In the United States Court of Federal Claims Nos. 24-1862, 24-1863, 24-1875 (consolidated) (Filed Under Seal: April 18, 2025) Reissued: April 28, 2025 1 ) RED CEDAR HARMONIA, LLC, ) ) Plaintiff, ) ) SYSTEMS ENGINEERING SOLUTIONS ) CORPORATION, ) ) Consolidated Plaintiff, ) ) and ) ) INTELLECT SOLUTIONS, LLC, ) ) Consolidated Plaintiff, ) ) v. ) ) THE UNITED STATES, ) ) Defendant, ) ) and ) ) IGNITEACTION, LLC, ) ) Defendant-Intervenor. ) ) John Davidson Levin, Womble Bond Dickinson, Huntsville, AL, for Red Cedar Harmonia, LLC; Samuel S. Finnerty, PilieroMazza PLLC, Washington, DC, for Systems Engineering Solutions Corporation; Tracye Winfrey Howard, Wiley Rein LLP, Washington, DC, for Intellect Solutions, Inc.
Sosun Bae, U.S. Department of Justice, Civil Division, Washington, DC, for defendant.
Elizabeth Newell Jochum, Blank Rome, Washington, DC, for defendant-intervenor.
1 An unredacted version of this opinion was issued under seal on April 18, 2025. The parties proposed redactions, which the Court adopts only in part owing to the “presumption of public access to judicial records.” Allied Tech. Grp., Inc. v. United States, 94 Fed. Cl. 16, 23 n.1 (2010) (quoting Baystate Techs., Inc. v. Bowers, 283 Fed.Appx. 808, 810 (Fed. Cir. 2008)). OPINION AND ORDER SMITH, Senior Judge
In this consolidated post-award bid protest, the plaintiffs—Red Cedar Harmonia, LLC (“Red Cedar”), Systems Engineering Solutions Corporation (“SES”), and Intellect Solutions, LLC (“Intellect”)—challenge the decision of the U.S. Census Bureau (“Census”) to award a small-business set-aside contract for IT support services to defendant-intervenor, IgniteAction, LLC (“IgniteAction”). Among other things, each protestor challenges the Agency’s evaluation of IgniteAction’s proposal as well as the protestor’s own proposal. Finding no error, the Court denied the protest soon after oral argument. See Order of February 18, 2025, ECF No. 60. 2 This opinion follows.
I. Background
A. Solicitation
The Census Bureau is the largest statistical agency in the federal government. Taylor R. Knoedl, Cong. Rsch. Serv., RL47847, The U.S. Census Bureau: An Overview (2023). Until recently, Census stored most of its data, including confidential data, in “on-premise data center infrastructure.” AR 6; see 13 U.S.C. §§ 9, 214 (requiring non-disclosure of confidential data). In 2022, Census launched the Census Acceleration to Secure Cloud (“CASC”) Initiative, an ambitious program to modernize data management practices by migrating data to secure cloud infrastructure. AR 11, 13. To address the logistical and cybersecurity challenges arising from the move, Census sought to “leverage industry knowledge and expertise” on “proven approaches and best practices in the adaptation and implementation of modern technologies” that would help “accelerate the . . . adoption of secure cloud capabilities” but also “continue to support existing IT and infrastructure.” AR6.
The instant solicitation resulted. On June 10, 2024, Census issued RFQ No. 1333-LB-24- JK0001 under Federal Acquisition Regulation (“FAR”) 8.405-3, seeking quotations for a single- award blanket purchase agreement (“BPA”) 3 for “on-premises and cloud infrastructure support services” supporting the CASC Initiative. AR 180, 508, 673. The Solicitation was issued as a small business set-aside, and to be eligible for consideration, quoters were further required to hold a General Services Administration (“GSA”) Multiple Award Schedule (“MAS”) contract under Special Item Number 518210C, Cloud Computing and Cloud Related IT Professional Services. AR 673; see also FAR 8.4 (Federal Supply Schedules). The BPA would be applicable
2 Unless stated otherwise, this Opinion cites to filings and docket entries in the lead case, Red Cedar Harmonia, LLC v. United States, No. 24-1862. 3 A BPA is a “simplified method of filling anticipated repetitive needs for supplies or services by establishing ‘charge accounts’ with qualified sources of supply.” FAR 13.303-1(a). BPAs are not contracts as such. They specify a “set of terms and conditions that will be applicable to future purchases” but do not commit the government to “any specific purchase.” J. Cibinic, Jr., et al., Formation of Government Contracts § 8.04[E] (5th ed. 2023). Contracting officers use them as a “convenient way . . . to negotiate an across-the-board discount for all of the supplies or services furnished by a single supplier.” Id.; FAR 13.303-2(d)(1) (“[s]ecuring maximum discounts”).
