General Dynamics Mission Systems, Inc.

B-416181: Jul 2, 2018

Additional Materials:

Contact:

Ralph O. White
(202) 512-8278
WhiteRO@gao.gov

Kenneth E. Patton
(202) 512-8205
PattonK@gao.gov

 

Office of Public Affairs
(202) 512-4800
youngc1@gao.gov

General Dynamics Mission Systems, Inc. (GDMS), of Fairfax, Virginia, protests the rejection of its proposal under request for proposals (RFP) No. W56KGY-17-R-0026, issued by the Department of the Army, Army Contracting Command--Aberdeen Proving Ground, for a computer hardware/software solution for the Army's Distributed Common Ground System (DCGS-A). GDMS challenges the evaluation of its product demonstration and technical proposal.

We deny the protest.

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:  General Dynamics Mission Systems, Inc.

File:  B-416181

Date:  July 2, 2018

Carla J. Weiss, Esq., D. Joe Smith, Esq., Grant Schweikert, Esq., and Joshua A. Violanti, Esq., Jenner & Block LLP, for the protester.
Jeffery M. Chiow, Esq., and Mark J. Linderman, Esq., Rogers Joseph O'Donnell, PC, for Raytheon Co. Intelligence and Information Systems; Anuj Vohra, Esq., and Hart W. Wood, Esq., Crowell & Moring LLP, for Palantir Technologies, Inc., the intervenors.
Debra J. Talley, Esq., and David A. Balaban, Esq., Department of the Army, for the agency.
Pedro E. Briones, Esq., Noah B. Bleicher, Esq., and Peter H. Tran, Esq., Office of the General Counsel, GAO, participated in the preparation of the decision.

DIGEST

Protest of the agency's technical evaluation is denied where the record shows that the evaluation was reasonable and consistent with the solicitation's evaluation criteria.

DECISION

General Dynamics Mission Systems, Inc. (GDMS), of Fairfax, Virginia, protests the rejection of its proposal under request for proposals (RFP) No. W56KGY-17-R-0026, issued by the Department of the Army, Army Contracting Command--Aberdeen Proving Ground, for a computer hardware/software solution for the Army's Distributed Common Ground System (DCGS-A).[1]  GDMS challenges the evaluation of its product demonstration and technical proposal.

We deny the protest.

BACKGROUND

Distributed Common Ground System-Army

The solicitation describes DCGS-A as a "system of systems" comprised of:  an extensive intelligence architecture that enables distributed processing and information sharing of data from sensors and other sources across the globe; software tools to help users analyze, update, and share intelligence products; servers that store data and intelligence; and a set of common standards that enable integration of new technology and processes.  See Agency Report (AR), Tab 10, RFP attach 3, Performance Requirements Document (PRD), at 6; see, e.g., Palantir USG, Inc., B-412746, May 18, 2016, 2016 CPD ¶ 138 at 2.[2]  The Test and Quality Division Chief for DCGS-A (hereinafter, division chief) explains that DCGS-A is primarily a software solution, with hardware components and over 100 software applications that aid the user in processing intelligence from something as simple as a weather feed, to more complex data such as full-motion video from drones.  Hearing Transcript (Tr.) at 19:23-20:16.[3]

According to the statement of work (SOW), the instant procurement, DCGS-A Capability Drop 1 (CD-1), seeks a combined, commercial hardware/software solution to enhance DCSG-A capabilities to meet the interoperability, security, training, usability, and data management capabilities at the battalion level.[4]  SOW at 4.  CD-1, known as the "battalion solution," also requires related support services and various deliverables, including product documentation (such as user manuals), software licensing agreements, monthly security patches/updates, and information assurance and cybersecurity controls, among other things.  Id. at 4, 11, 18-20.  The RFP states that the goal is to establish commercial item solutions that will satisfy the requirements specified in the solicitation's performance requirements and other documents referenced in the SOW.  Id. at 7.  The requirements are listed in a detailed performance requirement document (PRD) that specifies approximately 70 automated or computer-assisted capabilities, including various data management capabilities.  PRD at 6-11.

RFP Evaluation & Award Provisions

The solicitation was issued on September 29, 2017, pursuant to Federal Acquisition Regulation parts 12 and 15, and provided for the award of multiple, fixed-price, indefinite-delivery/indefinite-quantity (IDIQ) contracts for a 5-year base period and one 5-year option.  RFP at 2, 69; AR, Tab 2, Contracting Officer's Statement & Memorandum of Law (COS/MOL), at 10.  The RFP stated that award would be based on a best-value tradeoff among three evaluation factors:  technical, price, and past performance.  RFP at 69.  Offerors were advised that the technical factor was significantly more important than price, and that past performance would be evaluated on an acceptable/unacceptable basis.  Id.  Offerors were to provide separate proposal volumes corresponding to each evaluation factor and perform a live product demonstration at Aberdeen Proving Ground, Maryland.  Id. at 59, 64.  The technical evaluation factor included three subfactors, two of which are at issue here--technical product demonstration and technical solution narrative (hereinafter, demonstration and narrative subfactors, respectively).[5]  Id. at 69-72.

