Frank A. Bloomer-Agency Tender Official
B-401482.2,B-401482.3, Oct 19, 2009
Frank A. Bloomer, Agency Tender Official (ATO), protests the Department of the Army's award of a contract to The Ginn Group, Inc., of Peachtree City, Georgia, to perform public works functions at the United States Army Garrison at West Point, New York, following a public-private competition conducted under Office of Management and Budget (OMB) Circular No. A-76 under request for proposals (RFP) No. W912SU-08-R-0001. The protester basically alleges that the Army performed an unreasonable cost realism analysis of Ginn's proposal.
We sustain the protest on the basis that the cost realism analysis on which the cost comparison relied was materially flawed.
B-401482.2; B-401482.3, Frank A. Bloomer--Agency Tender Official, October 19, 2009
DOCUMENT FOR PUBLIC RELEASE
The decision issued on the date below was subject to a GAO Protective Order. This redacted version has been approved for public release.
William H. Venema, Esq., Epstein Becker Green Wickliff & Hall, P.C., for The Ginn Group, Inc.; and Angela B. Styles, Esq., and Gunjan R. Talati, Esq., Crowell & Moring, LLP, for Don Hale, designated employee agent, the intervenors.
Lt. Col. Vince T. Vanek, Geraldine Chanel, Esq., Peter D. DiPaola, Esq., Charlene T. Storino, Esq., and Raymond M. Saunders, Esq., Department of the Army, for the agency.
Paul N. Wengert, Esq., and Ralph O. White, Esq., Office of the General Counsel, GAO, participated in the preparation of the decision.
Protest challenging a public-private competition between an agency tender and a private sector proposal is sustained where: (1) the agency unreasonably accepted the private-sector offeror's revised fringe benefit ratios in its cost realism analysis; (2) the record provides no reasonable basis for the agency to accept the private-sector offeror's unsupported assumption that the firm could perform a significant portion of the workload 10 percent more efficiently; and (3) the agency unreasonably allowed the private-sector offeror to omit the labor cost associated with the material supply function from its cost proposal, and these errors prejudiced the protester.
Frank A. Bloomer, Agency Tender Official (ATO), protests the Department of the Army's award of a contract to The Ginn Group, Inc., of Peachtree City, Georgia, to perform public works functions at the United States Army Garrison at West Point, New York, following a public-private competition conducted under Office of Management and Budget (OMB) Circular No. A-76 under request for proposals (RFP) No. W912SU'08'R'0001. The protester basically alleges that the Army performed an unreasonable cost realism analysis of Ginn's proposal.
On September 27, 2006, the Army announced a public-private competition to decide whether to perform public works functions at the West Point Army Garrison with government employees in a most efficient organization (MEO) or to procure these services from a private-sector firm.
On September 2, 2008, the Army issued the RFP, seeking an agency tender and private-sector proposals to perform the public works functions. The RFP contemplated either the issuance of a performance agreement with the MEO, or award of a cost-plus fixed-fee contract with a private-sector offeror, for a base year and four annual options.
Each offeror was required to explain its approach to performing a performance work statement (PWS), and in particular, to describe its management approach and provide both a phase-in plan and a quality control plan. The RFP provided for evaluation on the basis of two non-cost factors: mission capability and past performance, which were equal in importance. Under the mission capability factor, the RFP listed four subfactors: management/organizing/staffing, technical approach, phase-in plan, and quality control plan. The agency anticipated the use of adjectival ratings of acceptable, marginal, and unacceptable, to rate each proposal under the mission capability factor and subfactors. The RFP also stated that all subfactors were critical subfactors, which meant that a marginal or unacceptable rating in any subfactor would carry forward to the factor rating. RFP at 90.
Under the past performance factor, the RFP anticipated ratings of low, moderate, high, and unknown risk. RFP at 90. The RFP also stated that the agency tender would not be required to submit past performance information, but would be assigned a neutral rating. RFP at 83.
Finally, the RFP provided that only technically acceptable proposals would be considered for a performance agreement or award, and that, among acceptable proposals, cost would be the controlling factor. RFP at 90. In short, the RFP anticipated issuance of a performance agreement or award of a contract on the basis of the lowest-cost technically-acceptable proposal. AR at 2.
The RFP instructed each offeror to provide a detailed explanation of its approach in the mission capability section of its proposal and to address as specifically as possible the actual methodology [the offeror] would use for accomplishing the PWS. RFP at 82. Additionally, offerors were instructed that any assumptions used in proposal preparation had to be separately identified in their proposals. Id.
A significant portion of the work in this RFP was expected to be generated by issuance of standard operating orders (SOO), service orders (SO), and individual job orders (IJO). As relevant to the protest issues, the PWS estimated that the contractor could expect to receive over 20,000 SOs annually, but did not provide specific information on the workload resulting from the SOs. For IJOs, the PWS distinguished between those that would be submitted, and a smaller subset that would be authorized--i.e., the PWS estimated that over 800 IJOs would be submitted, and that approximately 343 IJOs would be authorized each year. RFP amend. No. 10, Revised PWS, at 129. The RFP also provided a table estimating the hours required to perform the IJOs under each of 16 different trades, and required that offerors use this data to develop their proposals. Revised PWS at 147.
