Champion-Challenging Letter Strategies

In this series of tips, we will review best practices around Champion-Challenger testing of letters in collections strategies. It is best practice for an organisation to develop a minimum of 35-50 different collections letters, which subsequently need to be tested and monitored for results.

For organisations with sophisticated account management systems, testing and monitoring the effectiveness of different letters is conducted through Champion-Challenger testing. This is the process of selecting two statistically random sets of accounts and then performing different actions on the control (Champion) and test (Challenger) groups. Any difference in results must then be due to the different collections actions that were performed.

In Champion-Challenger testing of letter strategies, the following tests are recommended:

Letter Design In addition to the timing, frequency and tone of letters, organisations should also test the actual design of the letters themselves. This form of testing would include envelope design, letter headers, signatures and even the colour of the letter paper and ink. A number of organisations have run specific colour tests, with four sets of letters with the exact same wording, but printed on four different colours of paper, or using four different types of fonts or ink colours. What has been significant about this form of testing is that customers do have higher response rates to changed materials.

Accordingly, it is recommended that organisations not only test letter design on an ongoing basis, but also that they implement changed letter layouts, colours, fonts and so forth, on an ongoing basis. The moral of this tip is that even the most delinquent of customers like a change in letter layouts every so often!
Specific Letters The final test that we would recommend running is that of specialised letters as opposed to generic/all encompassing letters. The test here is whether a specific letter that best fits a delinquent customer’s specific circumstances produces improved results over a generic collections letter. There is definitely a cost associated with developing and generating different types of letters and so the test is whether an improvement in results for specific groups of accounts outweighs the costs associated with the specific letters.

Examples of testing of specific letters would be for new accounts, first payment defaults, different risk categories, different balances and so forth. Organisations have extended this form of testing to include language and even gender. Sending letters to a customer in their preferred language of choice certainly does improve collections responses and results, but there is a question over economic feasibility. Once an organisation starts generating letters in multiple languages, where is the line drawn as to how many languages and does this apply to all letters or just collections letters? In addition, if a collections letter is sent out in a secondary language, then there needs to be collections personnel that can answer the customers’ questions in that language, should they call the collections department.

This list of testing criteria to consider for collections letters strategies has covered three Tip of the Month articles. This underlines the wealth of data that is available to organisations for sophisticated letter testing and we would strongly encourage all collections departments to conduct ongoing letter tests and implement a complete new suite of letters on at least an annual basis.

Stephen J. Leonard is Managing Director of PIC Solutions, the largest customer management solutions company based in the Southern Hemisphere. He has over 15 years of risk management experience in the banking and consulting industries at Chase Manhattan Bank and Fair Isaac International. He holds an AS (State University of New York), BA (University of Toronto), MBA (Adelphi University – School of Banking, New York) and is a member of the UK and South African Institutes of Credit Management.