A better path to fair and reasonable pricing

The recent $75.5 million Justice Department settlement with Carahsoft and VMware once again highlights for software companies the inherent danger of selling products to the government. Clearly, the Federal Acquisition Regulation and the various fraud statutes, such as the False Claims Act, should penalize cheating and fraudulent behavior by government contractors because protecting taxpayer money is a cause of the highest order. Unfortunately, sometimes the regulations are so complex and onerous, they can easily ensnare contractors who have no ill intent.

The media reports on the VMware settlement indicate that the company had misrepresented its Commercial Sales Practices, which were submitted via Carahsoft in order to create fair and reasonable pricing on Carahsoft’s Schedule 70 contract with the General Services Administration. I have no special knowledge of VMware’s behavior, but I am familiar enough with the CSP process to know that it is ripe for faulty information.

The point of the CSP submittal is to give a contracting officer enough historic pricing and discounting information to determine that the pricing proposed for the GSA Schedule is acceptable. This is often referred to as a “vertical” price analysis because it only looks at the actual sales of the products in question and does not compare the prices “horizontally” against similar competitive products.

Logically, that approach makes sense because for commercial products it is widely assumed that the price buyers are willing to spend is the best market determinant for a fair price.  

To accomplish the pricing analysis task, GSA has created a form Schedule 70 aspirants must complete, through which they must disclose all their discounting practices to specified classes of customers in the preceding 12 months. And here is where things can quickly get off track. The form requires a software company to specifically list its discounting practices to the following classes of customers: Distributors, Dealers/Resellers, VAR/System Integrators, OEM, State and Local Government, Education/Non-Profits, and National and Corporate Accounts.

Although most of us intuitively know what those types of accounts might be, it is also a fact that few if any software companies classify their sales into those types of hard-and-fast categories -- nor do they necessarily have any type of standardized discounting for each.

For many software accounting departments, an inquiry with a request for the “standard discount” to the VAR/System Integrator channel over the previous 12 months will be met with stony silence.

Invariably, financial people will do what they are trained to do -- apply mathematics to the question and attempt to derive an answer. That answer will appear to be true to the best of everyone’s knowledge. But it will be formulaic and not based on some definitive concrete truth. It will depend on which customers were fit into the vague definition of VAR/System Integrator.

It will also depend on the algorithm used to determine the discount number, which can vary greatly depending on whether an average or median is used and whether unique one-time transactions are excluded.

At the end of the day, numbers will be supplied for all the categories of customers, and their ultimate accuracy will be dependent on the interpretation of who fits which category and the calculations that were used to determine discounting at one moment in time.

Unfortunately, now the trap has been set. The CSP submittals are the basis of representations that the government relies on to determine that pricing is appropriate. When, at a later date, it determines that there are flaws in the CSPs, it might use the False Claims Act to say the government relied on fraudulent information to make payments that it otherwise would not have made -- that the pricing was “defective” because the government relied on false information.

The False Claims Act levels an extreme penalty – as much as $11,000 for each transaction, regardless of the actual value of the transaction. A single transaction for $100 incurs the same fine as a transaction for $1 million. Damages also might be calculated and added to the fines. It adds up very quickly.

Once CSPs have been submitted, the software company is on the hook to monitor its representations for changes and to inform the government immediately when an update is in order. Now Murphy’s Law kicks in. Because the numbers were derived through a one-time-only process (because nobody else has ever asked for that data in that way before), the whole exercise can get lost in the day-to-day urgency of business.

Despite CSPs' importance, people move into new roles or forget how the numbers were achieved, business models change, and products come and go. Very quickly, it can reach the point where the original CSPs no longer make sense.

It's quite easy to imagine how a company could find itself uncertain of how the original numbers were derived. It isn’t intentional fraudulent behavior. It’s human behavior.

In the wake of the iconic $800 toilet seat and other procurement scandals, Congress passed the Federal Acquisition Streamlining Act in 1994. FASA’s major feat was to define the notion of commercial products, state a preference for the procurement of those products and lessen the regulatory burden on their providers. In FASA’s spirit of reducing the regulatory burden and inherent danger of contracting with the government, it would be helpful if the government requested important information in ways that gave commercial companies a better chance to accurately comply.

Because the CSP report is inherently an interpretive exercise, perhaps it could be abandoned for an alternative that achieves the government’s goal of price analysis while aligning more closely with how companies keep their own records.

One suggestion would be to do away with the current CSP discount form and replace it with a simple mathematical algorithm that could be easily reproduced in a standard report at any time. A simple request could involve asking a company to compile all the software sales it transacted under the GSA order ceiling of $500,000 and compare that revenue number to what would have been received if all sales had been done at list price. The difference between the two is the effective average discount rate for all non-large sales.

If the government could look at that rate and ensure that its proposed discount was equal to or better than that, would that not meet the objective of fair and reasonable pricing? Furthermore, the software company and the government would have an effective and easy way to measure compliance at any time.

The current CSP system does not work well for anybody. The interpretive nature of the exercise means the government could routinely be getting data that is well intentioned but suspect. Furthermore, the fact that any CSP can form the foundation of a very expensive False Claims Act undermines the spirit of FASA, has the potential to chill commercial companies’ willingness to contract with the government and could ensnare companies that are not willfully misrepresenting their pricing information.

It is time to move to a more simplified and fair way to measure pricing equity -- one that serves the purpose of getting the government a fair price while not unfairly trapping an otherwise responsible contractor.