Another requirement under the RFP was to provide material and supply services, including furnishing, maintaining, and replacing materials and supplies. Revised PWS at 58, 68-69. The RFP also provided for the service provider to replenish supplies, and be reimbursed for the cost of materials. Revised PWS at 55. In response to a question posed by an offeror, the Army amended the RFP to remove a requirement to estimate the cost of materials and supplies in each offeror's cost proposal; instead, the Army directed offerors to use a plug number of $500,000 per year as the reimbursable cost of materials and supplies. RFP amend. No. 10 (Answer to Question 10). Nevertheless, offerors were required to provide all labor to perform these services.
Each offeror was also required to submit a cost proposal, detailing the costs it expected to incur in performance. The RFP stated that the Army would conduct a cost realism analysis in accordance with Federal Acquisition Regulation (FAR) sect. 15.404-1(d), in order to determine the realism of the cost proposal. The RFP also informed offerors that the Army might adjust the cost of any private-sector proposal for purposes of evaluation pursuant to sect. 322 of the National Defense Authorization Act for Fiscal Year 2008 (now codified at 10 U.S.C. sect. 2461(a)(1)(G) (2006 & Supp. II 2008)). The RFP explained that this statute requires the agency to ensure that private-sector offerors do not obtain a competitive advantage over agency tenders by proposing lower costs for health insurance benefits or retirement benefits than the cost of those benefits provided to civilian employees by the Department of Defense. RFP at 80. The Army approach to implementing this statutory requirement will be set forth in greater detail below.
On November 24, 2008, the ATO submitted a timely public tender in response to the solicitation. Three other offerors, including Ginn, submitted timely proposals. The initial evaluation rated all offerors marginal or unacceptable. Therefore, on March 4, 2009, the Army opened discussions with all competitors, including the ATO and Ginn. AR, Tab 10A, Discussions Letter to ATO, Mar. 4, 2009, at 1; AR, Tab 15A. After receiving revised proposals, the Army conducted a second round of discussions. Second Discussions Letter to ATO, Mar. 16, 2009, at 1.
Set forth below is a more detailed explanation of the relevant portions of the evaluation record related to the three issues addressed in this decision.
Comparison of the Fringe Benefits
The RFP required each offeror to provide detailed fringe benefit cost information to permit the comparison required by 10 U.S.C. sect. 2461(a)(1)(G). Specifically, the RFP instructed the offeror to set forth in its cost proposal the direct labor costs for each of the five general functional areas of the PWS, and then show the addition of the corresponding costs for fringes/overhead, subcontracts, other direct costs, general and administrative costs, and profit. Further, the fringe/overhead costs were to be shown as three separate cost lines: one for insurance and health benefit costs, one for retirement benefit costs, and one for all other overhead and fringe benefits. The RFP stated that the all other costs line was for all other labor overhead costs not captured in either the health benefits or retirement benefits categories. RFP at 80'81.
The RFP instructed offerors that the health benefits and the retirement benefits costs must be consistent with the costs in the agency tender. RFP at 81 (citing OMB Circular No. A-76, Attachment C, para. B(2)(f)(1)(a)-(b)). Offerors were requested to express the fringe/overhead lines in two ways: as a percentage of the total direct labor costs, and as a dollar amount. Finally, the RFP instructed offerors that the all other category of direct labor fringe/overhead costs should reflect the offeror's accounting policies, procedures and practices and be in accordance with their indirect rate structure. RFP at 81.
To implement the requirements of 10 U.S.C. sect. 2461(a)(1)(G), the RFP stated that the Army would
[C]ompute the ratio of a private sector offeror's health insurance contribution to its direct labor costs, (including prime contractor and all subcontractor employees' total salaries and wages for the proposed contract); to determine if the ratio is equal to, or greater than, the standard health benefit cost factor used in the agency cost estimate.
Compute the ratio of a private sector offeror's civilian retirement contribution (the offeror and all subcontractors) to its direct labor costs; to determine if the ratio is equal to, or greater than, the standard civilian retirement benefit cost factor used in the agency cost estimate.
RFP at 92-93.
The RFP explained further that if either ratio (that is, the health insurance contribution or the retirement contribution) was less for the private-sector offeror than for the agency cost estimate, the Army would adjust--for purposes of the cost comparison--the private-sector offeror's costs upward. Id. at 93.
Ginn's initial proposal complied with these instructions--i.e., Ginn provided information on the hourly labor costs to perform each function of the PWS, and the associated cost of fringe benefits, which Ginn also calculated as a percentage of the total labor cost. AR, Tab 6, Ginn Initial Cost Proposal Spreadsheets. These ratios were shown in Ginn's initial cost proposal, thus:
Insurance & Health Benefits Cost
Retirements Benefits Costs
All other fringe/overhead
AR, Tab 6, Ginn Initial Cost Spreadsheets, Phase-in Sheet, at 1.