With respect to the demonstration subfactor, offerors were to demonstrate whether their proposed solution could perform 32 tasks listed in the RFP's product demonstration plan (PDP).  Id. at 64.  Each task corresponded to a performance requirement.[6]  Compare AR, Tab 15, RFP attach. 8, PDP, at 5, with PRD at 10.  The PDP included general guidelines for conducting the demonstration and a sample walkthrough (i.e., a narrative workflow or flowchart), but offerors were advised that the actual demonstration would use different scenarios and data, and that the PDP was provided for evaluation purposes only.  PDP at 6; app. C, Sample Demo., at 9; RFP at 70.

The RFP stated that the Army would evaluate the extent to which an offeror's solution was able to accomplish the PDP tasks, and listed factors that could be considered in this respect, including, "but . . . not limited to:  task complexity (e.g., number of steps); system speed; system ease of use; and system issues (i.e., error messages) . . . ."  RFP at 70; PDP at 3.  The RFP advised offerors that their solutions were not required to successfully complete all PDP tasks, but that solutions that successfully completed many or all of the tasks would be rated more favorably than solutions that completed fewer tasks.  RFP at 70.  The RFP provided in this regard that proposals would be assessed an adjectival rating of outstanding, good, acceptable, marginal, or unacceptable under the demonstration subfactor.[7]  Id. at 70-71.

With respect to the narrative subfactor, offerors were to provide a technical narrative, including a proposed implementation plan and a detailed description of hardware requirements and specifications, and complete a cross reference matrix (CRM) provided with the solicitation.  Id. at 61, 66.  The CRM table listed all of the approximately 70 performance requirements, not just the 32 requirements (demonstration tasks) described above.  AR, Tab 12, RFP attach. 5, CRM, at 1-12.  The offeror was to complete two columns in the table for each performance requirement.  CRM at 1-12.  In one column, the offeror was to "clearly indicate" whether the proposed solution meets or does not meet the requirement, and if it met the requirement, briefly describe how it did so.  See id.; RFP at 66.  In the other column, the offeror was to provide references, with hyperlinks, to the corresponding documentation, such as a user guide, specifications, or similar product document.[8]  Id.  Offerors were also to provide copies of these reference documents as appendices to their proposal.  RFP at 66.  Offerors were cautioned that they were to provide sufficient details, in a concise manner, to permit a complete and accurate evaluation.  Id. at 61.

The RFP stated that the Army would evaluate the technical narrative, including the CRM, to assess whether an offeror's proposed solution would meet all performance requirements within 22 weeks after award of the first delivery order issued under the IDIQ contract.[9]  See id. at 71.  The RFP provided that technical proposals could be assessed significant weaknesses or deficiencies under this subfactor, and would be evaluated on an acceptable/unacceptable basis.  Id.

The Army received proposals from eight offerors by the November 6 deadline, including from GDMS, Raytheon, and Palantir.  COS/MOL at 10.  GDMS proposed its GeoSuite software and demonstrated its proposed solution on December 8.  Id.  The relevant proposals were evaluated as follows:

  GDMS Raytheon Palantir
Technical (overall) Unacceptable Good Acceptable
Demonstration Unacceptable Good Acceptable
Narrative Unacceptable Acceptable Acceptable
Usability Maturity Acceptable Acceptable Acceptable
Past Performance Acceptable Acceptable Acceptable
Price $202,166,580 $305,419,199 $349,287,162

See AR, Tab 45, GDMS Debriefing, at 1.  Product demonstrations and technical narratives were evaluated by separate evaluation teams, which documented their respective findings in separate evaluation reports.  COS/MOL at 6; see AR, Tab 35, GDMS Demo. Eval. Rep., at 1-14; Tab 41, GDMS Tech. Narrative Eval. Rep., at 1-3.

GDMS's unacceptable rating under the demonstration subfactor reflected the evaluators' assessment of a weakness for each of eight demonstration tasks and a significant weakness for each of six other tasks (including three data management tasks/performance requirements), as well as three overall deficiencies.  AR, Tab 35, GDMS Demo. Eval. Rep., at 3-12.  GDMS's unacceptable rating under the narrative subfactor reflected the Army's assessment of two deficiencies with respect to two of the same data management requirements.  AR, Tab 41, GDMS Tech. Narrative Eval. Rep., at 1-2. 

