TITLE: B-298249.6, B-298249.7, B-298249.8, B-298249.9, B-298249.10, B-298249.11, B-298249.12, B-298249.13, B-298249.14, B-298249.15, B-298249.16, B-298249.17, B-298249.18, B-298249.19, B-298249.20, Multimax, Inc.; NCI Information Systems, Inc.; BAE Systems Information Technology LLC; Northrop Grumman Information Technology, Inc.; Pragmatics, Inc., October 24, 2006
BNUMBER: B-298249.6, B-298249.7, B-298249.8, B-298249.9, B-298249.10, B-298249.11, B-298249.12, B-298249.13, B-298249.14, B-298249.15, B-298249.16, B-298249.17, B-298249.18, B-298249.19, B-298249.20
DATE: October 24, 2006
******************************************************************************************************************************************
B-298249.6, B-298249.7, B-298249.8, B-298249.9, B-298249.10,B-298249.11, B-298249.12, B-298249.13, B-298249.14, B-298249.15, B-298249.16, B-298249.17, B-298249.18, B-298249.19, B-298249.20, Multimax, Inc.; NCI Information Systems, Inc.; BAE Systems Information Technology LLC; Northrop Grumman Information Technology, Inc.; Pragmatics, Inc., October 24, 2006

   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.

   Decision

   Matter of: Multimax, Inc.; NCI Information Systems, Inc.; BAE Systems
   Information Technology LLC; Northrop Grumman Information Technology, Inc.;
   Pragmatics, Inc.

   File: B-298249.6, B-298249.7, B-298249.8, B-298249.9, B-298249.10,
   B-298249.11, B-298249.12, B-298249.13, B-298249.14, B-298249.15,
   B-298249.16, B-298249.17, B-298249.18, B-298249.19, B-298249.20

   Date: October 24, 2006

   David S. Cohen, Esq., John J. O*Brien, Esq., Rowena Laxa, Esq., Laurel
   Hockey, Esq., and Catherine Kroll, Esq., Cohen Mohr, for Multimax, Inc.;
   Shelly L. Ewald, Esq., Timothy E. Heffernan, Esq., Louis B. Antonacci,
   Esq., Meghan M. DiPerna, Esq., and Justin M. Hargrove, Esq., Watt, Tieder,
   Hoffar & Fitzgerald, for NCI Information Systems, Inc; Drew A. Harker,
   Esq., Matthew H. Solomson, Esq., Chad E. Miller, Esq., and Patricia L.
   Stasco, Esq., Arnold & Porter, for BAE Systems Information Technology LLC;
   Anne B. Perry, Esq., John W. Chierichella, Esq., Jonathan S. Aronie, Esq.,
   Marko Kipa, Esq., Louis D. Victorino, Esq., and Keith R. Szeliga, Esq.,
   Sheppard Mullin, for Northrop Grumman Information Technology, Inc.; Karen
   R. Harbaugh, Esq., Robert E. Gregg, Esq., and Steven Tibbets, Esq.,
   Squire, Sanders & Dempsey, for Pragmatics, Inc., the protesters.

   Carl J. Peckinpaugh, Esq., and Helaine G. Elderkin, Esq., Computer
   Sciences Corporation; Richard J. Webber, Esq., Lisa K. Miller, Esq., and
   Craig S. King, Esq., Arent Fox, for CACI-ISS, Inc.; Richard O. Duvall,
   Esq., David S. Black, Esq., Eric L. Yeo, Esq., and Caitlin K. Cloonan,
   Esq., Holland & Knight, for Booz Allen Hamilton; Alexander J. Brittin,
   Esq., and Margaret A. Dillenburg, Esq., for Science Applications
   International Corporation; Gerald H. Werfel, Esq., Pompan, Murray &
   Werfel, for STG, Inc.; Richard J. Conway, Esq., David M. Adler, Esq., and
   Joseph R. Berger, Esq., Dickstein Shapiro, for Apptis, Inc.; and Grant H.
   Willis, Esq., and Peter F. Garvin, III, Esq., Jones Day, for Electronic
   Data Systems, the intervenors.

   Raymond M. Saunders, Esq., Karl M. Ellcessor, Esq., and Lt. Col. Frank A.
   March, Department of the Army, for the agency.

   David A. Ashen, Esq., and John M. Melody, Esq., Office of the General
   Counsel, GAO, participated in the preparation of the decision.

   DIGEST

   Where agency identified certain proposed hourly labor rates as
   significantly higher than independent government cost estimate (IGCE)
   labor rates, offerors reasonably deduced--incorrectly, as record
   shows--that rates not identified were not significantly higher than IGCE
   rates, which led offerors to leave those rates unchanged in their final
   proposal revisions; discussions therefore were misleading and protest is
   sustained on that basis.

   DECISION

   Multimax, Inc., NCI Information Systems, Inc., BAE Systems Information
   Technology LLC, Northrop Grumman Information Technology, Inc. (NGI), and
   Pragmatics, Inc. protest the Department of the Army*s award of 11
   contracts to other offerors, under request for proposals (RFP) No.
   W91QUZ-05-R-0004, under the agency*s Information Technology Enterprise
   Solutions-2 Services (ITES-2S) procurement for information technology (IT)
   services. The protesters assert that the agency failed to conduct
   meaningful discussions, and that its evaluation of proposals and resulting
   source selection were unreasonable.