In evaluating these proposed costs, the Army's evaluator observed that Ginn's fringe benefit rate did not include comparable costs for the firm's health insurance contribution (which was slightly higher) and retirement contribution (which was significantly lower) when compared to those provided by OMB for use in the agency estimate. The evaluator therefore concluded that where Ginn's fringe benefit ratios were lower than the OMB ratios, Ginn's costs for evaluation purposes should be increased to the OMB level, pursuant to 10 U.S.C. sect. 2461(a)(1)(G).
During discussions, Ginn was informed of the OMB ratios, and was informed that its proposed costs would be increased for purposes of comparison to the agency tender. AR, Tab 10B, Ginn Discussions, at 7. In response to this first discussions exchange, Ginn raised its (and Emcor's) total fringe benefit ratios to a total of 36.25 percent, but did not change any of the underlying retirement benefit ratio, which remained as shown in the initial proposal. See, e.g., AR, Tab 12, Ginn First Revised Cost Spreadsheets, Phase-in Sheet, at 1.
In the second round of discussions the Army stated that the retirement benefit ratios for both Ginn and Emcor were still below the OMB ratio, and would be increased for purposes of comparison to the agency tender. AR, Tab 15B, Ginn Final Discussions, at 3. In response to the second round of discussions, Ginn stated that it had revised its cost proposal to propose a retirement benefit cost factor of 26.1 percent. AR, Tab 17A, Ginn Final Discussions Response, at 2. However, the final cost proposal spreadsheets showed that this change had not affected the total fringe benefit rate; instead Ginn simply reduced the other fringe benefit rates by an offsetting percentage, as shown below:
Insurance & Health Benefits Cost
Retirements Benefits Costs
All other fringe/overhead
[Total of Ratios]
AR, Tab 17, Ginn FPR Cost Spreadsheets, Phase-in Sheet, at 1.
Despite these responses, Ginn explained that its [a]c[tu]al billing for performance will be in accordance with our DCAA approved Provisional Billing Rates as provided in Revision 1 of our pricing proposal. AR, Tab 17A, Final Ginn Discussions Responses, at 2.
In reviewing Ginn's FPR, the Army cost evaluator noted that Ginn had dramatically reduced the all other fringe/overhead cost component to 3.15 percent for both Ginn and Emcor, and had reduced the health insurance benefit component from 7.1 percent to 7.0 percent for both firms. The evaluator concluded that Ginn's compliance with 10 U.S.C. sect. 2461(a)(1)(G) thus appears to be little more than moving the numbers around for purposes of the evaluation. AR, Tab 23B, Final Ginn Cost Evaluation, at 4. Despite this conclusion, the Army's price negotiation memorandum (PNM) simply describes the mechanics of Ginn's change, but does not question it. AR, Tab 24, PNM, at 22. The PNM then states that the evaluators deemed both the agency tender and Ginn to have submitted prices [that] are realistic, reasonable, and complete. Id. at 25.
Ginn's Assumed 10 Percent Efficiency Improvement
Ginn identified several key assumptions in the Management/Staffing/Organization & Technical Approach volume of its proposal. One of these was that Ginn would perform IJOs and SOs 10 percent more efficiently than the government workload estimate. AR, Tab 6, Ginn Initial Proposal, vol. D at 48.
In the initial evaluation, the agency rated Ginn unacceptable under technical approach for several reasons. Of relevance here, the Army questioned Ginn's claim that it could achieve a 10 percent efficiency improvement, noting that the RFP did not provide specific data about the SO workload from which Ginn could determine how it might achieve increased efficiencies. AR, Tab 7, SSEB Initial Report, at 19.
During discussions, the Army asked Ginn to explain the supporting rationale for its claimed efficiency improvement. AR, Tab 10B, Ginn Discussions, at 5. In response, Ginn simply stated that it based its 10 percent efficiency improvement for IJOs on the data in the RFP, and applied this rate of improvement to SOs, given its experience with similar projects and scopes of service. AR, Tab 12, Ginn Discussions Responses, at 5.
Ginn's revised proposal attempted to further illuminate its key assumption, as follows:
The productive hour data provided by the Government is used as the basis for our proposed productive hours for these requirements. From our experience on similar projects, we have demonstrated the ability to perform this same work with lower hours. We have applied an efficiency factor of 10% where applicable to the hours indicated in the development of our estimate.
AR, Tab 12, Revised Ginn Proposal, vol. D, at 55.
In the final evaluation, the Army evaluators did not explain how Ginn had addressed their earlier concerns. Instead, the evaluators concluded under each evaluation factor that the weaknesses identified . . . were adequately addressed . . . or were corrected. PNM at 10-11. With regard to the technical approach subfactor, the final evaluation broadly concluded that Ginn rewrote the Technical Approach proposal section and provided specific details as to how they intend to successfully complete the PWS requirements. PNM at 10. The evaluators then rated Ginn acceptable. In short, the final evaluation does not mention Ginn's continued reliance on a 10 percent efficiency improvement as a key assumption, or how the agency's earlier concerns about this assumption were resolved.