2 for a base period of one year with four additional one-year option periods, during which time Census would procure services by issuing individual orders. AR 673. Census assigned an estimated value of $557,000,000.00 to the BPA. AR 674. The Solicitation stated that Census would evaluate the proposals on three factors— demonstrated prior experience, technical approach, and price. AR 761. Both the submission of offers and their evaluation would be done in two phases. 4
In Phase I, Census would evaluate the offerors’ prior experience. Offerors would submit the first volume (dubbed Volume I) describing their “demonstrated prior experience . . . as a prime or subcontractor on a maximum of three (3) contracts . . . similar [in] size, scope, and complexity” to the Solicitation “that have been performed or are currently being performed within the last five (5) years.” AR 765. Offerors were asked to describe their past performance in relation to five “Experience Elements”:
• 1001: “supporting On-Premises operational tasks,” • 1002: “monitoring . . . enclaves, both On-Premises and within Cloud environments,” • 1003: “designing, building, and maintaining a multi-cloud/hybrid-cloud architecture” that is “centralized” and “interoperab[le] with workloads hosted in a traditional data center,” • 1004: “migrat[ing] on-premises applications to the cloud” and “facilitat[ing] the development of cloud native applications,” and • 1005: “providing detailed cost analysis and enforcing compliance” with “prescribed financial design factors.”
AR 766–67. For each element, the offerors were also required to describe the tools and approach they used to the objective. AR 766–67. Based on the submissions, Census would assign a rating of either “low,” “some,” or “high” confidence to each offeror’s prior experience. AR 778. Quoters whose submissions “provide[d] the Government with the greatest levels of confidence” would be advised to participate in Phase II. AR 768. Quoters “not amongst the most highly rated” would be informed that they “are unlikely to be viable competitors,” but they would nevertheless be permitted to participate in Phase II. AR 768.
In Phase II, Census would evaluate offerors on the other two factors, technical approach and price. Offerors electing to participate in Phase II would submit two more volumes, Volumes II and III. In Volume II, they would propose “methodologies and approaches for fulfilling the requirements” of the Solicitation. AR 769. The factor was subdivided into five “Functional Areas”:
• 0001: “Operational Support (On-Premises),” • 0002: “Enterprise Multi-Cloud/Hybrid-Cloud Vision,” • 0003: “Integration and Migration Engineering,” • 0004: “Operational Support (Cloud),” and
4 The Solicitation provided for an optional third phase, but Census eventually made the award decision at the end of Phase II. AR 772–73, 1496.
3 • 0005: “Program and BPA Management.”
AR 769–70. Like the first factor (demonstrated prior experience), offerors’ technical approach would be assigned “low,” “some,” or “high” confidence. AR 778.
In Volume III, offerors would provide a “Business/Price Quote” consisting of a narrative explanation and filled out “BPA Rate Mapping” and “BPA Rate Card” schedules. AR 761, 770– 71. The rate mapping schedule specified a list of exactly fifty “BPA Role[s]” that the awardee’s employees would be expected to perform. AR 783–87. The roles described a type of position and in some cases also seniority level (such as “Backup/Recovery Engineer – Senior Level” and “Financial Analyst”), and each role was accompanied by a detailed “BPA Role Description” in narrative form. AR 783–87. Offerors—who, recall, each held a GSA MAS contract for Cloud Computing and Cloud Related IT Professional Services—were asked to (1) “map” each BPA Role to a labor category (“LCAT”) in the Federal Supply Schedule covered by their MAS contract, and (2) describe the minimum education and experience requirements associated with the chosen labor category for each BPA role. AR 783–87. Then, in the BPA Rate Card, the offeror would, for each BPA role and for each annual period: (1) the GSA MAS hourly wage for the chosen LCAT, and (2) the offeror’s proposed (and ideally discounted) hourly wage. AR 788–89.
For example, IgniteAction mapped BPA Role of “Backup/Recovery Engineer – Senior Level” to “ .” AR 1155. IgniteAction then described the job duties of employees in that LCAT and specified that such employees must have a degree and work experience. AR 1155. Finally, in the BPA Rate Card, IgniteAction stated that the posted hourly rate for that LCAT for the base period was $ , but IgniteAction would offer a discount and charge $ per hour. AR 1163.