   We sustain the protests.

   BACKGROUND

   The solicitation provided for award of multiple
   indefinite-delivery/indefinite-quantity contracts for a 3-year base
   period, with three 2-year options, to furnish IT services worldwide in
   support of Army enterprise infrastructure goals. Specifically, as set
   forth in the solicitation*s statement of objectives (SOO),

   ITES-2S contemplates services-based solutions under which contractors may
   be required to provide a full range of IT equipment. Therefore, end-to-end
   solutions to satisfy worldwide development, deployment, operation,
   maintenance, and sustainment requirements are included. Additionally
   included is support to analyze requirements, develop and implement
   recommended solutions, and operate and maintain legacy systems, and
   equipment. It is the intention of the Government to establish a scope that
   is broad, sufficiently flexible to satisfy requirements that may change
   over the period of performance, and fully comprehensive so as to embrace
   the full complement of services that relate to IT.

   SOO at C.2.0. The solicitation provided for work to be accomplished
   through the issuance of task orders, primarily on a fixed-price or
   time-and-materials basis, awarded generally on the basis of a competition
   among the ITES-2S contract holders. RFP sect. J, attach. 4, Task Order
   Procedures.

   Offerors were required to propose fully-loaded hourly labor rates for a
   minimum of 104 required labor categories at both a government site and a
   contractor site (for a total of 208 rates in the base year), which would
   be used by the contractors in competing for task orders (unless the
   contractors proposed lower rates). These rates, subject to an annual
   escalation rate proposed by each offeror, were applied to the annual
   estimated hourly requirements for each labor category, with the resulting
   totals combined with annual other direct costs (ODC) as specified in the
   solicitation and increased by a fixed markup proposed by each offeror for
   each ODC category, to yield an overall Total Proposed Contract Price
   (TPCP).

   Eight awards were contemplated (including up to four to small businesses),
   but the RFP also stated that *[t]he Government reserves the right to make
   no, one or multiple awards; the Government also reserves the right to make
   more than eight awards.* RFP sect. M.1. The awards were to be made to the
   offerors whose proposals were determined to be the *best value* on the
   basis of three evaluation factors: (1) mission support (with subfactors
   for performance-based approach, performance-based task approach, and small
   business participation); (2) performance risk (past performance, corporate
   experience, and financial; and (3) price. The non-price factors were
   significantly more important than price.

   Seventeen proposals were received. The Army entered into discussions with
   all 17 offerors, issuing written items for negotiation (IFN) and affording
   each offeror the opportunity to make an oral presentation. Subsequently,
   the agency determined that all proposals (one offeror withdrew) had no
   weaknesses or deficiencies, and that none of the TPCPs was unreasonably
   high, and requested final proposal revisions (FPR). Based on her
   evaluation of the FPRs, the source selection authority (SSA) selected
   11 proposals for award: Apptis, Inc., Booz Allen Hamilton (BAH), CACI-ISS,
   Inc., Computer Sciences Corporation (CSC), QSS, STG, Inc., EDS
   Corporation, General Dynamics, IBM, Inc., Lockheed Martin (LM), and SAIC,
   Inc.

   Upon learning of the awards, Multimax, NCI, BAE, NGI, and Pragmatics filed
   protests in our Office (B-298249.1, B-298249.2, B-298249.3, B-298249.4,
   B-298249.5). The Army subsequently advised us that its evaluation ratings
   for the financial subfactor under the performance risk evaluation factor
   failed to account for all information received during discussions, and
   that it thus would reevaluate the proposals and make new price/technical
   tradeoff decisions. We dismissed the protests as academic on May 12, 2006.

   The agency*s reevaluation led to the following results:

   +------------------------------------------------------------------------+
   |               |  Mission Support  |  Performance Risk  |     TPCP      |
   |               |                   |                    |               |
   |               |   (Performance,   | (Past Performance, |               |
   |               | Performance/Task, |     Corporate      |               |
   |               |  Small Business)  |    Experience,     |               |
   |               |                   |     Financial)     |               |
   |---------------+-------------------+--------------------+---------------|
   |      BAH      |       Good        |        Low         |$12,891,797,818|
   |               |                   |                    |               |
   |               |(Good, Outstanding,|(Low, Very Low, Low)|               |
   |               |       Good)       |                    |               |
   |---------------+-------------------+--------------------+---------------|
   |    Apptis     |       Good        |        Low         |13,048,326,171 |
   |               |                   |                    |               |
   |               |(Good, Good, Good) |(Very Low, Low, Low)|               |
   |---------------+-------------------+--------------------+---------------|
   |      QSS      |       Good        |        Low         |13,121,880,799 |
   |               |                   |                    |               |
   |               |   (Good, Good,    |(Moderate, Very Low,|               |
   |               |   Outstanding)    |        Low)        |               |
   |---------------+-------------------+--------------------+---------------|
   |   CACI-ISS    |       Good        |      Very Low      |13,391,706,671 |
   |               |                   |                    |               |
   |               |(Good, Good, Good) |(Very Low, Low, Low)|               |
   |---------------+-------------------+--------------------+---------------|
   |      IBM      |       Good        |        Low         |13,709,298,174 |
   |               |                   |                    |               |
   |               |   (Good, Good,    |(Low, Very Low, Low)|               |
   |               |   Outstanding)    |                    |               |
   |---------------+-------------------+--------------------+---------------|
   |      CSC      |       Good        |        Low         |13,788,431,819 |
   |               |                   |                    |               |
   |               |(Good, Outstanding,|  (Low, Low, Low)   |               |
   |               |   Outstanding)    |                    |               |
   |---------------+-------------------+--------------------+---------------|
   |      STG      |       Good        |      Very Low      |13,848,350,913 |
   |               |                   |                    |               |
   |               |   (Good, Good,    |(Very Low, Very Low,|               |
   |               |    Acceptable)    |     Very Low)      |               |
   |---------------+-------------------+--------------------+---------------|
   |  Pragmatics   |    [REDACTED]     |     [REDACTED]     |  [REDACTED]   |
   |---------------+-------------------+--------------------+---------------|
   |      EDS      |       Good        |        Low         |14,109,977,896 |
   |               |                   |                    |               |
   |               |   (Good, Good,    |(Moderate, Low, Low)|               |
   |               |   Outstanding)    |                    |               |
   |---------------+-------------------+--------------------+---------------|
   |Lockheed Martin|       Good        |      Very Low      |14,306,616,917 |
   |               |                   |                    |               |
   |               |(Good, Outstanding,|(Very Low, Very Low,|               |
   |               |    Acceptable)    |     Very Low)      |               |
   |---------------+-------------------+--------------------+---------------|
   |    General    |       Good        |        Low         |14,421,393,555 |
   |   Dynamics    |                   |                    |               |
   |               |(Good, Good, Good) |(Very Low, Low, Low)|               |
   |---------------+-------------------+--------------------+---------------|
   |     SAIC      |       Good        |        Low         |$14,554,346,258|
   |               |                   |                    |               |
   |               |(Good, Outstanding,|  (Low, Low, Low)   |               |
   |               |    Acceptable)    |                    |               |
   |---------------+-------------------+--------------------+---------------|
   |   Multimax    |    [REDACTED]     |     [REDACTED]     |  [REDACTED]   |
   |---------------+-------------------+--------------------+---------------|
   |      BAE      |    [REDACTED]     |     [REDACTED]     |  [REDACTED]   |
   |---------------+-------------------+--------------------+---------------|
   |      NGI      |    [REDACTED]     |     [REDACTED]     |  [REDACTED]   |
   |---------------+-------------------+--------------------+---------------|
   |      NCI      |    [REDACTED]     |     [REDACTED]     |  [REDACTED]   |
   +------------------------------------------------------------------------+

   Based on her review of the revised evaluation reports and a new series of
   77 head-to-head comparisons among the proposals, the SSA made awards to
   the 11 original awardees. Multimax, NCI, BAE, NGI, and Pragmatics
   thereupon filed these protests with our Office.

   CHANGE IN REQUIREMENTS

   Multimax, NCI, BAE and NGI assert that the Army improperly failed to amend
   the solicitation to reflect an increase in its requirements and permit
   offerors to submit revised proposals. The protesters maintain that the
   Army was required to amend the solicitation because it apparently decided
   to mandate, rather than merely authorize, use of ITES-2S for the
   acquisition of IT services. In this regard, in the source selection
   decision (SSD), the SSA explained her determination to award 11 rather
   than 8 contracts as based, in part, on the fact that *it has come to my
   attention that senior leadership within the Army is considering the
   possibility of mandating the use of ITES-2S for information technology
   services acquisitions. If that scenario were to arise, there is a
   possibility that more work might be performed under ITES-2S than was
   estimated when the Government developed its acquisition strategy.* SSD at
   28. According to the SSA, *[h]aving three additional contractors would
   provide the Army a larger base of contractors to handle this potential
   increase in workload.* Id. The protesters assert that, had they known of
   the loss of any future opportunity to market IT services to the Army
   outside of the ITES-2S contract vehicle, and of the additional work to be
   ordered under ITES-2S when it was made mandatory, they would have proposed
   lower prices.

   Where an agency*s requirements change materially after a solicitation has
   been issued, it must issue an amendment notifying offerors of the change
   and affording them an opportunity to respond. Federal Acquisition
   Regulation (FAR) sect. 15.206(a); Northrop Grumman Info. Tech., Inc. et
   al., B-295526 et al., Mar. 16, 2005, 2005 CPD para. 45 at 13; Symetrics
   Indus., Inc., B-274246.3 et al., Aug. 20, 1997, 97-2 CPD para. 59 at 6.
   This rule applies even after the submission of final proposal revisions,
   up until the

   time of award. Northrop Grumman Info. Tech., Inc. et al., supra; Digital
   Techs., Inc., B-291657.3, Nov. 18, 2004, 2004 CPD para. 235 at 3; NV
   Servs., B-284119.2, Feb. 25, 2000, 2000 CPD para. 64 at 17. Amending the
   solicitation provides offerors an opportunity to submit revised proposals
   on a common basis that reflects the agency*s actual needs. Dairy Maid
   Dairy, Inc., B-251758.3 et al., May 24, 1993, 93-1 CPD para. 404 at 7-9.