The chief engineer for DCGS-A CD-1 reviewed the technical evaluation reports and concurred with the evaluators' findings.  AR, Tab 42, Tech. Eval. Rep., at 1-2.  He assigned an overall rating of unacceptable to GDMS's technical proposal, because it was found unacceptable under two technical evaluation subfactors, did not meet RFP requirements, contained deficiencies, and presented an unacceptable risk of unsuccessful contract performance.  Id. at 2.

The source selection authority (SSA) for this procurement reviewed the evaluation results for all offerors, performed an integrated assessment of proposals, and conducted a best-value determination.  See AR, Tab 43, Source Selection Decision Document, at 98-99.  The SSA concluded that Raytheon and Palantir presented viable solutions that were "ready for award" and determined that their proposals represented the best value to the government.  Id. at 99.  The Army awarded contracts to Raytheon and Palantir on March 8, 2018, and GDMS filed this protest following receipt of a written debriefing.  COS/MOL at 23.

DISCUSSION

GDMS protests every weakness, significant weakness, and deficiency assessed against its technical proposal, arguing that the assessments were inconsistent with the performance requirements as stated in the solicitation.  The Army generally asserts that GDMS simply did not meet the requirements and that its arguments to the contrary largely reflect the protester's untimely disagreement with the product demonstration plan.

Although we do not address each of GDMS's arguments, we have considered all of the protester's contentions and find that none provide a basis to sustain the protest.[10]  In reviewing protests of an agency's evaluation, our Office does not reevaluate proposals, rather, we review the record to determine if the evaluation was reasonable, consistent with the solicitation's evaluation scheme, as well as procurement statutes and regulations, and adequately documented.  See Wackenhut Servs., Inc., B-400240, B-400240.2, Sept. 10, 2008, 2008 CPD ¶ 184 at 6; Cherry Road Techs.; Elec. Data Sys. Corp., B-296915 et al., Oct. 24, 2005, 2005 CPD ¶ 197 at 6. 

Product Demonstration Plan

As an initial matter, the protester contends that the Army unreasonably evaluated GDMS's demonstration by assessing weaknesses and significant weaknesses for tasks that could not be adequately demonstrated under the demonstration guidelines.  See Protest at 9; Protester's Comments at 1.  The protester also claims that discrepancies among the various demonstration documents call into question the agency's evaluation of GDMS's demonstration.  See Protester's Comments at 5-12.

As set forth above, offerors were to conduct a product demonstration as part of their technical proposal.  The solicitation included a product demonstration plan (PDP), with demonstration guidelines and a sample walkthrough, or workflow.  PDP at 6; Sample Demo. at 9-19.  The offeror was to bring two workstations (i.e., its proposed hardware) to the demonstration and could bring up to three personnel, including a system operator who alone would "attempt to perform all required tasks" on the offeror's proposed system in the presence of Army observers.  RFP at 64; PDP at 1.  The guidelines stated, among other things, that the moderator would direct the offeror through a series of steps similar to those in the sample walkthrough and direct the offeror when to move to the next step; that the offeror may ask for clarification of a step, but not solicit any feedback; and that the offeror may talk through the steps to clarify the functionality of its proposed solution.  See PDP at 6.

The division chief, who served as the product demonstration moderator and chairman of the respective evaluation team, explains that:

[t]he Army carefully designed and structured the steps [in the demonstration] to reflect a realistic mission scenario a user will encounter once the system is deployed in combat.  The demonstration steps were sequential and built upon one another in the same manner the system would be used in the field under realistic operating conditions and workflows.  The steps model the typical method or scheme of how an Army . . . Intelligence Analyst will use the system to support a combat mission from start to finish.  The steps were created, reviewed, and approved by Intelligence experts from Training and Doctrine Command (TRADOC), the Army proponent for Doctrine and concepts of employment for new systems.  The steps correspond to realistic Army conditions and reflect the actual requirements of the Army.  The steps enabled the Army to evaluate all thirty-two demonstration tasks.

AR, Tab 4, Decl. of Div. Chief, at ¶¶ 7-9.

The actual demonstration walkthrough was provided to GDMS on December 7, 2017, "so that GDMS would have it in advance of its actual Product Demonstration the following day," December 8.[11]  Id. ¶ 15.  The 32 PDP tasks were identical in the RFP's sample walkthrough and the final demonstration walkthrough.  Compare PDP at 3-5, with AR, Tab 37, Final Demo. Walkthrough (Final Demo.), at 1-10, and Tab 28, Task Mapping Matrix, at 1-6.  However, the division chief states that the final walkthrough reflected "immaterial changes" from the sample walkthrough, "mainly to renumber the steps and to change the sample data so that the offeror's product would have to analyze new data in real-time rather than regurgitate stale data."[12]  Decl. of Div. Chief at ¶ 16.

