Skip to main content
Analytics for the Repo Man

(Posted by Guest Blogger, and someone who's never yet spoken with someone from a collections agency - at least not as a "client" - Ian Turvill.)

We all know that analytics are used extensively in financial services, and particularly in areas such as credit and insurance, to make predictions about consumer behavior.  It is perhaps less well known that one where analytics has historically played a strong role, and where its impact is still growing, is in the business of collections and recovery.

Making decisions about collecting bad debt is complex: you have potentially many debtors, each with relatively small amounts outstanding.  (By which, I mean that the amounts outstanding can be small relative to the potential legal costs of going after payment through the courts.)  So, there is a fundamental resource issue involved: how do you align your resources against collections activities to get the greatest possible rate of recovery on your portfolio?

This is, of course, a problem that lends itself very well to predictive analytics, decision analytics, and Enterprise Decision Management as a whole.  This is borne out by an article "Knowing the Score on Productivity" that appeared in the June 1 edition of Collections and Credit Risk magazine.  (This is a link to the magazine.  Unfortunately, a subscription is needed to see the whole article.)

Here are some of the more tasty excerpts:

"Collection agencies are using scoring models to help determine how best to apply resources when crafting treatment strategies.  As collection agencies put a higher premium on boosting productivity and cutting costs, scoring models are becoming more important strategic as well as tactical tools used to meet those goals."

Excellent.  Just what we like to hear about the use and importance of scoring.

It continues:

"Such models are being used to provide such key metrics as the cost per contact for an account, portfolio yields, liquidation rates and potential recovery rates.   This information goes well beyond the traditional use of scoring to determine the propensity of the debtor to pay and provides collection managers with insight about how to effectively apply resources to work specific segments of portfolios."

This is very interesting.  As the article indicates, collections firms are very imaginative about the range of predictive indicators they can use to guide their business.

"Among the new ways agencies are using scoring data is to determine whether skip tracing needs to be applied to a specific account, whether an agent needs to be instructed to follow a script when talking to the debtor, and even which agents to assign to specific accounts based on the characteristics of the debtors. Scores also can help collection managers determine the frequency of contact."

What makes the article particularly interesting, from an Enterprise Decision Management point of view, is this final statement, with added emphasis from me:

"The two drivers that are making it possible for agencies to use their scoring models in a more sophisticated manner are better and more extensive data that can be fed into a model and improved IT architecture for their operating platforms.   The two go hand in hand. As agencies break down the information silos within their IT platforms, more data about the performance of the call center and the entire agency itself can flow across the enterprise.

These data include performance metrics on how previous treatment strategies reduced roll rates, increased recoveries and right-party contacts, and the average cost per contact method and recovery.

By introducing scoring models across a wide variety of performance indicators, including agent performance and the effectiveness of agent training programs, collection managers can gain a fuller understanding of how to apply their resources to an account in order to optimize their return on investment. In doing so, treatment strategies are based less on individual judgment."

My one suggestion for this article is that it should consider how the comprehensive decision-making capabilities that are now displayed within and across collections operations can also be extended back to the bank that made the loan in the first place.  The FULL realization of ENTERPRISE decision management would take the data on loan performance and feed them back into the originations process.  That said, the advances that the author, Peter Lucas, describes represent some very large steps forward, so I certainly won't fault him.

For those of you with a greater interest in this topic, I also recommend "Lifting the Rocks, Finding the Gold", a book by two consultants who focus on an analytical approach to managing the C&R process.

related posts