   We find that there was no material change in the Army*s requirements.
   While it is clear from the SSD that the SSA believed there was a
   possibility that use of the contract would be made mandatory, there is no
   evidence that this ever occurred. The Army states that *[t]here is not
   currently nor was there ever any proposal or initiative to make use of
   ITES-2S mandatory,* Agency Comments, Aug. 28, 2006, at 10, and the SSA
   asserts that she was never advised that use of the ITES-2S was to be made
   mandatory. Specifically, she explains, while she had understood from
   informal conversations prior to the initial awards in April 2006 that the
   Army*s Chief Information Officer (CIO) was *considering the possibility of
   directing Army personnel to use ITES-2S,* the Army CIO *did not consult
   with me directly on this matter . . . Nor were any written materials ever
   presented to me regarding this issue.* Declaration of SSA, Aug. 8, 2006,
   at 1.[1]

   The Director of the Army Small Computer Program (ASCP), which includes the
   ITES-2S procurement, also states that she was never advised that use of
   the ITES-2S was to be mandatory. She has submitted a declaration in which
   she states as follows:

   I am unaware of any Army policy or initiative that mandates the use of
   ITES-2S for Army IT services procurements. To my knowledge, the Army is
   not currently planning any proposed policy or initiative that would
   require ITES-2S to be used for IT services procurements. Nor, to my
   knowledge, has such a policy previously been under development. It is
   highly unlikely, in my view, that such a policy will ever be implemented,
   as there are serious questions about its desirability and feasibility.

   Declaration of Director of ASCP, Aug. 28, 2006.

   Finally, the Deputy Principal Director for Governance, Acquisition and
   Chief Knowledge Office, Office of the Army CIO, who is responsible for
   initiating and developing IT contracting policy for the Army, has
   submitted a declaration in which he states that he is *unaware of any
   proposal or initiative to make use of the ITES-2S contract
   mandatory--either . . . as of July 12, 2006, or at any time in the past.
   In fact, if such a proposal or initiative existed, my office would either
   have the lead or be actively involved.* Declaration of Deputy Principal
   Director, Aug. 28, 2006.

   The protesters note that current Army policy establishes such a strong
   preference for the use of ASCP contract vehicles such as ITES-2S, that its
   use is tantamount to mandatory. In this regard, for example, a draft
   version of the agency*s *Department of the Army IT Purchasing Guide: How
   to Procure Commercial Information Technology Hardware, Software and
   Services,* dated Summer 2006, provides that *the ASCP office is the
   primary source for hardware, software and services,* and that

   [i]f your purchase is other IT products [--i.e., other than software,
   desktops or notebooks--] or services and the cost is greater than $25K, or
   if the required item or service is available on an ASCP contract, but you
   have justification, such as better pricing, for purchasing the item or
   service from another source, you must request a waiver from ASCP. See
   [Department of the Army Pamphlet] 25-1-1, para 11-2.

   Army IT Purchasing Guide at 5-6.

   As noted by the Army, however, the draft Army IT Purchasing Guide reflects
   the same policy as it existed prior to the closing date for receipt of
   FPRs (March 27, 2006). Thus, for example, the November 2005 version of
   Army Pamphlet 25-1-1, referenced above in the draft Army IT Purchasing
   Guide, provides as follows:

   Army customers must look to meet their requirements using ASCP contracts
   before making commercial IT purchases from other sources.

                                   . . . . .

   If there is an existing ASCP indefinite delivery indefinite quantity
   contract or blanket purchase agreement (BPA) available, the system or
   service is obtained from that contract to the maximum extent practical.

                                   . . . . .

   The ASCP is the main source for commercial IT purchases greater than $25K.
   ASCP has an array of fully competed contract vehicles to meet most Army
   requirements. These contract vehicles must be considered before buying
   from contract vehicles from other sources. If an Army customer chooses
   other than an ASCP contract vehicle, ASCP must first grant a waiver.

   Army Pamphlet 25-1-1, sections 11-1, 11-2.

   We conclude that there is no evidence that, at the time of the awards,
   there was any likelihood of a material change in existing Army policy
   favoring use of ITES-2S. It follows that there was no significant change
   in the Army*s requirements that would necessitate amending the ITES-2S
   solicitation and affording offerors an opportunity to propose to the new
   requirements.

   PRICE

   The protesters assert that the Army applied an unreasonable, mechanistic
   formula in evaluating proposed labor rates. In addition, they assert that
   this resulted in a failure to conduct meaningful discussions. We agree.

