Profit-based Credit Marketing

This article discusses how appropriate account level treatment can be applied to customer segments to maximize the profitability of the credit portfolio overall. Initially, we will discuss the identification of specific ‘profit segments’. The specific actions taken are then illustrated within a targeted credit marketing strategy.

An existing account profitability model
In order to identify profitable account segments, it is necessary to calculate an account level profit. Having calculated the profitability of each account on the portfolio, groups/segments can be clustered with similar levels of profitability, revenues or costs for appropriate treatment.

We will not be discussing the components of the existing account profitability model in this article, but rather focus upon the application of an account level profit model within existing portfolio decision strategies. The development of the profit model is an extensive and complex topic requiring a separate article in its own right.

Segmenting for profit
With the account level profitability established, various levels of segmentation can be applied to identify account groups for specific account management decisions.

The diagram above shows risk and customer value segments.

As risk is associated with the likelihood of an account achieving a certain level of delinquency (referred to a ‘bad state’ in scoring terms), it also predicts future loss that will be incurred. Within profitability models of revolving card portfolios, the majority of losses incurred are due to written-off/charged-off debt. The vertical risk axis therefore also predicts future losses.

The horizontal axis depicts profit. Within a revolving card portfolio profit is primarily usage-based (the more the product is used, the more revenue will be generated). Revenue therefore indicates customer value, i.e. generally the most profitable customer is the one that utilises the facility extensively with a low anticipated loss.

It can be seen that four profit-based customer segments are identified:

• Choose to Lose
  This group is generating low revenue with risk being unacceptably high (indicating high future losses). As a result, this group will reduce overall portfolio profitability and are strong candidates for pro-active account closures.
• Remedial Class
  This group, although generating high revenue, also represents a higher risk. The challenge for risk managers is therefore to reduce the risk (i.e. restrict the bad debt losses) emanating from this group. As a result, this group would receive accelerated treatment within the collections area. Marketing activities are also likely to be suspended due to the high risk level.
• Under Achievers
  This group is not generating high revenue, although the associated risk is low. Under risk-based reward policies, very low-risk customers would have received generous limit increases and overlimit expansions. The reason for the low spend (revenue) is typically due to these low-risk customers being less responsive to limit reward strategies.  The challenge is therefore to encourage spend and/or cross-sell further products to improve the revenue being generated by this segment.
• Golden Group
  Need we say more? This group has the ideal combination of low risk and high utilisation (generating high revenue). As a result, these customers need to be protected against attrition and migration to competitors.

This is a basic segmentation to illustrate that specific and appropriate treatment is possible, where account groups are segmented by combining revenue and loss measures. In practice, sub-groups within each segment may be defined using further segmentation fields, such as whether the customer is a transactor or a revolver, whether the balance on the account is high or low, etc.

Illustrative example: Marketing Strategy for a Retail credit portfolio
In order to illustrate how the strategy would be applied operationally within the credit environment, consider the following example showing contrasting marketing strategies. Although all four segments are shown in the diagram, only the ‘Choose to Lose’ and ‘Under Achiever’ groups will be discussed.

Down the left hand side a sequence of segmentation criteria (denoted by the blue shaded rectangles) are used to ultimately derive the four targeted groups. The first three criteria were applied to eliminate policy exclusions from the strategy as follows:

• Block code – Eliminates accounts with exceptional circumstances such as deceased, fraudulent, etc.
• Delinquent cycle – This key is used to separate delinquent accounts from those accounts with an up-to-date status. This means that delinquent accounts will not receive marketing actions and the focus will be targeted towards accounts up-to-date.
• Time on books – This key defines mature and immature accounts in terms of the number of months that the account has been on the portfolio. In this case, a mature account is defined as being 12 or more months old. For the purposes of illustration, immature accounts (less than 12 months) are excluded from the marketing strategy.

The next two segmentation criteria are used to define the revenue and risk of selected accounts:

• Average utilization over the last 6 months As mentioned before, usage of the facility is strongly correlated to the revenue that will be generated. As a result the average utilization (that is the balance divided by the credit limit) over the previous 6 billing periods is used as a proxy, in the absence of a scientific profit model.
• Account risk – This is simply the internal risk assessment used for rating accounts in terms of projected default. In more sophisticated environments this would typically be a behaviour score, which may be coupled with a credit bureau score to predict the likelihood of a pre-determined ‘bad’ state.

From the flowchart it can be seen that the ‘Under Achiever’ segment has been defined, and appropriate actions have been assigned. In this case, the group will be offered discount vouchers to encourage spend on the retail card.

In addition, since the usage (and therefore revenue) was low, insurance is to be cross-sold at a discount to obtain further revenue. The premise here is that insurance would have been offered previously under normal pricing conditions resulting in low take-up. In order to improve the revenue of the group, insurance may be offered at a lower rate. Although there would be a reduction in the insurance product profitability on an overall basis, net revenue (assuming increased uptake) from ‘good’ risk accounts would be improved, thereby effectively improving profitability.

Appropriate actions can also be taken on the ‘Choose to Lose’ group, which have also been distinctly identified. In order to contain spend, no marketing activities (budget) will be applied as shown by the actions in the red shaded rectangle. There are also no benefits or associated incentives granted due to the cautionary stance being taken.

Summary
Although very different marketing treatment was taken upon each group, each was consistent with a single business objective – that of maximising profitability for the portfolio. If customers do not use the facility during the account management lifecycle, the portfolio will not be profitable. Emphasis should be placed upon taking appropriate actions on those groups that will yield an improved and/or sustained level of profitability.

Sharief Allie is a Senior Consultant with PIC Solutions, the largest customer management solutions company based in the Southern Hemisphere. He has over 9 years of risk management, business analysis and product development experience in the financial services industry and specialises in general credit risk management covering the entire credit lifecycle for various credit portfolios. These include retail credit, bank cards, personal loans, telecommunications and secured lending products. He was previously with Woolworths Financial Services, where he was involved in the successful implementation and analysis of account management strategies and StrategyWare. He has extensive experience of POS systems, imaging systems, predictive diallers and systems parameter maximisation of the Vision21 and VisionPLUS account processing systems. He holds a B. Com from the University of Cape Town and a Diploma in SAS programming. Member of the SA Institute of Credit Management and the South African Statistical Association.