Digital Debt Collection Trends in 2023 - From Culture to Strategies

I recently participated in a roundtable discussion organized by Germany’s gi Geldinstitute, where we explored debt collection trends and best practices. While the focus was on Germany, the ideas and direction are common across many countries in Europe and beyond, where high inflation and energy costs, along with other factors, are driving more people into delinquent payments.

Joining me for this discussion were Thomas Franke, Head of Debt Collection Service at the cooperative TeamBank, and Sven Krüssel, Head of Customer Service at VR Smart Finanz, where he looks after the self-employed and small to medium-sized enterprises (SMEs). Together, we explored the current state of digital receivables management at banks and our forecasts.

How Is Debt Management Culture Changing?

One of the changes in German collections is that many companies are trying to make the language more understandable to consumers, and to present themselves as solution providers. This is also effective because it does not put customers in a defensive attitude. You want to get the customer out of a crisis, and there's no use being too rude. One firm, for example, describes itself as the “Ikea of collections” and addresses its customers by their first name.

Lenders are also taking a lighter touch when in the past they might have sent someone into collection earlier. Especially given the pandemic, there are a wide variety of reasons why someone falls behind in payments, and many dig their way out of this situation again and again. Cancelling their account is to the detriment of both sides. That doesn't mean that you have to make one deferral after the other, but it’s better to take specific considerations into the decision logic.

What's Coming in the Next Few Months?

The panelists generally agreed that the next few months promise more difficulties with customers. This is prompting companies to give customers more ways to digitally deal with their situation, and avoid collectors calling on customers as the first or sole recourse.

Nobody in debt likes to be called. People appreciate it when they can deal with the topic privately and anonymously without having to justify themselves to anyone. For example, we see that people who use self-service portals to deal with their collections keep a much higher number of their promises to pay than those who just tell a collector on the phone that they will pay. It doesn't mean you get higher agreement rates, but the people who do promise are more likely to pay.

Digital processes always work when a problem can be standardized to a large extent and the entire dialogue tree can be anticipated. But if I'm in a specific situation where that can't be covered, then I need a personal contact. Conversations that can be standardized in call centers last 90 seconds or two minutes, because nothing happens there except "When can you pay?" But when you get to customers who have structural problems and can't get by with a deferral, then you have to talk.

How Far Is the Industry in Digitization?

Omnichannel communications is the goal for digital collections, but the industry is far from the standard adoption. In part this is due to banks’ interpretation of data protection rules. There are banks that say: I can send an SMS, but not say who I am, because that makes it clear that there is a contractual relationship. Other banks, on the other hand, send an SMS with the sender, which account is involved, and then it says that you are in arrears. Both are within the same legal framework, but the perception of what is allowed and what is not is different.

As Sven Krüssel noted, regulation is the big issue here because the banks are still unsure of what the rules allow. This means the data protection officers often want to play it safe with a conservative view. But they don’t bear the opportunity cost.

I recently spoke to the team leader of an American bank that uses artificial intelligence in voice dialog systems. He told me that he uses the turnover in his call center to cut jobs. Now chatbots are increasingly taking over the standard calls. The conversations that are left over and can't be automated are much more intense — these are the conversations that used to be escalated to team leaders. For these he hires new people with a more qualified profile who also earn twice as much.

What Are the Weaknesses of Automated Processes?

A weakness in automation is when you run the whole thing as just an IT project. After all, digital communication is also communication: you have to make processes lively and develop them agilely, then it works. When designing the process, you have to understand in detail how successful the individual dialogue steps are.

Then you should make adjustments and, for example, review it once a month in the dialogue team. What do I want to change? If you can adapt and experiment, that helps a lot.

You also need to avoid corporate jargon and legal language, and situations when the customer is offered options that are not of interest or are irrelevant, but also when too much is offered. Last but not least: there must always be an emergency exit in a digital conversation. How do I find my way back?

How Can You Improve Decision Trees for Automated Options and Discussions?

It is easier to develop complex strategies when you are able to break down the decision into parts. You shouldn't pack everything into a huge decision tree, but rather break it down into parts: what message do I want to convey, through which channel do I do it, at what time do I do it? If I can break it down into components, then it’s easier to maintain, and someone can still understand the logic in two years' time.

An example: I get the same six letters every quarter from my insurance company, on supplementary dental insurance, daily sickness benefit insurance, etc. In the last few years I've probably received the same letters 300 times. And I ask myself: how come? How come you get a customer experience that doesn't make sense? I suspect it’s because of the large number of possible combinations of products, customer types and contact channels through which their customers can be reached. You can only deal with the huge amount of combinations by breaking them down into manageable problems — you can't build decision trees for a million cases. How do I get out of combinatorial hell? The very existence of the buzzword hyper-personalization shows that it is difficult to get personalization right, otherwise we wouldn't need a separate word for it.

Where Do You See Digital Receivables Management in Five Years?

Not in five years, but in ten years, the promise-to-pay components will be automated in the dialog system. These systems will be able to communicate with natural language and what cannot be automated will be passed on to highly qualified employees. And the nice thing about automation is that you can simply ramp it up in the event of a crisis.

You can read the full transcript of our roundtable in German at gi Geldinstitute.

How FICO Can Help You Improve Debt Management

• Read the post Digital Collections: 7 Things That Can Make You More Successful
• Discover Can Digital Channels Work with High-Risk Collections Accounts?
• Read the post Even in a Crisis, the Best Collections Strategy Is to Avoid Collections 
• Find out why digital-first collections means thinking like a marketer
• Explore FICO debt collection solutions