   Price Evaluation Methodology

   Multimax, BAE and Pragmatics assert that the Army*s evaluation of proposed
   labor rates was unreasonable. In this regard, the Army reports that it
   employed a two-step approach to evaluating labor rates for purposes of
   determining price reasonableness, detecting unbalanced pricing, and
   identifying labor rates to question during discussions: first, it compared
   an offeror*s rate for a labor category to the IGCE rate for that category,
   and then it compared the rate to the mean of all offerors* evaluated rates
   for each labor category using a two-standard-deviation measure. The
   agency*s price evaluator explained the second step as follows:

   Next, the Price evaluation team calculated the mean of all offerors*
   evaluated labor rates for each labor category. The mean evaluated labor
   rates were then used to calculate the standard deviation from the mean. In
   order to determine the most appropriate measure of comparison, the
   following were calculated: mean plus and minus one standard deviation,
   mean plus and minus two standard deviations, and mean plus and minus three
   standard deviations. A comparison was made, using the three separate
   standard deviations, to determine which offerors* average labor rates for
   each labor category fall outside the range of each standard deviation. The
   majority of the offerors* average labor rates fell outside the range of
   one standard deviation and no offeror*s average labor rates fell outside
   the range of three standard deviations. Therefore, it was determined that
   two standard deviations was the most appropriate measure of comparison to
   use for the reasonableness assessment.

   Memorandum of Agency Price Evaluator to Source Selection Evaluation Board
   (SSEB) Chairperson, Nov. 29, 2005, at 2-3; Agency Comments, Sept. 21,
   2006, at 3.[2]

   Under this two-step approach, the agency would issue an IFN to an offeror
   questioning a proposed labor rate as significantly overstated (or
   understated) only if the rate both exceeded (or was lower than) the IGCE
   rate, and was more than two standard deviations greater (or less) than the
   mean rate of all offerors for that category. According to the contracting
   officer (who was responsible for conducting discussions and determining
   overall price reasonableness), the two-step evaluation was used to
   identify *extraordinary outlier rates,* that is, *rates that were
   significantly overstated or understated and which might pose a risk to the
   Government of paying an unreasonable amount during performance. . . .
   Rates that did not meet [both] tests were not considered outliers and were
   not questioned.* Second Declaration of Contracting Officer at 1; see
   Agency Comments, Sept. 8, 2006, 3-8, 16; Agency Comments, Sept. 21, 2006,
   at 3; Declaration of Agency Price Evaluator, Aug. 18, 2006, at 1.[3]

   The two-standard-deviation formula resulted in an extremely wide range of
   acceptable rates for the labor categories.[4] For example, looking just at
   the initial labor rates for government site labor, some of the more
   extreme ranges were as follows:

   +------------------------------------------------------------------------+
   |                           | IGCE Rate |        Acceptable Range        |
   |                           |           |                                |
   |                           |           |Mean-plus-two-standard-deviation|
   |                           |           |              Test              |
   |---------------------------+-----------+--------------------------------|
   |Program Manager-Senior     |$[REDACTED]|          $[REDACTED]           |
   |---------------------------+-----------+--------------------------------|
   |Project Administrator      |[REDACTED] |           [REDACTED]           |
   |---------------------------+-----------+--------------------------------|
   |Applications Systems       |[REDACTED] |           [REDACTED]           |
   |Analyst-Senior             |           |                                |
   |---------------------------+-----------+--------------------------------|
   |Systems Engineer-Senior    |[REDACTED] |           [REDACTED]           |
   |---------------------------+-----------+--------------------------------|
   |Software Engineer-Senior   |[REDACTED] |           [REDACTED]           |
   |---------------------------+-----------+--------------------------------|
   |Software                   |[REDACTED] |           [REDACTED]           |
   |Engineer-Intermediate      |           |                                |
   |---------------------------+-----------+--------------------------------|
   |Software Engineer-Associate|[REDACTED] |           [REDACTED]           |
   |---------------------------+-----------+--------------------------------|
   |IT Certified               |[REDACTED] |           [REDACTED]           |
   |Professional-Senior        |           |                                |
   |---------------------------+-----------+--------------------------------|
   |Disaster Recovery          |[REDACTED] |           [REDACTED]           |
   |Contingency Administrator  |           |                                |
   |---------------------------+-----------+--------------------------------|
   |Information Security       |[REDACTED] |           [REDACTED]           |
   |Specialist-Senior          |           |                                |
   |---------------------------+-----------+--------------------------------|
   |Information Security       |[REDACTED] |           [REDACTED]           |
   |Specialist-Intermediate    |           |                                |
   |---------------------------+-----------+--------------------------------|
   |Information Security       |[REDACTED] |           [REDACTED]           |
   |Specialist-Associate       |           |                                |
   |---------------------------+-----------+--------------------------------|
   |Systems                    |[REDACTED] |           [REDACTED]           |
   |Administrator-Senior       |           |                                |
   +------------------------------------------------------------------------+

   SSA Initial Briefing Materials, Nov. 28, 2005, app. A; SSA Final Briefing
   Materials, June 14, 2006, app. B.[5]