GDMS disputes that the changes between the RFP's sample walkthrough and the actual walkthrough (provided to GDMS on December 7) were immaterial, as the Army maintains.  Protester's Comments at 10-12.  According to the protester, "[s]ince the Agency claims that 'the [walkthrough] steps correspond to realistic Army conditions and reflect the actual requirements of the Army . . . it is unknown which version's steps properly reflect the actual requirements."  Id. at 11.  In addition, the protester claims that there were "significant discrepancies" between the actual walkthrough and the script (or slides) that the moderator read during the demonstration.[13]  Id. 8-9; see Protester's Supp. Comments at 3-7.  According to GDMS, such discrepancies call into question whether the Army's evaluation "aligned with the demonstration required of GDMS."  Protester's Comments at 7.

The protester also complains that the demonstration guidelines limited GDMS's ability to fully demonstrate the capabilities of its proposed GeoSuite software.  See Protest at 18.  For example, GDMS claims that it was impossible to demonstrate capabilities that required automation, because an offeror was forbidden from "skipping steps and moving between steps," without giving the agency adequate time to record the results of those steps, or without the moderator's direction.  Id. at 13; Protester's Comments at 13-14.  The protester thus asserts that "GDMS demonstrated all of the PDP tasks that were capable of being completed by an offeror who faithfully followed the Final Demonstration Script in conformance with the Demonstration Conduct Guidelines."  Protest at 8.

These arguments lack merit.  Notably, the protester concedes that "the Agency had the right to make changes" to the sample walkthrough provided to offerors in the solicitation.[14]  Protester's Comments at 11.  Moreover, one of the protester's allegedly "significant" discrepancies is based on the instruction in a PowerPoint slide that the offeror unplug its system, which does not appear in either the sample or actual walkthroughs.  Id. at 9; see Sample Demo. at 9; Final Demo. at 1.  However, GDMS concedes that:  "[t]here may be a valid reason why the Agency did not tell offerors that the demonstration would start with a system that was powered off and that they would need to perform the demonstration operating on battery power . . . ."  Protester's Comments at 9.

Based on our comparison of the sample and actual walkthroughs, we find that the demonstration steps (including their associated figures) appear largely the same.[15]  Compare Sample Demo. at 9-19, with Final Demo. at 1-10.  To the extent that there are changes between the walkthrough documents, the changes are consistent with the Army's assertion that it mostly re-numbered, or reordered, the steps.  More importantly though, such changes are consistent with the RFP's warning that the actual demonstration could present different scenario, and be based on different data.[16]  See supra n.14.

Finally, GDMS's assertion that it could not demonstrate all of the required tasks based on the demonstration guidelines, also lacks merit.  Here, too, the protester makes significant concessions, stating that GDMS does not "dispute[] that the Moderator followed the Demonstration Conduct Guidelines during the demonstration," and does not "take exception to the particular rules imposed for the demonstration."  Protester's Comments at 2, 12.  Furthermore, as the agency points out, the guidelines explicitly provided that offerors could talk through the steps to complete a task to explain the functionality of their proposed system to evaluators, and that offerors could ask for clarification of a particular step if it was unclear.  COS/MOL at 32, 45, citing PDP at 6.  In addition, the RFP explicitly stated that an offeror's proposed solution was not required to successfully complete all PDP tasks.  RFP at 70.  Thus, we agree with the Army that if GDMS believed it was impossible to perform certain tasks within the demonstration guidelines, GDMS should have protested those terms prior to the RFP's closing time for receipt of proposals, consistent with our Bid Protest Regulations.[17]  See COS/MOL at 32, 36, 49. 

In sum, the protester's objections to the demonstration guidelines and the changes between the sample and actual demonstration walkthroughs, provide no basis to question the Army's evaluation of GDMS's product demonstration.

Performance Requirement 13.2

Requirement 13.2 of the performance requirement document (PRD) states that the DCSG-A CD-1 solution "shall provide a computer-assisted capability that allows the All Source User to modify the correlation rules used to determine entity likeness and combine entities."[18]  PRD at 10.  PRD 13.2 is one of several, interrelated data management capabilities, which according to the PRD, "support[] content extraction and entity refinement that aids in situational awareness."  Id.

