by Ulrich Wiesner, senior consultant on FICO’s Analytics team in EMEA, specializing in collections and recovery
Many banks in both developing and mature credit markets are insufficiently prepared.
The economies in many countries have not seen a recession for more than a decade. Low levels of unemployment and low loss rates in retail credit have allowed banks to lose focus on their collection and recovery departments. Where investments have been limited to regulatory requirements, organisations are falling behind international good practice. The lack of prior investment shows across methodology, technology and organisational aspects. Even large organisations can be 10-15 years behind standards in markets such as the United Kingdom, Turkey or South Africa. A little bit of digitalisation won’t do enough to catch up.
Debt collection is a strange business. Most customers in arrears are not under severe financial stress and are eventually going to pay when being reminded. These customers primarily cause operational costs. Worst case, they are going to call to apologize for the late payment, creating additional operational efforts. A smaller segment of customers is in temporary or systematic financial hardship and will be unable to comply with the contractual payments required, so that some of these accounts will need to be terminated later.
The resulting risk-related costs caused by the latter customer group substantially exceed the operational costs caused by the first one, even in a prospering economy. In consequence, a career in debt collections does not just mean spending time in difficult customer conversations, it also brings with it responsibility for a large cost figure.
This seems to be the reason why banks frequently follow a strategy that is best described as dunning and typically consists of an automated chain of remainder letters, eventually culminating in the termination of the account.
In these situations, operational emphasis is often put on performance indicators such as volumes worked, unit costs and service level adherence, while result-oriented KPIs such as roll rates, impairments and net credit losses do not get as much attention. This results in avoidable account terminations, which ruin the customer’s credit record and destroy the customer relationship. In debt recovery, pragmatic payment plans are then suddenly offered to the same customers, and termination becomes questionable when strong recovery rates are achieved. This is not as evident in countries with more sophisticated collection approaches. Great recoveries are evidence of poor collections, unnecessary customer loss and significant, unnecessary profit erosion.
Communication like in the 1990s
Another challenge to many banks seems to be to communicate with customers in a contemporary way. Symptoms include the focus on letter strategies, insufficient modern telephony equipment available to collections departments, as well as limited use of digital channels, as if the internet and mobile communication never existed.
Frequently, collection communication is limited to a series of dunning letters until the account is ready to be terminated. Lack of productivity prevents approaching customers sufficiently early via phone, and data protection concerns — often dating back to the 1990s — still impact the way text messages (SMS) are used by many organisations. Hardly any collection department uses omnichannel communication platforms that can seamlessly hand off customers between alternative interactive communication channels such as voice, text message, email, and chat.
A lack of productivity in outbound telephony has a substantial negative impact. If outbound calls are to be dialled manually, collectors will hardly be able to make more than four right-party contacts per hour. In consequence, banks without appropriate technology in place typically do not actively call their delinquent customers, or they do so only very late into the process. A predictive dialler removes the burden of unsuccessful dialling attempts from collectors, and typically can contact ten customers per collector hour.
When text messages became popular in the 1990s, many organisations disregarded their use for collections due to data protection concerns and have not revisited this option since. Others use SMS only to ask customers to call back – some even without identifying the name of the bank. Such practice reflects concerns that an unauthorised third party could gain knowledge about the arrears situation or even the contractual relationship, for example, because the mobile phone is shared between multiple users, the number has not been captured correctly, or has been passed on to someone else by the network provider. Specifically, the concern about mobile phones being shared between multiple people feels like it is from a completely different era. Today, mobile phone users are in full control of access to data and messages stored on their phone. Appropriate wording of the message can prevent an unrelated third party from gaining any information from a misdirected message. Not using text messages as a communication channel means giving up a very successful, customer-friendly and cost-effective tool, which is superior to letters in terms of speed, accessibility and potential for interactivity. This is the reason why an increasing number of organisations are no longer using letters except where required by regulation.
Where SMS are used in debt collection, they are a successful and highly cost-effective communication tool. Combined with a back channel, they even allow customers to respond at any time, e.g., to set up promises to pay or to trigger a direct debit. Data protection needs to be taken seriously – but banks that abstain from using text messages in collections to avoid non-compliance need to ask themselves when they want to arrive in the 21st century.
Which map to use for navigation?
The way collection departments are integrated into the overall organisation of the bank has an underestimated impact on collection results. Typically, the collection department either reports to the risk department or to operations. This organisational decision can have a tremendous impact on the organisational culture and where the focus sits.
