"We have discovered serious problems with the FDCA (Field Data Collection Automation) program and I am personally involved in bringing key issues to the surface and developing a way forward. In short, the current situation is unacceptable. The American people expect and deserve a timely and accurate Decennial Census..."
So testified Carlos M. Gutierrez, Secretary of the U.S. Department of Commerce at a hearing yesterday in front of the U.S. Senate Committee on Homeland Security and Governmental Affairs on the status of the 2010 Census. Gutierrez finally awoke to the fact that the 2010 Census is in deep and very deep kimshe.
So serious is the trouble that in a highly unusual mid-session announcement, the U.S. Government Accountability Office (GAO) yesterday designated the 2010 Census Project as High Risk, which is in my opinion about 2 years late, since the program is already in trouble, not potentially in trouble.
The cause of the problem which the Census has been trying to paper over for quite some time is that it depends on 500,000 handheld computers to replace its paper-based collection system. As is always the case, it looked very easy to do on paper, but proved to be harder to do in reality.
The Census reasoning seems to have been along the lines of: if Fed Ex can use handhelds to track packages, why can we do the same for collecting Census data - should be dead easy, right? The idea in itself wasn't not outrageous, as long as the risks involved were clearly understood and managed. The GAO report makes clear - as the GAO has several times in the past - that they weren't (and from reading the report still aren't) on both accounts.
In Gutierrez's testimony, he goes on to state that the Census discovered late last year a "gap" as he calls it "between the capacity to get the work done and the amount of time remaining. One of the main reasons for this gap was significant miscommunication concerning technical requirements between the Census Bureau and Harris [the prime contractor]. The lack of clarity in defining technical requirements was a serious problem especially with regard to the testing and functionality of the handheld devices in a full Census environment. For example, discrepancies arose over data upload times, screen change speed and data storage capabilities."
So let me get this straight - with a little more than six months to go before a full scale dress rehearsal of the system, it was discovered that there was still major miscommunication between the Census and the contractor about basic performance parameters for the device to be used by hundreds of thousands of census takers? Weren't these parameters weren't spelled out in detail in the contract? Or did Harris follow the contract, and now the Census has figured out that what it specified won't do? Did Harris tell them there were problems, but the Census didn't listen? What the hell happened here?
Interestingly, back in November 2005, the Government Communications Systems Division (GCSD) of Harris achieved "a Capability Maturity Model Integration (CMMIÂ®) Maturity Level 3 rating. The Level 3 rating denotes superior process maturity within the division's program management, engineering, quality assurance, and other disciplines, and achievement of this rating has become a competitive differentiator on many government programs." I wonder if this rating helped Harris win the Census contract?
At the very least, I think the division's CMMI rating may need to be re-evaluated, or maybe better, the US government better start looking at what, if anything, SEI CMMI Level 3 actually means in practice.
Alas, the Census provided Harris with an updated set of requirements in mid-January 2008; hopefully they are the correct and technically feasible ones.
In the testimony yesterday, it came out that it may cost another $2 billion to "ensure" that the 2010 Census actually can succeed, on top of the $11.5 billion already allocated to the Census (of which $3 billion was for the IT portion of the Census). It also appears the probability of completing the Census on time is dropping rapidly unless there is a marked turnaround. The dress rehearsal in May will give better indication of the true risk status of the situation.
Gutierrez' also said yesterday, "There is no question that both the Census Bureau and Harris could have done things differently and better over the past couple of years."
What I really want to know is who in management is going to be held accountable for this excess level of risk mismanagement, incompetent communication, and rank amateurism in program and contract management. Or is it business as usual, with "mistakes were made," "we have learned from this experience," blah, blah, blah.
The folks at Government Executive have been following this slowly unfolding big time blunder in the making closely, and you can read more about it here, here and here.