Spending more to reduce debt-default rates isn’t always worthwhile. To maximize profits, businesses must balance bad-debt risk with the many obvious and not-so-obvious costs of aggressive credit reporting.

Credit investigations involve labor and information costs that consume a substantial portion of the typical credit-and-collections department’s budget. Midsize companies can spend tens of thousands of dollars annually on commercial credit reports alone, and large corporations that have bigger credit-reporting contracts and more credit analysts on staff can easily spend hundreds of thousands of dollars a year investigating their customers’ creditworthiness. Are these companies getting their money’s worth?

Most companies allocate substantial resources to the credit-approval process to minimize future bad-debt expenses, but they don’t always fully consider the costs of allocating those resources to credit decisions. "The more credit investigation you do, the lower your bad-debt expense is going to be, because you are better able to sort out customers with high bad-debt probability from customers with low bad-debt probability," explains Dr. Frederick Scherr, a professor of finance at the West Virginia University College of Business and Economics in Morgantown, W.Va. "However, there is an optimal amount you should spend on credit investigations. If you spend too much, you can drive bad-debt expenses to zero, but you will have a huge credit-investigation bill."

Sorting through information and analyzing relevant data takes time that a credit-and-collections department could spend on collections activities, which can mitigate a company’s slow-payment risk and bad-debt exposure. Credit managers, particularly those evaluated on their ability to contain bad debts, tend to justify their credit decisions by buying more credit information and doing more in-depth investigation than necessary. Credit managers who take this course believe that credit-investigation expenses are good costs, while bad debts are bad costs. "I have had credit managers tell me that they spend a ton on credit investigation because they want to avoid bad debts," Scherr says. "This fails to recognize the trade-off between the two. It is the sum of the two you are trying to minimize."

Another important consideration in creating a credit policy is the trade-off between risking bad debts and making additional sales. Too restrictive a credit policy can result in the loss of profitable sales, just as an overly generous policy can reduce profits because of increased bad debts.

The trick, then, to cost-effective credit investigations is gathering just enough information to make an appropriate decision. Credit-approval procedures should strike a proper balance between maximizing profitable sales and minimizing the sum of bad-debt expenses, carrying costs and credit-investigation expenses. This balance is best achieved when a credit department is charged with maximizing profits rather than minimizing bad debts, and when a credit staff has clear decision-making guidelines.

Evaluating Risk

Most credit executives have been trained to base their analysis of customers on the four Cs of credit: character, which indicates the debtor’s willingness to pay obligations; capacity, a measure of a company’s ability to operate successfully; capital, a measure of the means by which a business will pay its obligations; and conditions, the general and industry-specific economic conditions that influence the final credit analysis. These criteria are essential for evaluating a customer’s creditworthiness, but they disregard the importance of product value in decisions about whether credit should be extended.

Several factors influence product value from a credit standpoint. Foremost among these factors is gross profit margin. The higher a company’s selling margins are, the more risk it can assume in order to maximize profits. Likewise, order size is an important consideration; the larger the order, the greater the company’s exposure to risk. A large order for a low-risk customer may pose a greater exposure than a small order for a high-risk account.

In addition, a company’s capacity for increasing sales affects the level of risk the credit department should accept. Businesses that operate well below full capacity are paying overhead on orders they turn down. However, as production approaches full capacity, a company has less incentive to assume additional risk.

One last factor that affects a product’s value to the credit department is the speed with which it becomes obsolete. Products that are subject to frequent upgrades and improvements may lose value faster in inventory than as receivables. Obsolete inventory sitting in a warehouse can be as detrimental to a company’s bottom line as a bad debt.

When making credit decisions, consider your product’s intrinsic value so that you don’t miss opportunities to make profitable sales. By the same token, certain selling situations call for a more careful consideration of risk. Obviously, deciding which credit data is appropriate to each situation can get complicated. This complexity is one reason why many credit professionals gather all the information they can from credit reports, trade references, bank references, on-site visits, industry credit groups and the Internet. However, if you thoroughly understand your business’s risks and decision-making process, you can develop a focused approach to credit investigations that is not overly complicated.