Supply Support Staffing Cost
As mentioned briefly above, Ginn proposed to utilize a second subcontractor, Sunbelt, to perform the supply support function. Ginn described its approach as creating a private hardware store, and asserted that Sunbelt will provide staffing to support our material requirements as part of their operation without adding to the cost of the material items. AR, Tab 6, Ginn Initial Proposal, vol. D, at 48; AR, Tab 12, Ginn Revised Proposal, vol. D, at 54. However, Ginn also explained that this arrangement would be like walking into the local hardware store with the understanding that the store employees are included in the cost of the item. Ginn emphasized that the benefit of this technique was that it allows us to show their staffing on our organizational chart without pricing them in our cost proposal. Id.; AR, Tab 12, Ginn Revised Proposal, vol. D, at 54.
In response to a question about a related issue during discussions, Ginn explained that Sunbelt would provide five on-site full-time equivalent employees (FTE), while Ginn's team would provide one FTE to oversee that work. AR, Tab 12A, Ginn Discussions Responses, at 3. Ginn's revised proposal continued to assert that the cost of the Sunbelt employees were properly omitted from the cost proposal, and the cost proposal reflects no cost for those FTEs. AR, Tab 12, Ginn Revised Proposal, vol. D, at 54; AR, Tab 17, Ginn Final Cost Proposal Spreadsheets, Sheet C-5.1 DLT at 5 & Sheet C-5.1 Roll-up at 2, 4, 6, 8 & 10.
As noted earlier, the final evaluation did not address most of the specific issues identified in the initial evaluation, but included only the broad conclusions quoted above--i.e., that Ginn had rewritten its proposal and provided specific details as to how they intend to successfully complete the PWS requirements. AR, Tab 24, Price Negotiation Memorandum, at 10. Specifically, neither the initial nor the final evaluation addressed whether these costs were properly omitted from Ginn's cost proposal.
Source Selection Decision
The source selection decision document (SSDD) stated that Ginn's FPR had no weaknesses, had proposed a very logical and methodical approach, and was sound and provide[d] specific details, methods and procedures as to how they will successfully accomplish the requirements within the PWS. With respect to the cost evaluation, the SSDD simply stated that all of Ginn's proposed costs were realistic, reasonable, and complete. AR, Tab 26, SSDD, at 5.
The agency tender was calculated to cost $68,274,080. AR, Tab 26, SSDD, at 3. For purposes of making the comparison to the agency tender, Ginn's most probable cost, including costs of administration and one-time conversion costs, was calculated as $58,246,062. When the 10 percent conversion differential ($5,833,318) was added, Ginn's cost for comparison purposes was $64,079,380 and thus the cost comparison favored Ginn's proposal by $4,194,700. AR, Tab 25, Standard Competition Form Worksheets, at 3-4.
Based on this comparison, the source selection authority selected Ginn's proposal as the lowest-cost technically-acceptable proposal. AR, Tab 26, SSDD, at 7. On March 25, the Army notified the ATO that the cost comparison had resulted in a tentative decision to award a contract for the public works function to Ginn. AR, Tab 27, Memorandum to ATO, Mar. 25, 2009, at 1.
On May 6, the ATO received a debriefing. The ATO then filed an agency-level protest on May 18, which the Army denied in a decision dated June 29. The ATO then filed this protest with our Office on July 9. After receiving an agency report on August 10, the ATO filed a supplemental protest on August 20. This decision resolves both the initial and supplemental protests.
Our decision focuses on three areas where, in our view, the record does not support the reasonableness of the evaluation. In brief these issues are that (1) the Army unreasonably accepted Ginn's revised fringe benefit ratios in its cost realism analysis; (2) the record provides no reasonable basis for the Army to accept Ginn's unsupported assumption that the firm could perform the SO and IJO workload 10 percent more efficiently; and (3) the Army unreasonably allowed Ginn to omit the FTE's associated with the material supply function from its cost proposal. In light of these conclusions, the award to Ginn is not supported by the record.
In any federal procurement, including a procurement conducted pursuant to a cost comparison under OMB Circular No. A-76, source selection officials are bound by the fundamental requirement that their judgments be reasonable, consistent with the stated evaluation scheme, and adequately documented. DynCorp Int'l LLC, B'289863, B'289863.2, May 13, 2002, 2002 CPD para. 83 at 4. An agency is not required to base its technical evaluation on a company's reputation, or accept unsupported statements of capability, especially where an RFP requires the offeror to explain and support its proposed approach. L-3 Commc'ns Corp., B-299014, B-299014.2, Jan. 16, 2007, 2007 CPD para. 26 at 8.
In addition, when an agency evaluates a proposal for the award of a cost-reimbursement contract, an offeror's proposed estimated costs are not dispositive because, regardless of the costs proposed, the government is bound to pay the contractor its allowable costs. FAR sections 15.305(a)(1), 15.404-1(d); Palmetto GBA, LLC, B-298962, B-298962.2, Jan. 16, 2007, 2007 CPD para. 25 at 7. Consequently, the agency must perform a cost realism analysis to determine the extent to which an offeror's proposed costs are realistic for the work to be performed. FAR sect. 15.404'1(d)(1).