Although the solicitation does not expressly define correlation, the Army and the RFP employ this term to refer to the requirement that DCGS-A be able to identify and "merge," or combine, multiple instances of the same entity (such as a tank), based on similar attributes (such as the type of tank, whether the tank belongs to friendly forces or the enemy, the tank's recent locations, and any number of additional attributes).[19]  See, e.g., Tr. at 38:2-42:12, 58:18-59:16.  Correlation involves establishing "correlation rules" in DCGS-A to automatically identify like entities based on a pre-programmed, and modifiable, list of attributes.  See COS/MOL at 27; Decl. of Div. Chief at ¶ 25.a(2); Protester's Comments at 20; Tr. at 598:6-8 (GDMS's senior software engineer testifying that "[c]orrelation is determining likeness.  You're identifying the characteristics that determine likeness.").  During the hearing, the chairman of the technical narrative evaluation team explained correlation as follows:

[I]n layman's terms, what that means is if you have a soldier that observes a tank and then you have a second observation [of the tank] made by a sensor . . . that second observation is in close proximity [to] that first observation made by the soldier, so . . . the all source analyst who is not co-located with either the soldier or the sensor has to essentially look at this information on a screen and determine whether they are dealing with one tank or two tanks.  The all source analyst's job is to build the intelligence picture for the commander, and the commander has to understand whether they are dealing with two targets or one target.

Tr. at 691:18-692-17.  According to the chairman, "the purpose of correlation is to analyze multiple entries on a map, each of which contains different data entries and attributes (perhaps location, number and strength of enemy forces, weaponry, etc.), and combine [or merge] them" so there are no duplicates of the same entity in DCGS-A.  See Decl. of Sys. Eng'r at ¶ 12.  The Army explains that correlation rules are written prior to a mission, but that users must be able to modify them during the mission as new information about entities is learned.  See Tr. at 64:4-24, 692:22-694:17.

Thus, briefly stated, PRD 13.2 requires that the DCGS-A user be able to modify the correlation rules (the list of attributes) that identify and merge the same entities.  PRD 13.5, which is essentially the objective of 13.2, requires that DCGS-A automatically merge similar entities as new, and possibly duplicative, intelligence information is entered or updated in DCGS-A.[20]  See PRD at 10; Tr. at 115:6-116:24, 298:9-22 (PRD 13.5 or Task "25 is where you want the system to be doing this [merging] in the background.").

At issue here is GDMS's and the Army's fundamental disagreement as to whether the RFP's correlation requirements also include simultaneously merging similar entities (the Army's view)--or whether the process of merging similar entities is an entirely separate, albeit successive, step (GDMS's view).

GDMS's Technical Evaluation

Under the product demonstration subfactor, the Army assessed a significant weakness to GDMS's proposal for not demonstrating an adequate approach and understanding of Task 22, which corresponds to PDP 13.2.  AR, Tab 35, GDMS Demo. Eval. Rep., at 9.  The agency recognized that GDMS demonstrated a "solution that allowed the user the capability to modify and create conditions to identify entity likeness" and "to determine if two entities were the same."  Id.  However, the Army assessed a significant weakness because GDMS did not demonstrate the capability to create or modify correlation rules used to combine entitiesId. (emphasis added).  As the chairman of the demonstration evaluation team explains:

During the demonstration the protester successfully demonstrated a portion of the task by creating conditions to search for entity likeness. . . . Nevertheless, overall the protester failed the Task. . . . There simply was no ability for the user to modify correlation rules used to combine entities . . . . the protester demonstrated only an "option to merge," not an ability to modify correlation rules used to combine entities, which was the Army's specific requirement.  The user could manually merge the entities, but the protester did not demonstrate an ability to modify the rules to combine the entities.

Decl. of Div. Chief at ¶ 25.a(2) (internal citations omitted).  The evaluators found that this would compromise data integrity, could result in multiple instances of the same enemy units/equipment appearing in the data, and increased the user's cognitive workload while trying to determine which enemy unit/equipment is correct.  AR, Tab 35, GDMS Demo. Eval. Rep., at 9.  The Army concluded that this was a flaw in GDMS's demonstration that appreciably increased the risk of unsuccessful contract performance.Id.

Similarly, under the technical narrative subfactor, the Army assessed a deficiency because GDMS's technical narrative did not provide an approach to meeting PRD 13.2.  AR, Tab 41, GDMS Tech. Narrative Eval. Rep., at 1-2.  The technical narrative evaluators found--based on their independent review of GDMS's technical narrative, cross reference matrix, and GeoSuite's (GDMS's proposed software) user manual--that GDMS did not propose a technical solution to meeting PRD 13.2 that would allow a user to "modify the correlation rules used to determine entity likeness or to combine entities."  Id. at 1.  The evaluators also noted that GDMS did not "provide any indication how [it] intends to meet this requirement within 22 weeks after award of delivery order 1."  Id. at 2.  During the hearing, the chairman of the technical narrative evaluation team explained that the GeoSuite software suite includes a search function (described below) and two additional applications--a duplicate detection tool and a content merge tool--but no capability for a user to establish a rule that would merge entities.  See Tr. at 701:7-21.

