Did you know that 88% of companies are still using age and value to drive collections prioritization? Traditionally, most credit and collections departments prioritized collections based on aging. The customer who owes the most money for the longest period of time receives special attention. However, depending just on age to prioritize collections efforts may not always be the best approach, particularly if the company’s goal is to enhance collection efficiency, improve DSO, and reduce write-offs.

Over the last decade, credit scoring has evolved into one of the most sophisticated methods for automating risk analysis and assessing the collectability of a company’s accounts receivable portfolio. The model is intended to anticipate a customer’s inherent risk, such as the likelihood that the customer will become substantially overdue, go to write-off, or file for bankruptcy.

Statistical-based scoring measures unique risk probabilities for your accounts. That capability is what distinguishes it from credit bureau generic scores or in-house judgmental-based scoring. The score generated by statistical-based scoring effectively provides an estimate of the likelihood that a certain client will pay their bill on time.

A statistical-based scoring system’s standard output includes not only a credit score, but also the probability that the account will go bad, i.e., Probability of Bad (PBAD), within a specified period from the scoring date, usually six months, and an estimate of the cash value of the account that is at risk, i.e., Cash at Risk (CAR). When applied correctly, these variables will assist you in assigning collection resources to specific accounts, maximizing the return on investment (ROI) from collection activities.

There are numerous ways to apply risk-based collections – discrepancies in customers’ ability to pay, credit rating based on agency data, relative industry, geographical region, invoice valuation, age of balance, due credit limit, ability to use alternative methods of payment, historical payment behavior, and the commercial risk involved (the risk of losing the customer) must all be accurately determined in order to reap the highest return on investment for the company.

If a company can categorize its receivables into groups of customers – those with a high probability of on-time payment versus those who will be delinquent versus those with a high probability of loss – they will be able to apply specific treatments to each of those segments of the portfolio that will result in the most liquidity at the lowest cost.

Collection risk-based statistical scoring should be an intrinsic part of any collections operation and a crucial component of the cash flow forecasting process; whether it is to drive and predict free cash flow, manage bad debt, or even decrease counterparty risk around suppliers. The frequency of scoring is related to the percentage of the A/R portfolio that is past due.

Let’s look at an example situation that an organization can find itself in and see how this additional information might be employed to understand how statistical modeling drives collections. Since collection resources are limited and each overdue account cannot be addressed directly by a collector, a decision must be taken as to which accounts to call directly and which to handle via a less expensive approach, such as a dunning notice of some kind.

Assume you only have the resources to make one call and there are two accounts ABC and XYZ in dispute.

Account ABC owes $70,000 and has a PBAD (Probability of Bad) of 10%, which means it has a 10% probability of turning bad within 6 months, and its CAR (Cash at Risk) is $7,000. (PBAD times AR value). Account XYZ has a PBAD of 50% and owes $30,000, so their CAR is $15,000. What do you do? Whom do you call?

You approach Account XYZ in terms of risk-based collections. And here’s why, from both a statistical and a common sense standpoint. Account ABC is a reasonably low-risk account, with a PBAD of 10% indicating that the account will pay on time and may resent a collection call, which could disrupt otherwise solid customer relations. Furthermore, they only represent around half (100 times 7,000/15,000) of the risk that Account XYZ represents.

In other words, calling Account XYZ gets you almost 200% more bang for your buck than calling Account ABC. Account XYZ has a PBAD of 50% and is considered extremely risky. Accounts in this category should be closely monitored and called as soon as they are a few days late. In this case, the action recommended by risk-based collections is the inverse of what a historical collection decision would be, which would be to call Account ABC, the higher value account, because you would be unaware that Account XYZ actually represented a substantially greater risk.

Here is an Aging Bucket indicating Possible Risks

To summarize, various factors must be considered when deciding whether to use generic scores, credit bureau reports, and data, or a judgmental-based model improved with credit bureau data vs a statistical-based model. The ability to quantify risk is the primary distinction between statistical models and judgment models. More than any other characteristic, statistical-based models are an effective tool for credit and collection tasks.

By knowing and using the probability of the occurrence of specific credit and collection events, it is possible to optimize the allocation of resources available in a given credit and collection environment, thereby developing strategies that mitigate the possibility of negative results while simultaneously increasing the credit lines of low-risk accounts and providing the opportunity for additional revenue.