TheLectern

By Steve Kelman

Blog archive

How can we improve the use of past performance in contracting?

man graphs performance

Jaime Gracia, CEO of Seville Government Consulting, is one of the most thoughtful commenters around on government contracting issues as seen from a progressive contractor perspective -- i.e., one that is also mindful of the interests of the government and the taxpayer. I don't always agree with everything he writes, but I always take his views very seriously.

Jaime and I also share an interest in making government past performance evaluations a more meaningful part of the contracting process. Speaking for myself, I remain convinced that there is no single step the government could take to improve the performance of the contracting system than to make past performance an effective differentiator between excellent, acceptable and poor performance. We should penalize poor performers but also, importantly, reward the good ones. Customer decisions about whom to do business with based on a supplier's past performance are absolutely essential in the commercial market to making the free market work to satisfy customers.

Jaime has written an interesting FCW op-ed called "Why past performance must be part of acquisition reform," and I urge all blog readers to take a look.

Let me start with the parts of Jaime's column with which I agree. He urges that past performance evaluations to the extent possible be based on objective data about cost, schedule and performance rather than the subjective judgments of government officials. I strongly agree with this in general, though I assume Jaime would agree there are limits. There are situations where it is difficult to develop good objective measures at all (particularly of performance), and also those where changes in the original baseline make the evaluation of even objective performance information more subjective. At a minimum, we should recognize the validity of "inter-subjective data" -- the subjective judgments of a larger number of people, e.g. customer satisfaction surveys -- in judging past performance.

I also thought Jamie made a very interesting suggestion about greater public disclosure of project management information already provided the government. I would love to hear the reaction of both government people and contractors to this suggestion. Perhaps this at least could be done on a voluntary basis by the government -- with an agency disclosing its intention to do so in an RFP? How often would such information be considered proprietary?

However, Jaime and I continue to differ somewhat on the justification of the current level of "due process" protections for contractors in the past performance system. He is concerned about subjective or ill-willed negative judgments by government officials, particularly in the current "lowest price technically acceptable" contracting environment. My own reaction is that the bigger danger in the system is grade inflation, not unjustified dinging of contractors.

To some extent, this is an objectively testable proposition -- have we seen an increase in negative past performance ratings in recent LPTA years? (Of course, even if the answer is yes, this may be because performance has gotten worse. Though Jaime would counter, probably correctly, that this is the government's own fault. But if negative ratings aren't going up, this would argue that the system definitely suffers from grade inflation.)

At any rate, I continue to believe that the ability of contractor to contest ratings drives grade inflation, and that a sufficient due process protection is to allow the contractor to put their version of events in the past performance file.

Jaime -- and others -- let's continue this dialogue. I would love to see our community make improving the past performance system a high priority.

Posted by Steve Kelman on Apr 15, 2014 at 10:45 AM


Reader comments

Wed, Apr 16, 2014

Flip side of the coin: We're told to write past performance evaluations like OERs. Make sure the company looks good because any small negative comment might cost them future work and result in layoffs, etc.

Wed, Apr 16, 2014 Owen Ambur Silver Spring, MD

A minimal, first step that might be taken is to publish contract performance plan/report metrics in an open, standard, machine-readable format, like StratML. E.O. 13462 establishes the presumption that government information will be machine-readable by default. http://xml.fido.gov/stratml/carmel/EOOMRDwStyle.xml And OMB M-13-13 reiterates that open data standards should be used. http://xml.fido.gov/stratml/carmel/M-13-13wStyle.xml#_78e85ef4-b91c-11e2-bf2b-79d279ad226c Whether and how effectively contracting officers begin to do so will be a measure of their own performance. Moreover, contractors who are proud of their work could take it upon themselves to do so proactively.

Wed, Apr 16, 2014 a contractor

I think what is also being sugar-coated (or willfully ignored here) is that these contractor evaluations are performed against a set of expectations which themselves may or may not be rational. Consider -- a Federal buyer (programmatic or IT) articulates a set of loose requirements and the contracting shop then makes an award. This award is not made in a vacuum, but we see from the PR beating that GSA is taking that if the CO is not allowed to make the final call somewhat independently, the agency would get it's hand slapped. So there are at least three parties involved, and no clear-cut mechanism ensuring that each of the three parties have unified expectation of what constitutes "success" on the project. Now: Who fills out the contractor's performance rating, and based upon whose expectations and/or definition of success? Until all involved can address getting those definitions unified and documented then how can an assessment of contractor performance be done well? This is why the contractor appeal process has such teeth. These teeth are needed.

Tue, Apr 15, 2014 Bob Woods

A few years ago when I worked for another private sector company we did a statistical analysis, by agency, of ratings given to our company. We discovered a strong bias by agency toward contractor performance. Some agencies are simply not contractor friendly and some are. At the very least it would help to know which is which. Why don't we have a data base of agency performance on giving ratings? Normalizing these data would be helpful for us in industry to know where the "can't win" territories are. In following up for clients on past performance data we find a lot of reasons given by acquisition personnel on why someone got a bad rating. It is not well mapped to actual performance. Larger companies with lots of choices simply "shop" to get the rating that is most desirable.

Please post your comments here. Comments are moderated, so they may not appear immediately after submitting. We will not post comments that we consider abusive or off-topic.

Please type the letters/numbers you see above