GDMS protests the assessments under both technical evaluation subfactors, contending they are inconsistent with PRD 13.2 and related data management requirements as set forth in the RFP, namely, the correlation requirements.  Protest at 22-23.  As noted above, the protester interprets PRD 13.2 as contemplating two separate, but related actions:  first, modifying correlation rules used to determine entity likeness, then combining those entities.  Protester's Comments at 20, 36.  GDMS argues that this is the only reasonable interpretation of PRD 13.2 and that the Army "misunderstand[s] its own requirements."  Id.  GDMS finds support for its interpretation in an engineering diagram included with the PRD.[21]  Id. at 21.  In the protester's view, it met the requirement of PRD 13.2 because GeoSuite permits the user to establish correlation rules to determine entity likeness.  GDMS explains that GeoSuite does not use the term correlation, but that a GeoSuite user performs correlation by setting up an automated, recurring "search" based on specified criteria.  Id. at 38; Tr. at 412:2-17, 599:5-15.  GDMS explains further that the "automated search continually monitors the data based upon the search criteria with no additional input from the user."  Protester's Comments at 38.

We agree with the Army's view that correlation, as contemplated in the solicitation, requires that the DCSG-A CD-1 solution not only be capable of identifying like entities, but also merge like entities.

Significantly, GDMS concedes that its GeoSuite software does not provide the capability to set or modify correlation rules in order to merge like entities.  See id. at 36 ("[T]here is no modification of 'correlation rules' to 'combine entities' because 'correlation rules' are a precursor to combining entities, but not the means by which entities are combined."); Tr. at 414:9-22, 484:15-18.  In fact, with respect to PRD 13.5, GDMS concedes that GeoSuite does not provide an automated merge capability at all.[22]  Tr. at 760:21-24, 762:12-15.  These concessions support the Army's conclusions--reached by two independent teams of evaluators--that GDMS's solution is not capable of modifying (or even establishing) correlation rules to both identify and merge like entities.[23]

Therefore, the question for resolution here is not so much whether the Army evaluated GDMS's technical proposal consistent with the RFP's evaluation criteria and performance requirements (which it did), but whether the protester's interpretation of requirement 13.2 is reasonable.  We find GDMS's interpretation unreasonable.  In order for an interpretation to be reasonable, a solicitation must be read as a whole and in a manner that gives effect to all of its provisions.  See Northrop Grumman Info. Tech., Inc., B-401198, B-401198.2, June 2, 2009, 2009 COD ¶ 122 at 2.  Where the reasonableness of the evaluation turns on the agency's interpretation of a solicitation provision, the agency's interpretation of the provision must be consistent with the solicitation when read as a whole and in a reasonable manner.  Solec Corp., B-299266, March 5, 2007, 2007 CPD ¶ 42 at 2.

The RFP, read as a whole, supports the Army's view that correlation requires the capability to merge similar entities in DCGS-A.  For example, PRD 13.5, which is essentially the objective of PRD 13.2 as noted above, requires that DCGS-A automatically merge similar entities.  In this respect, the PRD defined automated as "[o]nce the baseline data is input, the computer performs a function continuously without input from an operator.  For example, the computer decides if two reported instances of tanks are the same and if so, merges them into one."  PRD at 6 (emphasis added).  Similarly, PRD 13.6, another related data management capability, requires a "correlation (merge) capability."  PRD at 10.  Moreover, the RFP provided that the Army would evaluate the extent to which an offeror's proposed solution, as demonstrated, was complex and required numerous steps, as well as the solution's speed and ease of use.  See RFP at 70; PDP at 3.

To the extent that GDMS relies on two text boxes in the PRD's engineering diagram (which the protester insists define correlation, see supra n.21; Tr. at 624:20-625:19), the PRD explicitly described the diagrams and their accompanying workflows[24] as "supplemental materials," designed "to illustrate activity based events that occur during the operational execution of a maneuver battalion. . ." and provide "a holistic view of the total end to end operational activities and inputs/outputs of the DCGS-A battalion system."  PRD at 9; app. A at 2 (emphasis added).  In this respect, GDMS misreads the diagram and its accompanying workflow, which reflect numerous data management processes, not simply PRD 13.2.  See PRD, app. A, at 32-33.  Likewise, GDMS overlooks the fact that the diagrams are introduced with examples and quotes from soldiers (the typical DCGS-A user) who "may not be comfortable using a computer beyond basic functions," and who express a desire that DCGS-A be easy to operate and provide quick analyses, among other things.  See id. at 3-7. 