Fringe Benefit Revisions
As explained above, in conducting a cost comparison pursuant to the use of OMB Circular No. A-76, Department of Defense activities are required to assess the cost of fringe benefits incurred by a private sector offeror, compare those to the OMB-generated cost factors that correspond to those costs for employees of the MEO, and make any needed adjustments to the private-sector offeror's costs before making the comparison. Specifically, the statute requires as follows:
A function of the Department of Defense performed by 10 or more Department of Defense civilian employees may not be converted, in whole or in part, to performance by a contractor unless the conversion is based on the results of a public-private competition that --
* * * * *
(G) requires that the contractor shall not receive an advantage for a proposal that would reduce costs for the Department of Defense by -
(i) not making an employer-sponsored health insurance plan (or payment that could be used in lieu of such a plan), health savings account, or medical savings account available to the workers who are to be employed to perform the function under the contract;
(ii) offering to such workers an employer-sponsored health benefits plan that requires the employer to contribute less towards the premium or subscription share than the amount that is paid by the Department of Defense for health benefits for civilian employees of the Department under chapter 89 of title 5; or
(iii) offering to such workers a retirement benefit that, in any year, costs less than the annual retirement cost factor applicable to civilian employees of the Department of Defense under chapter 89 of title 5[.]
10 U.S.C. sect. 2461(a)(1)(G).
The protester argues that the Army allowed Ginn to substitute nominally compliant fringe benefit ratios, which were achieved, in essence, by shifting fringe benefit costs from one category to another, without regard to its actual costs for non-health and non-retirement fringe benefits. The protester points out that those costs were initially identified as 15.63 percent of Ginn's labor costs, and 14.21 percent of Emcor's labor costs, and the protester argues that these costs will have to be reimbursed by the government when incurred. Protest at 12.
The Army only generally disputes the protester's argument on the merits, and mainly argues that the impact of the issue is not sufficient to overcome the cost comparison differential. Supp. AR at 26. In this regard, the Army asserts, in essence, that even if the protester's arguments are correct, it was not prejudiced by any error here. The Army submits a memorandum from a DCAA financial liaison advisor, which quantifies the recalculation of Ginn's fringe benefits as adding $4,097,427 in costs to Ginn's proposal. Supp. AR, exh. 3, Memorandum from DCAA Financial Liaison Advisor, at 1. The Army argues that this amount would not change the result of the cost comparison study. Supp. AR at 26.
Ginn, on the other hand, argues that the fringe rates used to comply with 10 U.S.C. sect. 2461(a)(1)(G) should be understood as comparison numbers for purposes of the evaluation. Ginn reasons that when the Army warned that 10 U.S.C. sect. 2461(a)(1)(G) required it to increase Ginn's and Emcor's retirement benefit rate, it was appropriate for the firms to make offsetting reductions in their other fringe rates because the whole effort was to comply with 10 U.S.C. sect. 2461(a)(1)(G) for evaluation purposes. Ginn also emphasizes that it was clearly understood that the rates used to calculate its FPR cost proposal were not the actual rates to be charged under the contract. Ginn Supp. Comments at 23. Ginn also argues that it is sufficient that Ginn's overall fringe rate was increased to 36.25 percent for evaluation purposes, and that there is insufficient guidance available to establish the proper way to calculate a private-sector offeror's fringe rates for comparison purposes. Ginn Supp. Comments at 26.
When the agency's cost evaluator reviewed Ginn's technique, he characterized the approach as simply moving the numbers around for evaluation purposes. AR, Tab 23B, Final Ginn Cost Evaluation, at 4. We agree.
In our view, the approach taken by Ginn to address the requirements of 10 U.S.C. sect. 2461(a)(1)(G) essentially renders meaningless the goal of the statute. Ginn's cost savings in offering lower-cost retirement benefits are simply relabeled as lower costs in other areas. In this regard, there is nothing in this record to support a conclusion that Ginn has actually reduced its employees' benefits in other areas--such as sick leave, vacation time, and other non-retirement, non-health insurance related costs. Without such a showing, these costs may still be incurred, and when incurred will become reimbursable under the cost-reimbursement contract anticipated here. Thus the Army could not reasonably conclude that the new ratios accurately reflect Ginn's accounting policies, procedures, and practices, and were in accordance with its indirect rate structure, as the RFP required. Cf. RFP at 81.
For the comparison of fringe benefits rates to be meaningful, either Ginn must show how its other costs will be reduced, or the comparison must hold steady the other fringe benefit costs, while increasing the insurance/health and retirement benefit fringe rates to the appropriate ratios. Since a cost realism analysis must account for all costs that will be incurred, and the analysis here provides no meaningful basis to accept Ginn's revised all other fringe/overhead rates, the cost realism analysis is unreasonable. See E.L. Hamm & Assocs., Inc., B'280766.3, Apr. 12, 1999, 99-1 CPD para. 85 at 10 (protest sustained where cost realism analysis provided no basis to accept offeror's significantly lower overhead rate rather than its higher DCAA-approved overhead rate).
