While credit providers have used consumer profile analytics for many years, the debt recovery sector has been slower to adopt these highly sophisticated techniques. However reliable analytics modelling can have a significant impact on your bottom line, through enhanced collection workflows, improved effectiveness of collections strategies and increased overall productivity. by Nicholas Harrak
While credit providers have used consumer profile analytics for many years, the debt recovery sector has been slower to adopt these highly sophisticated techniques. However reliable analytics modelling can have a significant impact on your bottom line, through enhanced collection workflows, improved effectiveness of collections strategies and increased overall productivity.
by Nicholas Harrak
Workflow automation, productivity and effectiveness in debt recovery are critical issues facing our sector today. But, unlike credit providers who have long used consumer profile analytics to determine credit worthiness, the debt recovery sector has been slower to implement sophisticated analytics modelling that can predict both propensity to pay and the most effective recovery strategies.
Admittedly this is not a quick or easy process to implement. It took recoveriescorp two years to develop a strategy and build our internal capability to analyse the predictive characteristics of debtor profiles and create propensity to pay scoring models.
We now regard analytics as absolutely essential to our business sustainability. Internally, this has improved our workflow strategies, automation, productivity and effectiveness in debt recovery. Externally, this has added significant value to our relationships with credit providers who already run their own leading-edge analytics programs. Analytics have also enhanced our relationships with debtors, as we engage more effectively with them to achieve better payment outcomes.
These are the steps taken to establish our program and, importantly, to maintain its relevancy.
There are many elements that can predict propensity to pay, including traditional ones such as the total amount due and postcodes demographics. If you’re dealing with thousands of accounts each month, there’s a vast amount of intelligence about the outcomes of your collections processes that you can harness and analyse.
The key is develop your organisations internal capability to analyse and create modelling that will give you answers to questions such as: what is the most effective strategy for collect these debts? How do we segment the portfolio? Where do we invest resources? At which point should we make a telephone call, send a letter, or commence legal action?
Without accurate and effective modelling, it would be inefficient and uneconomical to perform this type of analysis on every account. But with an effective modelling strategy, you can accurately and efficiently segment and effectively treat large volumes of accounts.
Analyse your data and challenge the outcomes. If your modelling tells you that Profile A are better payers than Profile B, ask yourself; how can you improve your outcomes on Profile B? What type of skill set should you allocate to Profile A and B?
A revealing case study for us involved a review of a business as usual letter that we were sending to debtors at a particular point within a process. The return generated by this letter had led us to believe that the letter was very effective.
When we challenged this process, our modelling showed that two debtor profiles responded positively to the letter, and two did not. The result was not what we had expected. This type of analysis allowed us to segment the profiles, continue sending the letter where it was effective, and develop other strategies for the two non-responsive groups.
The above case study shows the importance of using data to challenge your business as usual outcomes to segment your accounts and align them with the right collection strategies.
Today it’s about connecting with the customer in the way that they want to be engaged. For example, some of our customers will respond positively to a collections strategy that utilises our self-service portal where they can enter their own payment arrangements
Others will respond to a strongly worded letter, but it’s the type of letter that, if sent to a different customer profile, will result in very negative feedback that’s time consuming to deal with and may impact negatively both on the collection agency and the client. An effective analytics strategy will show you how each profile type responds.
The benefits of analytics led us to revisit our purpose. Our goal is to ‘Connect. Engage. Succeed.(c) with our own people, our clients and debtors.
Modelling shows us how best to do this with debtors, by aligning skill sets with different profile types.
Once you understand how you should approach each category, you can tailor your strategies only where necessary without disturbing those areas where the modelling shows that your strategy is effective.
This works equally well for the credit provider who may be very effective at collecting debts from certain profiles, however find other profiles challenging. An effective analytics and modelling strategy will assist the credit provider to determine the profile of accounts that they should retain, outsource or send to legal action. It also provides valuable information which can be incorporated in the outsourced agency’s strategies.
Once the modelling is working, it is essential that you adopt a regular framework to maintain and calibrate the data to ensure that the model remains accurate. At recoveriescorp we have resources allocated to perform this task on a monthly basis, other organisations have different intervals.
For example, payments by a particular profile might begin to degrade. You will need to identify this trend quickly, and look at your processes to determine why it is occurring, what sort of impact it will have and what you may need to change.
This will enable you to stay ahead of the game, rather than discovering in six months that a trend has changed and your collections have reduced as a consequence.
Maintaining the data also means you can challenge your own processes and prove new strategies on a percentage of a debtor profile. The remaining percentage becomes your control group, ensuring that you can test a number of strategies and compare outcomes before making long-term changes to your workflow treatment strategies
The global financial crisis and the Queensland floods are key examples of economic and environmental factors that can influence your existing modelling. For example, during the GFC, our modelling showed that households with high disposable income were impacted differently to those with low disposable incomes. Accordingly, different collection treatments were required to tailor our customer engagement strategy and reduce the potential adverse impact for our client partners.
Likewise, the Queensland floods very quickly changed the profile of existing debtors in Queensland and also affected the profiles of new debtors in that state. Once again the modelling required enhancement to accurately reflect this change.
Most clients are involved in one industry only, such as banking, communications or insurance. We are quite privileged to manage a number of accounts across a diverse range of industry segments. This is where analytics can add significant value. While the demographics, predictive elements and trends vary across industries, we are sometimes able to identify trends in one industry that will impact on another.
For example, our experience shows that the usual order of default begins with credit cards first and mortgages last. So if credit card delinquency increases and those customers are not rehabilitated, that’s a red light for looming mortgage defaults!
There’s a vast amount of commentary out there about big data and analytics, much of it very theoretical, so there’s a tendency for some people to become overwhelmed.
But once you’ve gathered your data, it’s essential to start determining how it can be applied to improve your business. You’ve got to act, otherwise all this information may be wasted. This comes back to challenging your processes.
We have found that improvements of 10% to 20% are achievable, either in recovery rates or in terms of productivity or direct cost savings.
That’s a very significant gain in an era when business sustainability is demanding increased workflow automation, greater efficiencies and more effective collections strategies.