In our view, these various solicitation provisions, read as a whole, leave little doubt that an offeror's solution was to simplify DCGS-A data management performance requirements.  Indeed, GDMS concedes that different offerors could propose solutions that were more automated and that "there are obvious ways that are simpler or more straightforward flows for an operator to do an activity."[25]  See Tr. at 658:24-659:10, 671:21-672:7.

We therefore find no basis to question the Army's evaluation that GDMS' technical proposal did not meet the requirements of PRD 13.2.  The protester's allegations to the contrary only reflect its disagreement with the agency's evaluations, which provides no basis to question the reasonableness of the Army's judgments.  See Citywide Managing Servs. of Port Washington, Inc., B-281287.12, B-281287.13, Nov. 15, 2000, 2001 CPD ¶ 6 at 10-11.  Since the RFP provided that the assessment of one deficiency under any technical subfactor would render the entire proposal unacceptable and awardable, see RFP at 72, we also have no basis to question any of the other deficiencies, significant weaknesses, or weaknesses assessed against GDMS's technical proposal.

Best-Value Tradeoff

Finally, GDMS challenges the Army's source selection decision, arguing that the best-value tradeoff was unreasonable insofar as it relied on the allegedly flawed evaluations above.

As discussed above, the record does not support GDMS's assertions that its technically evaluation was flawed.  Therefore, we have no reason to question the SSA's reliance on the evaluators' assessments in conducting his tradeoff and best-value determination.  While GDMS disagrees with the SSA's decision, the protester's disagreement provides no basis to question the reasonableness of the agency's judgments.  See id.

The protest is denied.

Thomas H. Armstrong
General Counsel



[1] The awardees are Raytheon Co., of Garland, Texas, and Palantir Technologies, Inc., of Palo Alto, California.

[2] Our citations are to the conformed version of the RFP provided in the agency report.

[3] On June 19-20, 2018, GAO conducted a hearing, on the record, during which testimony was provided by three witnesses:  the division chief, a DCGS-A systems engineer, and GDMS's senior software engineer.

[4] The RFP states that as technology evolves and new warfighting requirements emerge, the DCGS-A capability set will need to be updated to meet these future requirements.  AR, Tab 6, RFP attach. 1, SOW, at 4.  The Army states that it anticipates conducting future capability drops in this respect.  Army Letter to GAO, June 11, 2018; see, e.g., DCGS-A CD-2 Draft RFP, available at https://www.fbo.gov/notices/ a86be14f96d3508e9a20f56917377f52 (last visited June 28, 2018).

[5] GDMS does not challenge the evaluation of its proposal under the other technical subfactor (usability maturity) or the past performance or price factors.  GDMS also does not challenge the Army's evaluation of the awardees' proposals.

[6] The division chief explained that these 32 tasks/performance requirements were selected so that evaluators could focus on an offeror's solution itself, rather than the remaining 40 or so performance requirements which could involve other Army systems.  See Tr. at 33:19-34:11.

[7] The RFP defined unacceptable as not meeting solicitation requirements, "and thus, contain[ing] one or more deficiencies, and/or risk of unsuccessful performance is unacceptable.  Proposal is unawardable."  RFP at 72.  Deficiency was defined as a material failure of a proposal to meet a requirement or a combination of significant weaknesses in a proposal that increases the risk of unsuccessful contract performance to an unacceptable level.  Id. at 70.  Significant weakness was defined as a flaw that appreciably increases the risk of unsuccessful contract performance.  Id.

[8] Offerors were to submit proposals electronically.  RFP at 59.

[9] According to the DCGS-A systems engineer who served as the evaluation chairman for the narrative subfactor (and who testified at the GAO hearing), "[o]fferors were not required to have a ready-made solution when submitting their written proposal for this Technical Subfactor, rather the solicitation allowed offerors to explain in their written proposal how they could meet the requirements within 22 weeks of award."  AR, Tab 3, Decl. of Sys. Eng'r, at ¶ 6.

[10] Moreover, the RFP provided that the assessment of one deficiency under any technical subfactor would render the entire proposal unacceptable.  RFP at 72.  As a result, we need not address the protester's challenges to all of its assessed deficiencies, since we find the Army reasonably assessed one deficiency under the technical narrative subfactor with respect to GDMS's proposed solution to meeting performance requirement 13.2, as discussed below.

[11] The demonstration could not exceed 2 days; the offeror was to set up and configure its system on day 1, and perform the demonstration on day 2.  RFP at 64. 