Ginn's Efficiency Assumption
With respect to Ginn's assumption that it will achieve a 10 percent increase in efficiency in performing SOs and IJOs, the protester argues that the Army should not have accepted this assumption because Ginn provided no support for this claim. Supp. Protest at 5-8. The Army argues that with respect to SOs, the evaluators were not trying to make a value judgment as to [Ginn]'s claim of a 10 [percent] efficiency factor because such a judgment was not necessary. Supp. AR at 13. With respect to Ginn's assumption of 10 percent efficiency over the IJO workload figures in the RFP, the Army argues that offerors were not required to use the IJO workload stated in the RFP, and asserts that the evaluators found Ginn's assumption, and resulting staffing levels to be acceptable because they met the standard. Supp. AR at 16.
In our view, the record provides no support for the Army to accept Ginn's assumed 10 percent improvement in efficiency, either as a technical approach or more importantly, in support of using lower staffing numbers in its cost proposal. As the Army noted in its initial evaluation, the RFP did not provide sufficiently detailed SO workload information for an offeror to assess its efficiency. With respect to the IJO workload, as noted above, the PWS stated that offerors shall use the workload listed below to develop their individual proposals. Revised PWS at 147. The solicitation also required offerors to address as specifically as possible the offeror's actual methodology to accomplishing the PWS. The evaluators questioned the basis for Ginn's assumption of increased efficiency, and raised the issue in discussions. Yet Ginn's revised proposals provided no factual support for its increased efficiency assumption. In short, the Army had no basis to accept Ginn's key assumption that it could perform either the SOs or the IJOs with 10 percent fewer labor hours and a correspondingly lower cost. By nevertheless accepting the cost proposal based on this assumption, the Army's cost realism analysis was unreasonable. See Northrop Grumman Info. Tech., Inc., B-400134.10, Aug. 18, 2009, 2009 CPD para. 167 at 6-7 (protest sustained where agency did not consider evaluator's concern that awardee's claim of staffing efficiencies to justify lower staffing was unsupported).
In attempting to calculate the effect of this issue on the cost comparison, the protester asserts that the additional cost to perform SOs without the 10 percent efficiency factor was more than $500,000, and the additional cost to perform IJOs was more than $1 million. Supp. Protest at 7-8. Although the Army and Ginn dispute the protester's argument, neither has provided an alternative calculation for assessing the degree to which the protester was prejudiced. Our Office reviewed Ginn's FPR cost proposal and we are unable to determine the precise amount of the additional cost for Ginn's performance of SOs. However, since IJOs were priced as a separate line item, based on our review of Ginn's FPR cost proposal spreadsheet, we conclude that the 10 percent reduction in labor costs for efficiency unjustifiably reduced its costs by more than $1 million, over the life of the contract.
Supply Support Cost
Finally, the protester argues that it was unreasonable for the Army to accept Ginn's claim that it could omit the cost of the FTEs for supply support provided by Sunbelt in the cost proposal. The protester argues that under the terms of its proposal, Ginn has stated that it will include those labor costs in the price of supplies at Sunbelt's private hardware store, and thus the realistic cost of performance by Ginn had to include those costs. Supp. Protest at 4.
The Army asserts that it is undisputed that [Ginn] proposed not to charge the Army for the cost of the Sunbelt employees and not to charge more for the material items in order to recoup the cost of those employees. Supp. AR at 10. The Army also argues that there is no support for the protester's claim that the Army will pay a marked-up cost for supplies. Therefore the Army reasonably expects that Sunbelt would be willing . . . to provide labor and still charge reasonable prices in exchange for serving as the exclusive source of materials and supplies at West Point. Id. The Army also argues that it intends to incorporate the relevant portion of Ginn's proposal into the contract, and will hold Ginn to those terms. Id.
Contrary to the Army's argument, Ginn's proposal states that Sunbelt is basically providing a private hardware store, and will provide staffing without adding to the cost of the material items. However, Ginn acknowledges that--as in any retail environment--the costs of the store employees are included in the cost of the item. AR, Tab 12, Ginn Revised Proposal, vol. D, at 54. This latter statement suggests that Ginn's costs for materials and supplies, which will be reimbursed by the Army, will include labor costs. In our view, the analysis must distinguish between reasonable retail prices at a hardware store and the cost of supplies that a contractor would be allowed to recover under a cost-reimbursement contract, which the plug number represents. We also think the Army cannot reasonably argue that it can steadfastly limit the prices at Ginn's private hardware store when the proposal appears to provide otherwise. In short, the Army's cost realism analysis does not reasonably account for the cost of performing supply services.
The protester asserts that the wage cost of the FTEs, based on the applicable wage determination, would be $610,128.79 over the term of the contract. Supp. Protest at 5. Again, while the Army and Ginn dispute the protester's position on the merits, neither has provided an alternative calculation for purposes of assessing the prejudice from this flawed evaluation.