   As illustrated by the chart, the upper end of the range was significantly
   above the IGCE for some of the labor categories, and in some instances was
   nearly, or more than, twice the IGCE (such as $[REDACTED] versus the
   $[REDACTED] IGCE rate for Application System Analyst-Senior, $[REDACTED]
   versus the $[REDACTED] IGCE rate for Software Engineer-Senior, and
   $[REDACTED] versus the $[REDACTED] IGCE rate for Information Security
   Specialist-Senior). Likewise, the lower end, in some instances, was below
   the federal minimum wage or was even a negative number (such as
   $[REDACTED] for Project Administrator, $[REDACTED] for Information
   Security Specialist-Senior, and $[REDACTED] for Information Security
   Specialist-Associate). There is no indication that the agency ever
   reviewed the results of the formula to assure that the prices at the
   extreme end of the ranges reflected reasonable pricing; rather, the agency
   mechanistically applied the formula and accepted the results without
   further analysis. We conclude that the agency*s methodology did not
   provide a valid means for identifying *outlier* (questionable) rates, and
   this aspect of the evaluation therefore was unreasonable. See generally
   Metro Mach. Corp., B-297879.2, May 3, 2006, 2006 CPD para. 80 at 9-10
   (mechanical application of an agency*s own estimates for labor hours or
   costs to determine evaluated costs, without the exercise of informed
   judgment by the contracting agency in independently analyzing the
   offeror*s proposed costs based upon its particular approach and
   circumstances, was unreasonable); The Jonathan Corp.; Metro Machine Corp.,
   B-251698.3, B-251698.4, May 17, 1993, 93-2 CPD para. 174 at 11-13; United
   Int*l Eng*g, Inc. et al., B-245448.3 et al., Jan. 29, 1992, 92-1 CPD para.
   122 at 11. We therefore sustain the protests of Multimax, BAE and
   Pragmatics on the basis that the Army failed to reasonably evaluate
   proposed labor rates.

   Misleading Price Discussions

   The price evaluation also is problematic because the results of the
   agency*s analysis were used to determine which prices to bring to
   offerors* attention during discussions as possibly being unreasonably
   high, and thus potentially determined which prices offerors would adjust
   in their FPRs. In this regard, NGI and Multimax assert that the price IFNs
   they received during discussions failed to reasonably advise them of
   numerous labor rates that should have been brought to their attention as
   significantly overstated and, moreover, that they were misled into
   believing that only the few rates identified in the IFNs significantly
   exceeded the IGCE rates. The protesters maintain that, had they known that
   numerous of their proposed rates not identified during discussions were
   significantly higher than the corresponding IGCE rates, they would have
   reduced those rates in their FPRs, just as they reduced the rates that
   were identified in the IFNs as overstated.

   It is a fundamental precept of negotiated procurements that discussions,
   when conducted, must be meaningful; that is, discussions must identify
   deficiencies and significant weaknesses in each offeror*s proposal that
   could reasonably be addressed so as to materially enhance the offeror*s
   potential for receiving award. PAI Corp., B-298349, Aug. 18, 2006, 2006
   CPD para. 124 at 8; Spherix, Inc., B-294572, B-294572.2, Dec. 1, 2004,
   2005 CPD para. 3 at 13. An agency fails to conduct meaningful discussions
   where it fails to apprise an offeror that its prices were viewed as
   unreasonably high. Price Waterhouse, B-220049, Jan. 16, 1986, 86-1 CPD
   para. 54 at 6-7. Further, an agency may not mislead an offeror--through
   the framing of a discussion question or a response to a question--into
   responding in a manner that does not address the agency*s concerns;
   misinform the offeror concerning a problem with its proposal; or misinform
   the offeror about the government*s requirements. Metro Mach. Corp.,
   B-281872 et al., Apr. 22, 1999, 99-1 CPD para. 101 at 6. In conducting
   exchanges with offerors, agency personnel also may not *engage in conduct
   that . . . favors one offeror over another,* FAR sect. 15.306(e)(1); in
   particular, agencies may not engage in what amounts to disparate treatment
   of the competing offerors. Front Line Apparel Group, B-295989, June 1,
   2005, 2005 CPD para. 116 at 3-4.

   Here, in the discussions questions issued to NGI and Multimax (as well as
   to other offerors such as BAE) questioning the proposed rates for
   particular labor categories as significantly overstated, the agency
   advised as follows: *Your proposed labor rates are significantly higher
   than the Independent Government Cost Estimate (IGCE) rates for the
   following labor categories . . . . The offeror should consider revising
   the price proposal. If you do not revise the identified rates, please
   provide an explanation for the basis of the rate.* See IFN to NGI no. 51,
   Nov. 3, 2005; IFN to NGI no. 298, Jan. 6, 2006; IFN to BAE no. 50, Nov. 3,
   2005; IFN to BAE no. 297, Jan. 6, 2006; IFN to Multimax no. 49, Nov. 11,
   2006. Thus, there was no reference in the IFNs to the agency*s reliance on
   the two-standard-deviation calculation, but instead only to the proposed
   labor rate being *significantly higher than* the IGCE as the basis for the
   IFN.

   The Army*s price discussions with NGI and Multimax (as well as with BAE)
   were inadequate because, due to the agency*s reliance on the
   two-standard-deviation formula to identify *outlier* rates--and the broad
   range of acceptable prices resulting from the formula--it failed to bring
   to the protesters* attention numerous rates that reasonably should have
   been considered significantly overstated. In this regard, the record shows
   that proposed rates that were not questioned in the IFNs could actually
   exceed the IGCE rates by a greater percentage than the rates that were
   identified. Thus, for example, although NGI*s rate for [REDACTED] at the
   contractor site was [REDACTED] percent higher than the IGCE rate, this
   rate was not identified in an IFN because it was within the wide range of
   acceptable prices established under the formula. At the same time,
   although NGI*s proposed rate for [REDACTED] at the contractor site was
   only [REDACTED] percent higher than the IGCE rate, because it fell outside
   the range established by the two-standard-deviation test, NGI was advised
   that its rate was *significantly higher* than the IGCE. There simply is no
   reasonable basis for bringing the former rate to the offeror*s attention,
   but not the latter.

