Credit score improvement

Abstract

When it comes to automating the decision process for extending credit to consumers, the market has spoken. These days, score-based, computer-driven credit decisions are the norm, because in an unforgiving marketplace, fast, reliable credit decisions bring competitive advantages and higher sales.

But businesses have been far slower to adapt automated credit solutions when dealing with other businesses. Barely more than a third of B2B receivables managers were using an automated system to drive credit decisions, according to a 2003 survey by the Credit Research Foundation (CRF).

The majority of businesses still extend credit to suppliers, contractors and myriad other businesses the old-fashioned way: processing applications by hand, checking bank and trade references, consulting standard-issue credit reports and other data, passing paper to and fro for approvals, and finally rendering a decision.

So why has B2B been slow to come to the party?

One reason is that despite all the glittering promise of automated decision making - the prospect of completing a task that used to take five to eight days in the time it takes to order a drive-through hamburger - the reality has a checkered history.

Far too often, expected dramatic increases in efficiencies have failed to emerge or did so only after a painful transition. Frequently, credit managers bought into the vision of automated credit systems only to find themselves dealing with unanticipated headaches: clunky, bulky software; training challenges; and weak technical support to efficiently work through glitches. "That salesman sold me a bill of goods!" has been the distressingly common lament of credit managers.

Investment in Credit Decision Systems Ramping Up

According to the CRF survey, the reasons credit departments gave for not employing credit scoring - that is, not using a predetermined formula for determining creditworthiness - include fears that it would disqualify too many reasonable risks. Or conversely, that it would approve too many bad risks. Other reasons stated were that it's difficult to implement, or simply because, "I hate scoring."

While a quarter of respondents were very comfortable with credit scoring and confident it would help, another quarter were unconvinced it works. Twenty-one percent intuitively believed it would help but couldn't justify the return on investment; more than a third cited concerns about the cost of automating their systems - typically an expense of between $5,000 and $25,000. And nearly 29 percent feared a lack of IT support.

With that said, there are clear signs that the way B2B credit managers do business is changing fast, pushed by improvements in the way credit decision-making systems work, strong pressures to do more with less and Sarbanes-Oxley demands to keep better track of accounts receivables.

Fully 42 percent of credit departments not using credit scoring now, plan to have a system ramped up and running within five years, according to the CRF survey.

Not surprisingly, the largest businesses are paving the way. Businesses with sales of $1 billion or more are most likely to use credit scoring, as are those with 7,500 or more accounts, according to the CRF.

Businesses Need to Be Fast and Accurate to Compete

The most common reason for automating is efficiency, along with other factors such as corporate governance concerns, rising workloads, increased competition, customer service and senior management concerns.

And despite fears perhaps sown during earlier, glitch-hobbled years of credit scoring, there are encouraging signs of improvement. More than 80 percent of credit managers who have deployed credit-scoring systems say they're satisfied with the results, according to the CRF.

The benefits for those who have adopted B2B credit scoring are evident in the CRF survey. Forty-five percent reported a decline in day's sales outstanding (DSO). More than half said it reduced bad debt. Twenty-three percent said it enabled a reduction in head count, and more than half said it sped turnaround time on new customer applications and increased the frequency of customer reviews and credit line updates.

Credit Decisions Blend Art and Science

There's good reason for such salutary effects. As any credit manager knows, making decisions about creditworthiness is an art as well as a science. Businesses selling big-ticket, high-demand products can afford to be very picky about credit. Those selling products with a high margin might be more willing to take calculated risks in the interest of maximizing sales.

It all depends on the business. Top credit managers get to where they are by knowing their industry inside and out and embracing both the science and art of credit decision making in all its nuances.

It is these top performers who devise or strongly influence the rules that govern automated credit decisions.

Statistically Derived Versus Rules-Based Models

Credit-scoring models can be developed in a couple of ways, with advantages to each. The scoring model might be statistically derived. The advantage of statistically derived models is that they are created objectively and are based on actual payment performance data. Using data available in a prior period, such as the previous 12 months, essentially creates these models. The model development process then looks for characteristics and patterns within only the most predictive set of data to determine the most appropriate weighting for these data elements. This process removes biases toward data that may not be truly the most predictive. It also ensures that the data that is the most predictive is weighted properly to give the best assessment of whether an account is likely to go bad.

Conversely, the model might be rules-based, developed from the experience of seasoned analysts. This method requires no historical data and typically assumes a looser definition of a "bad" account. In this case, the analyst decides which data to include in the model, which consequently reflects the analyst's preferences.

Once credit is approved, terms are set and credit lines established, another benefit to automated systems comes into play. The system can develop management reports to track applicants, refreshing data on active accounts to update risk scores and credit lines, prioritizing collections and spotting trends such as accept and decline ratios.

Hosted Application Systems Reduce Costs and Simplify Implementation

On a nuts-and-bolts level, automated credit-scoring systems work best in a networked environment, including a local area network or intranet. But a third option, a hosted application, offers measurable benefits, including reduced IT costs and support needs, faster implementation and a simpler update process.

Companies connect over the Internet to the hosting service, which then draws data from credit reporting agencies such as Experian and fires back a report to the company.

Current pressures driving more and more companies to automate their B2B credit decisions are only likely to heighten. As e-commerce grows, so will the need for speed. Regulatory and corporate governance demands for precise reporting and monitoring of accounts receivables aren't going away, providing all the more reason to go with the automated flow.

There's also geographic flexibility to consider. With wireless technology, key credit decisions can be made from anywhere with an automated system doing the heavy lifting.

Tips for Taking the Automated Plunge

Before you take the automated plunge, however, here are a few points to bear in mind for an easier transition:

* You will need partners. IT must be involved from the get-go. Show sales and marketing how it will help them. Make sure finance understands the return on investment.

* Understand how your current processes work. Do you have a written credit policy? If so, review it carefully.

* Know where you want to go. Do you know what your decision rules are? Are you prepared to turn over the majority of your decisions to an automated system? On the other hand, are you prepared for the reality that not all decisions can be system-generated?

* This is not a "magic pill." If there are bottlenecks elsewhere in the approval process, automated decision-making won't help that. Changes do not occur automatically, so keep three words close to heart: Test, test and test. Prepare to take a step backward before moving forward.

* No technical effort is ever perfect. Expect glitches. What you get may not be exactly what you envisioned. If you forgot a requirement, be ready to negotiate. Make sure you have adequate vendor and internal IT support.

* This is not a one-time effort. The predictive power of any model, if not dynamic and constantly updated, will deteriorate over time.

* Don't forget the "people factor." Engage the people on the floor early. Use the opportunity to elevate positions and skill sets.

* Lastly, if in doubt, consider retaining a consultant.