To succeed in a protest against a cost comparison under OMB Circular A-76, the protester must demonstrate not only that the agency failed to follow established procedures, but also that its failure could have materially affected the outcome of the cost comparison. Trajen, Inc., B-284310, B-284310.2, Mar. 28, 2000, 2000 CPD para. 61 at 3; Dyneteria, Inc., B-222581.3, Jan. 8, 1987, 87-1 CPD para. 30 at 2.
As noted above, the cost comparison here resulted in a decision that Ginn's proposal offered the lowest costs by $4,194,700. In our view, the errors identified above call into question the savings that the Army calculated would be achieved by awarding a contract for public works services to Ginn. These are (1) $4,097,427 (for the increased retirement benefit cost), (2) more than $1 million (for the unrealistic 10 percent efficiency assumption), and (3) approximately $610,128.79 (for the supply function). Although the exact figures are not entirely certain from this record, the sum of these amounts significantly exceeds the cost comparison differential. Accordingly, in our view the protester was competitively prejudiced by the errors in the evaluation, and we sustain the protest.
In sustaining a protest challenging an agency's evaluation of proposals, we ordinarily recommend that the agency review the procurement and take appropriate actions to correct any improprieties. Under the specific circumstances here, our recommendation would be for the Army to reasonably reevaluate Ginn's costs, including reassessing whether Ginn provided a reasonable basis in its proposal for the savings in the areas identified in our decision, including holding discussions with Ginn and the other offerors if additional supporting information would be needed for a meaningful reevaluation.
The protester points out that such a recommendation in this case would be barred by sect. 8023 of the Consolidated Security, Disaster Assistance, and Continuing Appropriations Act, 2009, which states as follows:
None of the funds appropriated by this Act shall be available to perform any cost study pursuant to the provisions of OMB Circular A-76 if the study being performed exceeds a period of 24 months after initiation of such study with respect to a single function activity or 30 months after initiation of such study for a multi-function activity.
Consolidated Security, Disaster Assistance, and Continuing Appropriations Act, 2009, Pub. L. No. 110-329, div. C, title VIII, 122 Stat. 3619, 3626 (Sept. 30, 2008). As relevant here, the effect of the provision is to bar the Army from using funds appropriated under the statute to perform any multi-function study (as was performed here) more than 30 months after the study was initiated. In our view the 30-month deadline has plainly passed for this study.
Accordingly, because any recommendation by our Office to correct the evaluation here would result in the Army expending funds to continue to perform the study at issue, we will not make such a recommendation. New Dynamics Corp., B-401272, July 8, 2009, 2009 CPD para. 150 at 13-14; see also Rosemary Livingston--Agency Tender Official, B'401102.2, July 6, 2009, 2009 CPD para. 135 at 13-14, recon. denied, Department of the Navy--Request for Modification of Remedy, B-401102.3, Aug. 6, 2009, 2009 CPD para. 162 at 4. Since it appears from the record that the contract may not have been awarded, see Supp. Contracting Officer's Statement, at 1, we recommend that the Army cancel the RFP and not proceed with the award of a contract to Ginn. However, if a contract has been awarded to Ginn, we recommend that the Army terminate it for the convenience of the government.
The protest is sustained.
Lynn H. Gibson
Acting General Counsel
 The marginal rating was to be used for a proposal that demonstrates shallow understanding of requirements and proposes an approach that does not clearly meet or, in some instances, falls below minimum PWS requirements. . . . Such a proposal was not . . . eligible for award unless these deficiencies and/or major weaknesses are corrected. RFP at 91.
 The PWS defined an SOO as operating instructions for plant operation, operator maintenance, recurring preventive maintenance services, or other services where specific work and manpower requirements are relatively constant and predictable in advance. Revised PWS at 39. This category is not otherwise relevant to the issues addressed in this decision.
 The PWS defined an SO as any maintenance action requiring a maximum of 40 hours of labor, typically small minor maintenance and repair jobs and emergency work such as broken plumbing or electrical failures. SOs also include temporary installations, new installations, temporary services, support for special events, vigils, and security and fire alarm acceptance tests. Revised PWS at 62-63.
 The PWS defined an IJO as work expected to take between 40 hours and 300 hours, and described these orders as appropriate for project oriented work such as repairs, modifications, replacements or installations. The PWS noted that IJOs can vary significantly in nature and scope and may involve multiple crafts and locations, and also advised that an IJO could take up to 5 weeks to complete. Revised PWS at 63.
 In this decision, citations to the Revised PWS are to the final version issued as part of amendment 10 to the RFP.
 The RFP identified several elements of supply services, including stock control, managing residential furnishings, and managing a self-help center. Revised PWS at 68-69.
 Ginn proposed to use two subcontractors to provide these services. The first, Emcor, was to perform several aspects of the work, and Emcor's rates for fringe benefits are discussed in this section of our decision. The second subcontractor, Sunbelt Industries, Inc., was to perform only supply support services, which are discussed separately below. Ginn did not provide fringe benefit information for Sunbelt.
 In some instances, Ginn's proposal refers to its second subcontractor as Sunbelt Industries, Inc., and in others as Sunbelt Technologies. While we note the discrepancy, we need not address it further.