Learning from Decision Trees

Studying credit decision trees is a good way to understand the credit decision-making process. Decision trees map in flowchart form every step in the credit-approval process. A decision tree begins with the first decision a company makes when it receives a credit request. The decision has three possible outcomes: approve credit, deny credit and gather more information. A probability of default is associated with each outcome. The gather-more-information outcome leads to another decision that has three outcomes, and the tree grows from there. In textbooks, credit decision-tree illustrations have a dozen or so branches, but in real businesses they can contain a couple hundred distinct decision points. For example, decision trees for companies that sell products with different margins need to reflect the entire decision-making process for every product’s gross profit margin.

"What is presented in textbooks is a very, very scaled-down model of how credit managers would actually think about the credit-approval process," says Scherr. Decision trees are "really not intended to be practical. It requires too much initial investigation to set up a model that adequately captures the process that is going on. You can make the tree simpler by making fewer distinctions within it, but then you have lost representativeness, basically throwing away data. The decision trees you see in a textbook are intended to show you where the trade-offs are between size of order, margin, bad debt and credit-investigation expense," he explains.

Scherr points out that as you study these trade-offs, several relationships become obvious:

  • The larger the order, the more credit investigation you should do.
  • There is an optimal amount of credit investigation for each size of order.
  • Complete investigation is not always the way to go; you can make correct credit decisions but blow all your margins on credit-investigation expenses.
  • Higher-profitability lines require less credit investigation, because increasing the number of customers is worth the credit-risk trade-off.

By understanding these relationships and the standard risk levels presented by the different types of customers in your receivables portfolio, you can begin structuring your credit-approval process for maximum efficiency. Although credit decision trees are too cumbersome for everyday use, the thought process behind the decision matrix provides a relatively simple system for making credit-approval decisions. Scherr explains, "There is a way to look ahead only one decision, which I call ‘decision tree on the fly.’ Rather than building the whole tree, you ask, ‘What can happen as a result of the next step of the credit investigation? What outcomes am I likely to see? How valuable is that information to me? Is it worth spending the additional dollars?’" You can answer each question mathematically or intuitively.

Reaping Powerful Results

Pioneer Balloon in Wichita, Kan., has used these principles to revamp its sources for credit information and methods for making credit decisions. "With a new account, our first step is to run a one-page credit-score report," explains Mark Borofsky, Pioneer Balloon’s corporate credit manager. "The score and the size of the order determine whether we are going to give open terms or investigate further. In some markets our margins are higher, so we are a little looser extending credit. Others are tighter, so we are looking for a higher score. If there is a large order but there are not a lot of trade lines reported, we will run a full profile on the company, even if the score is high."

Borofsky and his team sometimes ask a prospective customer’s permission to get a personal credit report if the credit agency’s database contains insufficient company information, and they review financial statements for large distributors. However, they haven’t checked references for more than two years. While competitors are checking references, Borofsky is approving orders. "What people fail to realize is that the average cost to check all the references for one account is $30 to $60," he notes. Pioneer Balloon requires new customers to submit a signed credit application that includes six references, but the company checks the references only if the customer starts making late payments.

When they cannot say yes to an order and cannot justify gathering more information, Borofsky’s staff negotiates alternative selling terms. "If the credit report comes back and it raises questions, it is not that you do not sell to the customer, it is that you sell to them safely. The idea is not to stop selling to high-risk accounts, but rather to take precautions. We sell to high-risk accounts, but we manage them with purchase money and blanket UCCs [Uniform Commercial Code financing statement filings], by billing directly to distributors’ customers," and by using other appropriate means, Borofsky says.

Pioneer Balloon uses risk-analysis software to support its credit decision-making process. In addition, the company shares customer payment histories with its credit-reporting agency every month, and in return, the agency scores Pioneer Balloon’s entire receivables portfolio. This combination of internal and external credit data provides for easy recognition of customer and portfolio trends.

