The Impact of Scoring on Collections

Collectability score is a result, not a process, and is often misunderstood. The biggest problem most firms have with scoring is implementation, but more on that later. There are also vendors who sell a so-called “credit score” as a collection score after the account has gone into default. However, if those scores proved their lack of value because the credit score failed to recognize the debtor as a bad risk, how reliable would their score be on the debt collection side? Not at all!

Essential Components of a Good Scoring Process
There are several key elements that are essential components of a good scoring process; unfortunately, many scoring services are missing some or most of these elements:

• Skip element: Does the scoring process first identify the most recent address of the debtor? If not, the score is relying upon an address that may be old and out-of-date.
• Telephone number: Does the scoring process use the updated or verified address to locate the best telephone number of the debtor?

If the two above items are not done properly, the scoring algorithm has less validity. Without the ability to make a right-party contact, any score has less value. You cannot rely upon the address and telephone that the client provided, because one of the reasons the debtor is past due is, perhaps, because he or she moved. Also, some scoring programs just use the header information from one of the credit reporting agencies—information that is frequently out of date or obsolete.

• Scrub element: Does the scoring process attempt to locate not only bankruptcies and SSDI information, but judgments and liens as well? If not, the score may not reflect that there are judgments and liens that would directly affect the ability to collect the debt. Also, the analytics should “date weight” all bankruptcy and SSDI dates. More recent dates could mean that claims against an estate have a higher probability of yielding some form of payment.
• Asset identification: Can the scoring process identify not only real property, but such items as cars, boats, planes and other high-value items?
• Type of debt: Each type of debt has its own characteristics, and some are easier to collect than others.
• Time since charge-off: The earlier the better. This needs to be weighted.
• Number of agencies: This is critical; how many agencies have worked the paper? The higher the number, the harder it is to collect.
• Balance due: This also is critical and directly affects possible work rules. See also the section on Probability of Payment below.

Scoring Algorithm
The effective scoring algorithm should produce two scores:

• The Locator score indicates the probability of a right-party contact.
• The Collectability score should reflect the probability of effecting some form of payment from the debtor.

Using these two scores, the file could them be sorted, first by collectability score, and second by the locator score. The accounts could be segmented into a logical collection strategy at that point.

All of us in the collection industry are well aware that it is far easier to "effect" a payment on a smaller balanced account. Most debtors can more easily write a $50 or $100 check compared to a larger balance account where the collector is demanding a much higher payment. Therefore, if a scoring program or process factors in a POP factor, the program will tend to give a higher score to the lower balance accounts, which is obviously not the desired result.

Account Segmentation
Account segmentation is the issue that creates the most problems for agencies new to using scoring because it requires a discipline and a paradigm shift to not handle all accounts in the same manner. If the score is not properly and logically implemented, then scoring was a waste of time and money.

Account segmentation decision elements include three areas:

1. Who (collector or group of collectors) should work the account.
2. How the account should be worked (calling and letter campaigns).
3. When should the account be worked/called.

For example: if the accounts are segmented into quintiles you have five areas: the top 20 percent, the second 20 percent, the middle 20 percent, the fourth 20 percent and the bottom 20 percent. If we assume the Top 20 percent has the most potential dollars in it, you should assign that group of accounts to your top collectors. Create a calling campaign to call only between the hours of 6–9 p.m., local time, Monday through Thursday, Saturday afternoons and all day Sunday. A letter per week should be sent.

What’s the result: We answered all three elements above. Each group of accounts should have a similar segmentation plan.

Using the above to its logical conclusion:

• The second 20 percent is worked by experienced (not yet top tier) collectors, calling 8-10 a.m. Monday and Wednesday, and 6-9 p.m. Tuesday and Thursday plus one weekend call. One letter is sent each week.
• For the middle 20 percent, average collectors call between 8 a.m.–9 p.m., Monday through Thursday, plus one letter every 10 days.
• In the fourth 40 percent group, rookies and trainees call Monday through Friday, 8 a.m.–9 p.m., local time. Letters are sent as necessary.
• Lastly, in the bottom 20 percent, there usually are not enough potential dollars in this group to justify calling, so make this a letter campaign.

Accounts could be segmented in numerous ways. Some of the most popular in addition to the quintile scenario above are segmenting them into quartiles, i.e., 25 percent groups, or segmenting them into thirds, i.e., 33 percent groups. The final element of segmentation is the ultimate Decision Tree: Which accounts will be worked in-house, which ones should be outsourced, which ones should be referred to legal, and which accounts should be sold or archived.

Alan Beshany is Vice President of Decisioning Sciences at Great Seneca Financial Corp. He can be contacted at beshanya@greatseneca.com or 720-838-8074.