 Ginn used the term TGGT or The Ginn Group Team to refer to both itself and its primary subcontractor, Emcor.
 The conversion differential does not represent 10 percent of the agency tender's total proposed costs because, among other things, the amount of the solicitation's plug number for supplies is not included in the calculation of the conversion differential. 10 U.S.C. sect. 2461(a)(1)(F)(i); OMB Circular No. A'76, Attachment B, para. D(5)(c)(4)(c) (The conversion differential is . . . the lesser of 10 percent of the MEO's personnel-related costs . . . or $10 million . . .).
 During the public review period that followed, the ATO requested information from the Army about the Ginn proposal. In response to the ATO's request, on April 7 the Army provided a CD-ROM with certain information about Ginn's proposal. An agency-level protest filed on April 22, based on the facts gleaned from that information, was denied by the Army on June 2. The ATO then protested to our Office. We resolved that protest in an earlier decision. Frank A. Bloomer--Agency Tender Official, B-401482, July 20, 2009, 2009 CPD para. 174.
 The protester here also objected to the terms of the RFP and to the conduct of discussions. While the Army has argued that many of these issues were untimely, and all of them lack merit, in light of our decision below we do not address these issues.
 The Army notes that, although the 15.63 percentage rate is reflected in Ginn's narrative, the firm actually used a lower amount, 14.90 percent, in calculating its spreadsheets. The Army argues that the lower amount from the spreadsheets is proper, for cost analysis purposes, but provides no explanation for this conclusion. We note that 14.90 percent appears close to the mean (14.92 percent) of the rates that Ginn and Emcor submitted in the initial proposal. Since, at the conclusion of this decision, we find that the impact of the errors identified above is sufficient to show competitive prejudice, and the contemporaneous record does not clarify which factor is proper, we note the discrepancy, but we express no views on a proper resolution.
 As mentioned above, Ginn's proposal states that its actual provisional billing rates established by DCAA, which were reflected in Ginn's first revised proposal, will be charged during performance of the contract. Ginn Supp. Comments at 21; AR, Tab 17A, Final Ginn Discussions Responses, at 2.
 Ginn does acknowledge the Army's calculation of the significance of this issue as $4,097,427. Ginn Supp. Comments at 26.
 In its comments, Ginn argues that it has compared its workload on another contract to the same industry standards that were used to prepare [Ginn]'s initial proposal and found that its efficiency assumption was entirely appropriate. Ginn Supp. Comments at 12. As support, Ginn furnished a declaration from one of its senior executives, making essentially the same claim. In our view this broad claim, based on comparison to unspecified industry standards for workload, does not provide a reasonable basis for Ginn to assume, or the Army accept as reasonable, a 10 percent improvement in efficiency across the board.
 In our view, the Army has, in essence, allowed Ginn to hide five FTEs in the RFP's plug number for materials and supplies. When using normalized costs in a cost realism analysis, it is essential to segregate cost factors that are company-unique (here, the inclusion of labor and overhead in setting retail prices at a hardware store), from those that are expected to apply similarly to all offerors (here, the allowable cost of materials and supplies). E.g., Bendix Field Eng'g Corp., B-246236, Feb. 25, 1992, 92'1 CPD para. 227 at 17. If the Army intended that offerors could use the solicitation's plug number for both the cost of supplies and also the associated costs of the labor for this function (i.e., providing supply services), it should have amended the RFP to inform all offerors of this interpretation, in order to provide an equal basis for the competition.
 We note that the protester argues that the effect of the fringe benefit comparison error would have been $4,777,371.41. We need not resolve which calculation is correct because even under the Army's lower figure the protester was prejudiced.
 We note that the provision was extended to fiscal year 2010. See Continuing Appropriations Resolution, 2010, Pub. Law No. 111'68, Div. B, sect. 104, 123 Stat. ___, ___ (Oct. 1, 2009) (extending appropriations restrictions).
 For standard competitions, as here, the start date is the day that the agency makes a formal public announcement of the cost comparison study--in this case, September 27, 2006. The end date is the day that all standard competition form certifications are complete, signifying a performance decision. OMB Circular A-76, Attach. B, B.1, B.2. The Army completed the certifications on March 25, 2009. AR, Tab 25, Standard Competition Form, at 4-5. We conclude that the 30-month period ended approximately 2 days after the March 25 certification, or March 27, 2009.
 The Navy has argued in another protest of an A-76 competition that the Department of Defense has issued memoranda concluding that, notwithstanding the expiration of the time limit specified in the appropriations statutes, the agency may take actions to correct flaws in an A-76 competition, such as amending the solicitation and making a new award decision. Department of the Navy--Request for Modification of Remedy, B-401102.3, Aug. 6, 2009, 2009 CPD para. 162 at 3 n.1. Those arguments were not raised by the Army in the record here, and therefore we believe it would be inappropriate to speculate on their potential applicability. Although Ginn argues that we should find the ATO responsible for causing the competition to exceed the time limit, Ginn Supp. Comments at 27, the current language of the statute provides no exceptions to the time limit.