A New Era in Account Management Decisioning

Defining the Status Quo

In the late 80’s and 90’s, automated credit risk management expanded its focus from credit acquisitions to incorporate full credit life cycle management. In addition to application scores that assessed a potential customer’s likelihood of repayment, behaviour scorecards were developed, enabling lenders to use historical usage information on their accounts to predict future behaviour.

These behaviour scores could then be used in account management strategies, focusing on controlling specific elements of account management, for example increasing or decreasing revolving credit limits, delinquent and overlimit collections, marketing activity and performance-based pricing.

At the time, the biggest challenge with implementing automated account management decisioning was the data scale required for the calculation of behaviour scores. To build behaviour scores that are robust as well as predictive, scorecard developers typically used historical observation data covering twelve months, which is a data window that incorporates all seasonal impacts.

To keep the data footprint down to manageable levels, remembering that all of this processing was performed on mainframe or equivalent platforms, the twelve months of data was summarised into monthly snapshots, coinciding with the billing cycle of the accounts being processed.

For this reason, behaviour scores were calculated monthly, at the time of account billing. The majority of account management decisions were also taken at this time, with the notable exception being collections, which required daily updates to track the effectiveness of the actions taken or to accelerate actions if required. However, even collections strategies relied on a limited set of daily (intra-cycle) data to make their decisions.

Since its introduction, automated account management decisioning has remained largely consistent in approach. Monthly cycle data is used to calculate the behaviour scores, which are in turn used in account management strategies to determine the actions that are to be taken.

The lack of change in approach to account management is understandable, as the credit products which traditional account management strategies are set to work with have not changed in any meaningful structural way. Most of these products run on monthly cycles, with a payment expected every cycle, and delinquency ageing is calculated from cycle point to cycle point. This matches with the use of monthly data snapshots used in account management systems, as these can be aligned with the payment cycles of the products.

However, while this method of account management may work well for instalment loans, asset finance, and credit cards, it is not particularly useful for managing products that do not have monthly cycles. Short-term micro-loans, sometimes called payday loans, have a loan term of a few days to a few weeks, with a single instalment.

These loans should not require ongoing account management, given the fact that they have a single repayment. However, the fact that we know the vast majority of customers apply for a new loan every month, which effectively makes the loan more of a revolving product in behaviour. This means that the recurring loans should probably be treated as an ongoing account.

Another product that does not fit the monthly cycle is the overdraft facility, where customers can hold a negative balance on a transactional account. In the mobile money space, overdraft facilities are increasingly being offered to wallet holders, allowing lower risk customers access to a relatively small credit facility on the mobile wallet.

What these account types have in common is that they are typically low-value credit products but offered to very high volumes of consumers. It is a portfolio type that demands extremely exacting account management oversight, as profit margins are tight, and the number of accounts is far too great to be managed manually.

Added to this is the fact that these accounts do not have a typical monthly cycle. Loans or advances are typically repaid within days or weeks of being offered, and it is critical to constantly observe, monitor, and react to the usage of the account, not simply at monthly cycle points.

Mobile wallet overdrafts are very interesting as they have a wealth of transactional information that can be used to evaluate customer behaviour. As the customer transacts using the wallet, all debits and credits flow through the account, in much the same way as they do in a traditional chequing account.

To effectively manage these accounts, a new type of account management decision engine is required. While the decisions are often the same as made by traditional account management decision engines, the data used to make the decisions is very different, and the timing of when the decision engine should be invoked needs to be considerably more flexible.

With the development of Event-Driven Account Decisions Management (ED-ADM), Credit Risk Connection has configured its ADEPT Decisions Platform to deal with both of these challenges.

Matching Data to Product

ED-ADM uses a configurable data dictionary, enabling subscribers to submit and use whatever data is most relevant for the product being managed. This data layout is not inherently cyclical, and thus data can be focused over any timescale that is predictive for the decisions being made.

For example, subscribers that have mobile wallet products can bring in a lot of recent debit and credit transactional data to model the financial health of the customer, which is invaluable when assessing them for facility increases, new product offerings and collections activities.

The scale of the ED-ADM data dictionary means that subscribers do not have to sacrifice longer-term data indicators for the inclusion of more recent transactional data. When assessing customers that have been on the book for longer, historical data snapshots can be included alongside the recent transactional data, bringing in increased robustness to behaviour scorecards alongside the highly predictive transactional data.

Applying Account Management Decisioning when it is needed

Equally important to using appropriate data, is to make decisions at the appropriate time. The introduction of event-driven system calls in ED-ADM means that account holder actions can be evaluated and acted upon immediately after any changes occur on their accounts. Whenever changes occur in the collections management system, accounts can be flagged for evaluation within ED-ADM. A prioritisation layer within the calling system ensures that the most important account behaviours are actioned first.

With these components in place, high-priority actions such as account blocking and unblocking, collections activity, and customer request handling can be dealt with quickly, leveraging the full power of the decision engine.

Scores can be calculated in real-time and then used within complex strategies that determine exactly what action should be taken. These outcomes are fed back directly into the collections management system for action, ensuring that updates are applied as soon as possible after the decision is made.

Using this approach, ED-ADM can perform decisioning in areas that traditional account management systems could not, due to their monthly or daily batch processing. For example, transaction evaluation for potential fraud can now be implemented within ED-ADM, using relevant transactional information, and purpose-built fraud prevention models. Blocking accounts when they cross non-payment or transaction velocity thresholds can also be implemented in ED-ADM, as can unblocking once the account moves out of these high-risk categories.

Collections activities can be updated and applied immediately once payments are made, activity occurs on the account, or information is gathered from the collections agent. This contrasts with overnight updates that happen with traditional batch-processing decision engines.

Not Only for Non-Traditional Products

Although the focus of this article is on how ED-ADM addresses the needs of non-traditional lending products, such as payday loans and mobile wallets, the benefits of using ED-ADM are not limited to just these two products.

Traditional transactional products, such as credit cards, current accounts and revolving loans can also benefit from the use of more granular transactional data and event-driven account management decisioning. Even instalment-based loans, such as BNPL,  can benefit from reactive real-time collections strategies. And all of this is on top of what is normally provided by traditional cycle-based account management decision engines.

ED-ADM is truly the best of both worlds!


Mobile lenders are already reaping the benefits of our new account management approach. Should you wish to find out more about how the new ED-ADM platform can take your business to the next level of account management, please reach out to us at hello@adeptdecisions.com.

We will be delighted to set up a software demonstration of our ADEPT Decisions Platform and discuss the potential benefits that it can bring to your specific business model in 2024.

About the Author

Jarrod McElhinney is the Head of Client Solutions and has been with ADEPT Decisions since 2017, playing a key role in designing and managing the platform, and ensuring that all subscribers realise direct business benefits from our solutions.

About ADEPT Decisions

We disrupt the status quo in the lending industry by providing fintechs with customer decisioning, credit risk consulting and advanced analytics to level the playing field, promote financial inclusion and support a new generation of financial products.