Risk-related costs usually exceed operational costs of collections, a primary focus on operational costs can have a disastrous impact on the overall results. For that reason, more advanced collection organisations have a strong focus on the risk-related performance indicators, which are also mirrored in annual personal goals of the respective stakeholders. In risk-focused organisations, across all levels, operations are managed based on numbers related to credit losses. In contrast, an operations culture dominated by unit costs and service levels will struggle to get sufficient focus on risk results as a primary goal.
South Africa is facing a coronavirus-driven jobs bloodbath, with the Business for SA (B4SA) formation forecasting that up to four million jobs could be lost this year.
The suddenness at which the COVID-19 has hit the world, leaves limited time to strategically re-focus. Currently, the economic impact of the pandemic is difficult to estimate.
Most collection departments have not seen inflated volumes yet, but as soon as the regulatory payment holidays expire, collection departments will likely face substantial volumes of customers who will be unable to return to their contractual instalments. Should the world economy head into a persistent recession, collection managers might need to handle account volumes they have not seen in their entire careers. This will pose big operational challenges, and the resulting credit losses will be substantial.
May customers who are now feeling financial stress are good customers in bad circumstances. Previous crises in other countries suggest that such customers aim for a quick rehabilitation and typically will financially stabilize earlier rather than later. Banks able to identify them and offer appropriate solutions are more likely to avoid unnecessary losses, strengthen the customer relationship and hence secure future revenues.
There is not much time left to prepare for a potential tsunami. This requires emphasis on several tactical measures which we at FICO have flagged as critical to any bank’s collections’ strategy during this time:
- Higher volumes in collections require a stronger focus of available staff on relevant accounts, as well as a higher level of automation of standard processes. Automated communication platforms allow a fully automated, interactive communication with customers via voice, text and chat. Such cloud-based solutions can be launched with minimal integration effort within weeks and provide almost unlimited scalability. They allow for fully automated processing of low-risk cases, including capturing payment commitments. With additional integration effort, they even allow agreements that do not required detailed discussion or explanation, such as due date modifications, simple payment plan adjustments, or a repetition of an agreed direct debit. At the same time, such platforms allow customers to address their issues at any time, even outside standard office hours, and without the perceived and uncomfortable need to justify financial delays to a human collector.
- Increasing volumes will force operational units to focus their actions on important cases. If the propensity of a customer rolling to the next arrears bucket is known, then the higher loss-incurring accounts can be targeted more easily. There may not be enough time to develop and implement scorecards or machine learning models ahead of the wave. But there should be enough time for the development of data-driven segmentation trees. In addition, existing regulatory models such as propensity to default can improve segmentation quality, if such score values can be brought into operational systems.
- Customers with structural financial difficulties require appropriate forbearance measures to address their financial hardship. In the past, many banks have used such tools only by exception. In consequence, the range of tools in place is often limited, and related policies are not up to date. However, in most cases a successful restructure of a loan is economically more favourable than the termination of the loan, even in case of financial concessions. Banks should review their toolkit sooner rather than later and update the respective guidelines and policies. To assess which forbearance tools should be offered to which customers, it is also important to understand the impact of such measures under IFRS 9. To ensure appropriate conversion rates, a successful hardship framework should include streamlined operational processes and avoid bureaucracy and unnecessary paperwork.
- Even employees with a strong affinity to numbers struggle with estimation of the impact of payment plan changes on Net Present Value, Expected Loss and capital requirements. In consequence, mathematical methods can substantially improve the selection of the appropriate forbearance measure. Such methods allow derivation of strategy trees that, depending on the nature and extent of the financial stress, select the tool that is most appropriate in context of a given business goal. Should the crisis turn into a persistent recession and result in substantially inflated NPL levels, such methods should be seriously considered to improve the impact and success rates of restructures. As a short-term measure, organisations should at least temporarily reduce the instalment to an amount that seems affordable in light of the customer’s available income. Debt payment to income metrics used in originations can act as a guideline of what seems acceptable for a given product or portfolio. Again, data-driven segmentation can help to separate potentially successful restructures from hopeless cases.
Mid-term, many organisations cannot afford to further postpone updating their technical debt collections infrastructure. Only a contemporary collection system enables banks to swiftly adapt segmentation and treatment chains to changing volumes, capacities and risk distributions.
Organisations should not wait for calmer times to return before they address respective shortfalls. And only the use of a predictive dialler enables the level of productivity that allows an organisation to approach every customer in time. Every manual call wastes the opportunity of two additional calls that could have happened in the same period.