article banner

Data analytics to support collections and recoveries

Lenders are expecting an increase in collection and recoveries activity later this year. Nikhil Asthana take a look at how data analytics can help firms plan ahead, streamline processes and support vulnerable customers.

Last year the government introduced a number of initiatives to help businesses during COVID-19 lockdowns, most notably the Bounce Back Loan Scheme (BBLS) and Coronavirus Business Interruption Loan Scheme (CBILS). Despite this intervention, many businesses will not recover, leading to an increase in defaults. Collections and recoveries require specialist skill sets, and lenders will be under pressure to meet demand, while actively preventing fraud and making sure they are treating customers fairly.

We explain the eight key areas where a two-stage data-driven approach to collections and recoveries can streamline your processes, support vulnerable customers.

Collections and recoveries: a data-driven approach

A data-driven approach to collections and recoveries can help lenders manage the process and meet compliance obligations for the British Business Bank (BBB). There are eight key areas where efficient use of data can have a positive impact on CBILS and BBLS recoveries:

  1. Making sure client data is up to date and drawing on external sources to fill in the blanks

  1. Identifying vulnerable customers to improve safeguarding activity

  1. Detecting customers in difficulty and promoting early intervention

  1. Establishing a re-payment schedule and tracking missed payments

  1. Monitoring each stage of the collection process to ensure compliance

  1. Checking for key fraud indicators using a combination of internal and external data to develop a prediction model

  1. Developing a predictive customer data model for efficient resourcing and ensuring teams stay within their operational capacity

  1. Generating and tracking key performance indicators using a balanced dashboard and workflow tool to track workload, and manage customer contact and responses efficiently


Figure 1: Data led collections and recoveries approach

Streamlining processes

A data-driven approach to collection and recoveries consists of two stages, as outlined below. Essentially these two stages aim to put together all the information needed to optimise the collections and recoveries customer data model. This allows firms to gauge the scale of operational resources and use data analytics to gain further insight on how best to manage the process for each customer (through cohort analysis).

Stage 1: Data profiling

Data profiling is a valuable tool to assess the information you hold on your clients and check it is complete, accurate and consistent. This can be done prior to any collections and recoveries activity and can give greater customer insight by identifying any interconnectedness across the data. Drawing on information from the lenders internal data store, client data and external sources (such as Companies House), data profiling can also highlight any errors in the information or issues with inconsistent data formats (data quality issues).

Effective data profiling can help firms monitor the potential for fraud, identify appropriate customer contact strategies (particularly for vulnerable customers) and support early intervention. While some of these clients will not ultimately default, data profiling will help lenders manage the workload further down the line for those who do, and can help to meet collection timescales. Using intelligent external data sources can also provide indicators to identify potential fraud and therefore reduce collections issues.

Stage 2: Data analytics

Integrating data analytics into the BBLS and CBILS collection processes can help to target customer outreach by cohorts. It can also improve operational workflows and management information for improved oversight and compliance reporting.

A data-led approach can help differentiate between complex and simple cases, and highlight specific circumstances such as vulnerable customers, financial distress or issues around multiple products. Relationship managers and operational teams can use this data to inform their approach, and to support intelligent customer contact. This includes outreach via appropriate digital channels, reducing resourcing needs and keeping costs low.

Using predictive models

The above techniques can also be applied for predictive purposes, to gauge when resources will be needed and to help with forward planning. To achieve this, firms can review the current customer data and assess key areas that may impact the likelihood of a cure, or the ability to recover losses if not. Key considerations for predictive modelling include:

  • The assumptions in developing each client cohort
  • The quality of data, including the use of external data
  • Status of collateral, such as home ownership or a business asset
  • Customer circumstances
  • Contact strategies based on criteria such as how many years they have been a customer, the number of products and the total balance in their accounts
  • Differentiating workflows by segregating different types of case, considering processes and cycle times and allocating them to on-shore/off-shore teams depending on the scenarios for each group

This data will feed into a predictive model to gauge the potential for default and identify which customers may need additional support. In turn, that will inform resource allocation to make the most of special skill sets and effectively manage the high volume of activity.

The holistic data risks impacting your business Uncover the critical data risks impacting your organisation