[12] Offerors could request sample data before the actual demonstration for information purposes.  RFP at 64; AR, Tab 16, RFP attach. 9, Sample Demo. Data.

[13] During the demonstrations, the agency projected PowerPoint slides onto a screen that the moderator read to direct the offeror through each step of the demonstration.  See RFP, Tab 36, Demo. Slides, at 1-70; Decl. of Div. Chief at ¶ 19; Supp. AR, Tab 55, Supp. Decl. of Div. Chief, at ¶¶ 5-9.

[14] The RFP explicitly advised offerors that the actual demonstration would use different scenarios and data, and that the product demonstration plan (including the sample walkthrough) was provided for evaluation purposes only.  Sample Demo at 9; see RFP at 70.

[15] As stated above, the 32 PDP tasks are identical in the sample and actual demonstrations.

[16] We also compared the actual demonstration walkthrough and the moderator's PowerPoint slides and find that GDMS's concerns in this respect are overstated.  In our view, the differences between the slides and the actual walkthrough are not significant, but reflect inconsequential changes, for example, to omit the titles of lengthy and complex data files.  Compare Final Demo. at 1-10, with Demo. Slides at 1-70.

[17] Our timeliness rules specifically require that a protest based upon alleged improprieties in a solicitation that are apparent prior to the closing time for receipt of initial proposals be filed before that time.  4 C.F.R. § 21.2(a)(1); see AmaTerra Envtl. Inc., B-408290.2, Oct. 23, 2013, 2013 CPD ¶ 242 at 3.

[18] The PRD defined computer-assisted as:  "The operator tells the computer to perform a function and the computer performs the function until complete.  For example, the user selects a location and tells the computer to determine if [an entity such as] a T-80 tank can see friendly defensive positions considering terrain masking (intervisibility)."  PRD at 6.  (An entity could be a person, tank, building, or any other element.  See Tr. 38:2-6.)

[19] The chairman of the technical narrative evaluation team testified that correlation and correlation rules are Army vernacular used by DCGS-A writers and developers.  See Tr. at 735:7-16.

[20] The PRD defined automated as "[o]nce the baseline data is input, the computer performs a function continuously without input from an operator.  For example, the computer decides if two reported instances of tanks are the same and if so, merges them into one."  PRD at 6.

[21] GDMS points out that the engineering diagram associated with data management depicts "correlation rules modification" and "correlation" in different text boxes, and as sequential steps.  Protester's Comments at 21.

[22] GDMS was also assessed a significant weakness under the demonstration factor, as well as a deficiency under the narrative subfactor, because its GeoSuite software does not provide an automated capability to merge similar entities as new information is entered into DCGS-A.  See AR, Tab 35, GDMS Demo. Eval. Rep., at 10; Tab 41, GDMS Tech. Narrative Eval. Rep., at 2.  Both sets of evaluators found that GeoSuite provided a manual process for merging entities that prompts the user when GeoSuite finds instances of entities that appear similar, and then requires the user to manually merge the entities.  See id.

[23] As previously discussed, the demonstrations and narratives were evaluated by separate evaluation teams.  COS/MOL at 6.  According to the chairman of the demonstration evaluation team, his team was comprised of seven individuals, including the chairman, four DCSG-A users, and two subject matter experts.  Tr. at 27:11-21.  He testified that the evaluators discussed the demonstration at length; that they reached the same understanding of what they saw; and that they arrived at a consensus on each one of the findings, with no dissent.  Id. at 316:13-20; see Decl. of Div. Chief at ¶ 22.  According to the chairman of the narrative evaluation team, his team was comprised of system engineers, software engineers, and DCSG-A users.  Tr. at 734:3-7.  He testified that "each team member evaluated proposals independently" (including reviewing GeoSuite's user manual and "every single requirement in [GDMS's] cross reference matrix"), that the team discussed their individual findings, and that there was no dissent regarding the findings and assessed deficiencies.  See id. 710:22-24, 734:10-22.

[24] To be clear, the "operational workflows" associated with the PRD's engineering diagrams are distinct from the demonstration walkthrough.

[25] In fact, during his testimony, GDMS's senior software engineer recognized that:

[T]he user has to be able to work potentially in a vehicle, has to work in a combat situation, all those things are taken into account for system speed and ease of use.  The simplest, quickest, least number of clicks . . . to complete every step . . . [for] an operator and analyst do this job, effectively . . . in a simple manner . . . .

Tr. at 613:11-614:2.

Jul 20, 2018

Jul 19, 2018

Jul 18, 2018

Jul 17, 2018

Looking for more? Browse all our products here