   The above example is not an isolated one. NGI calculates that
   [REDACTED] of its proposed labor rates that were not identified during
   price discussions were similar to this example--they exceeded the
   corresponding IGCE rate by a higher percentage than one or more of the
   rates identified in its price IFNs. NGI notes further that [REDACTED] of
   its unquestioned rates exceeded the IGCE rates by a greater percentage
   than did some rates that were questioned in other offerors* IFNs.
   Likewise, the record indicates that Multimax was not advised that its
   proposed rates for a significant number of labor categories were higher
   than the corresponding IGCE rates, despite the fact that these proposed
   rates deviated from the IGCE by a greater percentage than rates that were
   identified in discussions with Multimax or other offerors. Multimax
   calculates that [REDACTED] of its unquestioned rates (only [REDACTED] of
   its rates were identified in IFNs) exceeded the IGCE rates by a greater
   percentage than the rates that were questioned in other offerors* IFNs
   (during initial price discussions).

   We conclude that not only were offerors not adequately advised of all of
   their significantly overstated rates, but the agency*s failure to identify
   the additional rates actually misled the offerors into believing that
   those rates did not require further adjustment.[6] In these circumstances,
   we conclude that the agency failed to conduct meaningful discussions with
   the protesters.

   The Army questions whether the protesters have established that they
   suffered competitive prejudice as a result of the agency*s approach to
   discussions. A reasonable possibility of prejudice is a sufficient basis
   for sustaining a protest. McDonald-Bradley, B-270126, Feb. 8, 1996, 96-1
   CPD para. 54 at 3; see Statistica, Inc. v. Christopher, 102 F.3d 1577,
   1581 (Fed. Cir. 1996); Creative Info. Tech., Inc., supra; The Jonathan
   Corp.; Metro Mach. Corp., supra, at 10. The record includes the
   protesters* calculation of the price reductions they would have proposed
   had the agency identified the additional labor rates during discussions;
   these calculations show total price reductions of $[REDACTED] for NGI and
   $[REDACTED] for Multimax. Based on these reductions, the protesters*
   prices would have been lower than one or more awardees with equal or lower
   technical ratings. Since the protesters* calculations are consistent with
   their responses to the price IFNs they received--i.e., they lowered their
   rates in response to the price IFNs--and their lower prices may have
   affected the award determination, we find that there is a reasonable
   possibility that the protesters were prejudiced by the agency*s actions.

   The Army argues that, had it advised the protesters of additional labor
   rates that significantly exceeded the IGCE rates, it also would have
   provided similar information to the awardees, thus affording them a
   similar opportunity to reduce their rates. The Army suggests that the
   result would have been a *wash.* However, the Army*s speculation as to
   what would have occurred if all offerors had been provided the same
   opportunity to revise their proposals is insufficient to rebut the
   apparent prejudice to the protesters from the misleading discussions. See
   Creative Info. Tech., Inc., supra; The Jonathan Corp.; Metro Mach. Corp.,
   supra, at 10. Accordingly, we sustain NGI*s and Multimax*s protests on the
   ground that the agency provided inadequate, misleading discussions.

   Unequal Discussions

   NCI and BAE assert that they were not afforded meaningful discussions with
   respect to their proposed labor rates, and instead were denied the
   opportunity provided other offerors to enhance their proposals* potential
   for award by addressing labor rates that significantly exceeded the IGCE
   rates. We agree.

   During discussions, the Army identified as significantly higher than the
   corresponding IGCE rates proposed rates that were as little as 3.67
   percent (during the initial price discussions) or 3.27 percent (during the
   second round of price discussions) higher than the IGCE rates. Meanwhile,
   as alluded to above, the agency did not identify other labor rates that
   were very much higher than the IGCE rates.[7] For example, [REDACTED] of
   NCI*s initially proposed labor rates was identified as significantly
   higher than the IGCE rates, despite the fact that NCI proposed rates that
   exceeded the IGCE by as much as [REDACTED] percent. Likewise, only
   [REDACTED] of BAE*s proposed rates were identified as significantly higher
   than the corresponding IGCE rates, despite the fact that other of BAE*s
   proposed rates exceeded the IGCE by as much as [REDACTED] percent. In
   contrast, one of the awardees had 69 of its proposed rates identified as
   significantly higher than the IGCE rates, another had 27 rates identified,
   and another 20 rates.

