratoriums for both the consumers and SMEs. Bureau data became a challenge as many of the leading indicators became redundant since people ‘froze’ in time, which placed risks on many of the account origination and account management scores.
- Income reduction was not well known. Whilst many organizations had the ability to use transactional data for segmentation, that didn’t provide a view of consumers currently in work, who have been furloughed (reduced working time) and would not return to work at the end of the period. This provided no known stability of income in the near term.
- The ability to obtain accurate industry sector employment data became a challenge. This was critical to understand income risk and job redundancy across the sectors.
- The payment holidays didn’t follow the traditional forbearance guidelines, as almost everyone qualified. This left gaps in understanding the customer’s situation for the next action once the period is over.
- Certain portfolios performed much better than anticipated due to the impact of strategic defaulters (or “buffer gatherers”) — people who didn't need the payment holiday but applied for it anyway, to gather funds. When the programs expired, they simply continued paying, which resulted in better performance seen compared to what would have been expected if they were hardship cases. This placed additional doubt on whether this can be attributed as a positive or just a “calm before the storm”.
- In terms of capacity management in the operation, this provided a challenge for the best size to cope effectively with the amount of work in place. One participant related his experience as the following: “Hope for the best, prepare for the worst.”
Some of the suggestions to mitigate or better understand these blind spots: -
- Digital channels in the past 5 years have been used to “tell”, “go” and “get” customers to conduct collections activity. While this is a very cost-effective method, these channels can be used for so much more. This helps build your view of the customer (digitally) and pass through information that allows you to become more proactive. Why search for data elsewhere when the data can be obtained directly from the source? A change in the messaging can allow you to capture information, such as:
- Reason for delinquency
- Duration of impact
- Disposable income
- Revise your restructure toolkit. Having just one or two options are no longer enough. You are now aware of the “strategic defaulters”, so what is the best way to cater for them? How do we solve for customers that “use the magic word Covid and receive a payment holiday without any questions”?
- Further to “strategic defaulters”, make an attempt to start having sight of collaborators vs. non-collaborators. Customers across different risk grades took up payment holidays – who will be willing to communicate with you around changes in their circumstances and keep to their arrangements thereafter? Understanding whether people are willing to collaborate with you has become as important as understanding the risk of the customer.
Topic 2: How are you using advanced analytics?
- There were various degrees of advanced analytics capabilities across the countries, with some at an infancy level. Many organizations stated that performing stress testing on their back books was very critical to understanding the robustness of their models.
- Organizations sought external data sources to provide guidance on where their portfolios were heading. Metrics such as salary deposits, levels of unemployment data, industry codes, house and commercial estate prices and other alternative data, such as kilometers travelled, were leveraged on the secured portfolios.
- Some organizations went further, mainly in the UK, and used Open Banking data to segment customers during the period. Income and expenses were a key attribute that required attention in this space.
Some of the suggestions on the levels of analytic sophistication and how this could be leveraged were as follows:
- Portfolio profiling and scenario modelling were seen across the board, but most of the analytics were descriptive, and more rarely predictive. Not many had taken analytics into the prescriptive space, utilizing mathematical optimization to determine the best approach to the following:
- Capital adequacy – what is the best offer to the client?
- Capacity optimization – how many agents do you need?
- Contact optimization – which channels and when should they be applied?
- Next best action – should you work, place, sell or soak the account?
- A huge amount of data can be sourced by omnichannel communications, as discussed above. Can any of it be leveraged further using the open banking data?
Topic 3: How easy have you found bringing real-time data into your automated processes?
- Bureau data has always existed in most organizations and continues to remain accessible.
- Open banking data is a new concept and still in trial stages and thus has become more of a batch process.
- Frequency of payment data updates has increased but this is not used extensively at the moment.
- There are challenges around bringing through government data due to frequency of updates and availability in some regions.
- Self-service channels that are operating 24/7 contain data which is obtainable but large.
- With the move into more prescriptive analytics, organisations are seeing challenges. They are able to run optimisation models but due to the size of the data and the processing requirement for the changes, these models are currently being executed in a batch manner – when they can. Many have collections systems that cannot support the delivery of prescriptive analytics.
Suggestions around methods of collecting the right data and automating it were as follows:
- With the rush to collect new data around the client, ensure proper planning is in place.
- Is this a short- or long-term problem?
- Are these static requirements (once-off)?
- What is the frequency for updating the data?
- Ensure operational teams understand why data is being collected. This assists in the conversations with the clients and facilitates deeper conversations in determining collaborators vs non-collaborators.
- Platforms have become more advanced. As the collections system separated from the core banking system, two things started to innovate/evolve faster than the C&R system – digital analytics and omni-channel customer engagement. Without a configurable workflow management system, it will always be a tremendous challenge to support these.
- Data marts are inflexible, and many organizations are on the move to cloud solutions. Cloud provides a fast turn-around and does not require client processing and infrastructure from the organization. Data streaming provides additional benefits for the organization for the use of vast amounts of data on the go. Costs are always a concern in this space, and planning around which data is required real-time will be important.
Invention or Adoption?
There was comfort that, across the regions, all are facing the same problem – there is “no magic answer” to this problem. But there was a keen interest in understanding further methods in leveraging analytics and optimization. Data and tools were noted as critical dependencies for success. The focus on digital omnichannel communications was key even further into in-App self-service.
Finally, focus has to be shifted to educating customers during this time, as many people now in arrears have never been in collections before. This is critical to managing the debt tsunami.
Bruce summed up the current state by noting, “People say need is the mother of invention. That’s not where we are today — the capabilities needed to improve collections and recovery have already been invented, and proven to be successful. This is a case of need being the mother of adoption. Failure to adopt what is needed will cost far more than the investment.”
We recently held three webinars showing how to user these tools to improve debt collection today. I urge you to watch them: