High unemployment rates of the post-pandemic era in many recovering regions and increasing inflation rates are signaling an economic recession. With the pressure of increasing energy prices on consumer cash flow and households in many advanced economies facing a cost of living crisis, collections managers know it’s time to transform their capabilities to make their operations more agile and resilient to maintain, if not to improve, the collections performance.

As in the transformation of other credit risk decisioning areas, collections managers have different pathways in front of them based on the level of maturity of their available collection kit. However, it’s possible to discuss 5 key areas with the greatest impact on collections performance in order to start developing a successful roadmap:

Bring in self-service capabilities

Digitalization penetrated into many marketing and customer originations functions, however, in collections, there is still a way to go before it matures. Start considering how much your operations are able to utilize the organization’s existing digital channels. Is your customer services department able to inform a delinquent customer when they reach out to them for other problems or requests? Are you able to notify the customers about their missing payment when they use an ATM for another purpose? What about when they use the internet or mobile banking? How automated is your SMS channel? Are you able to use interactive messaging to capture customers’ payment promises? How automated are your email and letter sending processes? Are you able to use the push notifications for outbound collections?

Later you can consider how automated is your outbound IVN service. Is your IVN able to capture the customer’s payment promise? How about your inbound IVR service? Is your IVR able to understand with what purpose the customer is calling and let the correct collections team answer that call? How automated are your inbound team services? Are they able to solve the customer problem over the phone or are they routing to another team in collections?

Then you can start making more digital contact channels available. Are you able to use chatbots as delinquency reminders? Is the chatbot aware the customer is delinquent when the customer contacts for another service request? And then you should design an omnichannel contact strategy to use all these digital contact channels in the most effective way with minimal deterioration to customers’ experience and loyalty.

Leverage customer360 data for a better understanding of customer payment behavior

Debt management sits at the end of the customer lifecycle - if we ignore the insight feedback function to originations strategy. This means a significant amount of customer data is collected during marketing, origination and customer management activities before the customer account lands in collections. How much of this data flows into the collector screens?

Are your collectors able to see what is the last time a customer spends money or makes payments into their account? How much deposit or savings the customer has in other accounts? What about credit bureau information? Has the customer other debts in the market? What if you’re outsourcing early collections activity, are you able to see what collections actions are taken earlier? Is your organization utilizing open banking? Are you able to see customers’ accounts in other banks? Customers’ spending, saving and payment behavior have changed significantly during and after the pandemic. Are you able to use the most recent data to understand and segment your customers?

Use advanced analytics for more granular segmentation and treatment strategy

Which customer to contact when and at which frequency have always been the most important questions of collections. In collections terminology, we name the answer to this question as a treatment strategy. Treatment strategies may vary from a single contact in 30 days to many, commonly limited by the regulations in many countries. Because of the increasing importance of keeping a balance between compliance and performance, it’s getting more and more important for collections managers to know which set of customers should be treated with which strategy.

Collections strategies must be designed at the delinquency bucket level and both effectiveness of the strategy and the productivity of the operations should be considered. A strategy that will require an outbound call contact with all the accounts in that delinquency bucket and will utilize all collection resources is definitely not a good strategy. Bucket-level strategies are most effective when a collection model dedicated to that bucket is used. Delinquent customer behavior would be different in each bucket and hence predictive data characteristics or the weight of those data characteristics would be different as well.

The use of more granular segmentation and differentiated treatment strategies would lead to more effective use of available collection tools, reduction in manual workloads and hence better collection capacity management.

Many collections managers today still rely on the predictive power of traditional collections roll models or expected collection amount models for customer segmentation. Isn’t it time to use machine learning models either for benchmarking existing model performance or developing a challenger model to optimize performance? Machine learning models excel at capturing non-additive relationships in data. Organizations that use machine learning models report significant improvement in the predictive power of the models and are hence able to generate more granular customer segmentation.

Develop new payment relief programs

At the beginning of the pandemic, we observed a wave of customers with financial difficulty applying for payment relief programs. We have also seen the struggle of financial institutions to meet this demand from the deterioration of customer service KPIs. Why did this happen? First of all the process of capturing and implementing a payment relief request, whether it’s a restructuring or payment delay, is not automated in many organizations.

The lucky situation with the pandemic was that the majority of the programs were government initiatives and financial institutions were not mandated to check the eligibility or affordability of the customers. However, we have observed the consequences of that in the delayed recognition of losses after the fact. In the next wave, financial institutions will need to make it smarter and with a better customer experience. In order to do that, personalized payment relief programs need to be introduced. Every customer’s financial problem is different and you can’t solve them with a standard approach. For example. you can’t offer an extended payment plan for a farmer who receives the next payment in 6 months. Payment relief programs tailored to different financial difficulties of the customers need to be introduced.

Modernize collections decision technology to innovate at scale

So far, we discussed what data and analytics capabilities can be introduced to collections and what process automation and strategy improvements can be considered. However, it’s important to realize that all these changes can only be done with a strong enabler collections decision technology.

In order to process customer 360 data, you need strong data ingestion, orchestration and aggregation capability. To be able to apply machine learning models with the consideration of responsible and ethical AI, your technology should cover the entire model development and governance process integrating with other applications. You can only develop new payment relief programs and rapidly deploy them in operations with a flexible strategy management tool. To be able to monitor the performance of the granular segmentation and treatment strategies in detail, you need a good data visualization and reporting capability.

Collections managers need to work with reliable technology partners which will deliver these capabilities in a single, integrated platform in order to orchestrate them all, do it at scale, keep operations open to future innovations and at the same time reduce the total cost of ownership.

Learn more about SAS® Intelligent Decisioning for collections and SAS® Risk Modeling.



About Author

Zeynep Salman


Zeynep Salman is a credit risk professional with direct experience managing originations, customer management and collections teams for consumer and small business portfolios. She joined SAS in 2022 and is currently leading risk decisioning advisory activities across EMEA. Zeynep is passionate about driving automation, seamless customer experiences, convergence of credit and fraud evaluations across customer lifecycle, AI driven customer engagements and working with clients to support near and long-term strategic roadmaps to drive value. Before joining SAS, Zeynep held many key roles at financial institutions such as Citibank, HSBC, Toyota Finance and UniCredit, as well as software vendors such as FICO.


  1. Nice read! When experiencing economic downturn, having better collection decision engine becomes especially important.

Back to Top