   Again, this example is not an isolated one. NCI calculates that
   [REDACTED], and BAE [REDACTED], of their unquestioned labor rates exceeded
   the IGCE rate by a higher percentage than rates brought to the attention
   of other offerors. NCI has calculated that it would have lowered its labor
   rates in a manner that would have reduced its TPCP by $[REDACTED] had it
   been advised of these additional overstated rates; BAE calculates that it
   would have reduced its price by $[REDACTED]. These price reductions--which
   have not been brought into question by the agency--would have resulted in
   the protesters* prices being lower than those of one or more awardees with
   equal or lower technical ratings. NCI Comments, Sept. 13, 2006, ex. 1; BAE
   Comments, Sept. 8, 2006, at 7, chart 3. Accordingly, we sustain the
   protests of NCI and BAE on the ground that the agency failed to provide
   them meaningful discussions.

   RECOMMENDATION

   We recommend that the Army reopen discussions with offerors, consistent
   with our conclusions above, and then request revised proposals. The agency
   may then determine the number of awards it deems appropriate. If the
   evaluation of revised proposals results in a determination that one or
   more of the current awardees* proposals no longer represent the best value
   to the government, the agency should terminate such contracts. We also
   recommended that Multimax, NCI, BAE, NGI, and Pragmatics be reimbursed
   their costs of filing and pursuing their protests, including reasonable
   attorneys* fees, with regard to the price evaluation and discussions
   issues. 4 C.F.R. sect. 21.8(2)(1) (2006). In accordance with 4 C.F.R.
   sect. 21.8(f)(1), the protesters* certified claims for such costs,
   detailing the time expended and costs

   incurred, must be submitted directly to the agency within 60 days after
   receipt of this decision.

   The protests are sustained.

   Gary L. Kepplinger

   General Counsel

   ------------------------

   [1] The SSA has further stated that there was no intention to raise the
   $20 billion contract ceiling set forth in the solicitation. Declaration of
   SSA, Aug. 7, 2006, at 1.

   [2] According to the Department of Defense*s (DOD) Contract Pricing
   Reference Guides, the standard deviation is a measure of dispersion of the
   samples or observations using the square root of the variance (with the
   variance of a sample being the average of the squared deviations between
   each observation and the mean). In a distribution that is approximately
   normal, plus or minus one deviation will include approximately 68 percent
   of the total observations in the sample; plus or minus two deviations will
   include approximately 95 percent of the total observations; and plus or
   minus three deviations will include approximately 99.7 percent of the
   total observations. However, of particular importance here, because all
   values are squared, a single observation that is far away from the mean
   can substantially affect both the variance and the standard deviation. DOD
   Contract Pricing Reference Guides, vol. 2, ch. 3.3.

   [3] The record indicates that, in a limited number of instances during the
   first of the two rounds of price discussions, the agency questioned an
   offeror*s rate for a particular labor category as being overstated even
   though it did not exceed the IGCE rate for that category. Specifically,
   during the initial round of price discussions, the Army questioned 123 of
   the offerors* over 3,500 labor rates as being significantly higher than
   the IGCE rates for the categories questioned (as well as 28 labor rates
   for being significantly understated); in 9 of these cases, however, the
   rates were questioned as being significantly in excess of the IGCE even
   though they only exceeded the range of acceptable rates established by the
   mean-plus-two-standard-deviation test, and in fact were not higher than
   the IGCE rates for those categories. Agency Comments, Sept. 8, 2006, at
   10; Agency Comments, Sept. 21, 2006, at 3. (During the second round of
   price discussions, the Army questioned 69 labor rates as being
   significantly higher than the IGCE and 25 as being significantly
   understated. Agency Comments, Sept. 8, 2006, at 10.)

   [4] The extremely wide range of acceptable labor rates appears to have
   resulted from the fact that, as noted above, rates that were far from the
   mean disproportionately increased the range described by the
   two-standard-deviation calculation.

   [5] Application of the formula also resulted in a wide range of acceptable
   hourly rates for some labor categories at contractor sites, such as, for
   example, $[REDACTED] to $[REDACTED] for Program Manager-Intermediate,
   where the IGCE was $[REDACTED]. SSA Initial Briefing Materials, Nov. 28,
   2005, app. A; SSA Final Briefing Materials, June 14, 2006, app. B.

   [6] We note that, with respect to NGI, the agency confirmed during
   discussions that only the identified labor rates were overstated.
   Specifically, during NGI*s oral presentation, the contracting officer
   explained the IFN to NGI as follows: *We*re basically identifying the
   proposed labor rates that are significantly higher than the IGCE rates.*
   Videotape of NGI Oral Presentation/Discussions at 1:15-1:16. The
   contracting officer confirmed NGI*s interpretation of his explanation in
   this regard when answering the following questions from NGI:

   NGI : . . . I was just curious what significantly meant in terms of the
   delta between the cost estimate and what you saw here.

   Contracting Officer: Yeah . . . we generally . . . I mean . . . we can*t
   really give you that.

   NGI: Right, but I guess my question is, is it just these [REDACTED] rates?

   Contracting Officer: It*s just the rates we*ve identified.

   Id. at 1:17-1:18.

   [7] The reference above to labor rates identified during discussions as
   being overstated but exceeding the IGCE rates by as little as 3.27 and
   3.67 percentage does not account for the proposed labor rates identified
   during discussions as being overstated but which in fact were less than
   the IGCE rates. The Army maintains that these rates were erroneously
   identified as being overstated. Agency Comments, Sept. 8, 2006, at 10;
   Agency Comments, Sept. 21, 2006, at 3.