As a result of this new credit-investigation process, Borofsky’s team has shrunk from seven to five employees and has reduced its budget for credit reports by more than $30,000. At the same time, the department actually recovered more bad debts than it wrote off last year, because employees spent more time on collections and less time on credit checks.

Pioneer Balloon has realized a competitive advantage from its new credit-approval process, which handles roughly 2,000 new applications yearly. "We have a reputation in the ad premium industry of approving credit more quickly than anybody. The advantage we have over our competitors is that they take up to a week to clear references. In most cases, we are done in a couple of minutes, whether the order is for $5,000 or $200,000," Borofsky claims. "We mapped an order through our entire system and were able to knock out over 15 steps. That just speeds up the order flow." These improvements could not have been achieved without a thorough understanding of the credit-approval process.

Maximizing Software Support

Cardinal Distribution in Dublin, Ohio, the wholesale pharmaceutical unit of Cardinal Healthcare, uses risk-analysis software to standardize its credit decision-making process and rein in bad-debt expenses. The company manages a large receivables portfolio with many high-exposure customers. Financial analysis is a critical factor in most credit decisions for businesses with this type of receivables portfolio. Cardinal uses specialized ratio-analysis software to ease that labor-intensive task, then imports the resulting data into the risk-analysis package for storage. "We are really not changing the analysts’ jobs so much as creating an electronic credit file and making their jobs easier. It helps with consistency and standardization, taking our due diligence to the next level," observes Jeffrey Carpp, director of financial services. The data for each customer is readily shared up and down the corporate approval ladder.

The credit-information database Cardinal is building relies on a number of data sources specific to the health care industry. The company acquires financial norms from health care-related financial information providers. From those norms, "we have created models for four different industries with different customer types and segments within the health care industry," says Carpp. "You can define things to break down to specific financial ratios that are unique to retail organizations or charitable organizations or whatever you want to look at." Over time and with a sufficient number of organizations in your database, you can do a comparative analysis to validate your models and fine-tune your decisions.

Missing Scoring Opportunities

Pioneer Balloon uses generically-derived payment scores as the first step in its credit approval process. At the other end of the spectrum, Cardinal Distribution relies on financial analysis software and comparative data models to provide the parameters necessary for most credit decisions.

Other companies attempt to automate their credit-approval processes by developing their own expert systems. These systems can help streamline decision-making, but they do not necessarily help businesses make better decisions. "Expert systems use the credit manager’s expertise to help automate the decision process to bring greater efficiency to either new credit decisions or new credit authorizations on existing customers. When we take the same data credit managers use in their expert systems and apply statistical methods to it, we come out with a much stronger predictor, because mathematics is so strong in this area," explains Michael Banasiak, president and senior consultant of Predictive Business Decision Systems Inc. (PBDS) in Tinton Falls, N.J.

PBDS retrospectively analyzes expert systems to compare their ability to predict risk with a statistical model’s risk prediction. The statistical behavioral-scoring model always does better, usually much better. "In one case, an expert system using accounts receivable information, financial statement information, collateral details, commercial and consumer bureau data, everything you could imagine, did no better than random. The behavioral-scoring system was able to do much better just using historical payment data," Banasiak says.

The Bottom Line

This information should serve as a warning for companies contemplating highly sophisticated credit-approval systems. Expert systems are no smarter than the person who designs them. In addition, Pioneer Balloon’s Borofsky observes, "I think a lot of people over-guideline credit." Simpler may not always be better, but complex systems need to be carefully designed if they are going to provide quantifiable benefits.

For companies with many small-balance accounts or high margins, generic credit-scoring products may prove both economical and effective. For companies that deal with more homogeneous customer segments, customized behavioral scoring models can provide an effective foundation for a credit-approval process. Businesses with large exposures or tight margins generally require more financial analysis. These companies can use ratio- or risk-analysis software to provide the mathematical support necessary for sound credit decisions. Every company must consider a unique set of credit-approval issues. Whatever route you choose for your credit decision-making process, be careful not to waste time or money gathering and analyzing more credit information than you really need.