The Solicitation stated that Census would evaluate the Business/Price Quotes as follows. A price submission would be deemed “complete[] and accura[te]” if “the Quoter has properly understood the price quote instructions” contained in the RFQ and rate card attachment. AR 778. The quoters’ submissions would “be checked for mathematical correctness to include . . . arithmetic” accuracy and Census would also check whether “all prices were summarized correctly.” AR 778. Census would then determine the total evaluated price using the “Rates Only for Price Evaluation” (“ROPE”) method, under which the total evaluated price is determined by multiplying the rates provided by the offeror in their Rate Card and a “pre- determined estimated number of hours for each labor category.” AR 779. Prices would be evaluated for reasonableness. AR 779.
Finally, Census would make an award to “the responsible Quoter whose quote provides the greatest overall value to the Government, price and all other factors considered.” AR 776. The Solicitation advised offerors that “Factor 1 [demonstrated prior experience] is more important than Phase II . . . evaluation factors [technical approach and price].” AR 768 (emphasis deleted). It also advised that “[w]hen combined, non-price evaluation factors are significantly more important than Factor 3 (Price),” and that price will become “more important in the assessment” only if competing quotes “approach parity” along the non-price factors. AR 776.
4 The Court granted the motion the following day. Order Granting Motion to Complete the Administrative Record, ECF No. 38.
Red Cedar, SES, and Intellect filed their respective motions for judgment on the administrative record (“MJARs”) on December 11, 2024. ECF Nos. 41, 44, 45-2 [hereinafter, respectively, “Red Cedar MJAR,” “SES MJAR,” and “Intellect MJAR”]. In addition, two plaintiffs—Intellect and SES—amended their respective complaints. ECF Nos. 42, 43. On January 3, 2025, defendant and defendant-intervenor filed their respective cross-motions for judgment on the administrative record (“CMJARs” or “MJARs”). ECF Nos. 48, 49 [hereinafter, respectively, “US MJAR” and “IgniteAction MJAR”]. On January 15 and 16, 2025, plaintiffs filed their respective response-and-reply briefs. ECF Nos. 51, 52, 53 [hereinafter, respectively, “Intellect Reply,” “SES Reply,” and “Red Cedar Reply”]. On January 28, 2025, defendant and defendant-intervenor filed their respective reply briefs. ECF Nos. 54, 55 [hereinafter, respectively, “US Reply” and “IgniteAction Reply”].
The Court held oral argument on February 6, 2025. On February 18, 2025, the Court denied the plaintiffs’ MJARs and granted defendant’s and defendant-intervenor’s CMJARs. Order of February 18, 2025, ECF No. 60. The Clerk of Court entered judgment on February 19, 2025. Judgment, ECF No. 61. The Court’s Order ruling on the motions stated that an opinion would follow in due course. Order of February 18, 2025, at 1. This is that opinion.
II. Standard of Review
The Tucker Act, as amended, confers jurisdiction upon this Court to “to render judgment on an action by an interested party objecting to a solicitation by a Federal agency for bids or proposals for a proposed contract or to a proposed award or the award of a contract or any alleged violation of statute or regulation in connection with a procurement or a proposed procurement.” 28 U.S.C. § 1491(b)(1). The Court’s jurisdiction therefrom includes bid protests arising from a BPA award pursuant to FAR Part 8. Labat-Anderson Inc. v. United States, 50 Fed. Cl. 99, 104 (2001).
Rule 52.1 of the Rules of the United States Court of Federal Claims (“RCFC”) governs motions for judgment on the administrative record. In ruling on such motions, the Court’s task is to determine whether the agency concerned acted in compliance with the legal standards governing the decision under review, given all disputed and undisputed facts in the record. R. Ct. Fed. Cl. 52.1; see also Supreme Foodservice GmbH v. United States, 109 Fed. Cl. 369, 382 (2013) (citing Fort Carson Support Servs. v. United States, 71 Fed. Cl. 571, 585 (2006)). Our Court’s review is generally limited to the administrative record, and we must make findings of fact as if conducting a trial on a paper record. R. Ct. Fed. Cl. 52.1; Bannum, Inc. v. United States, 404 F.3d 1346, 1354 (Fed. Cir. 2005).
This Court reviews bid protests in accordance with the standards set forth in the Administrative Procedure Act (“APA”). 28 U.S.C. § 1491(b)(4) (incorporating by reference 5 U.S.C. § 706); Axiom Res. Mgmt. v. United States, 564 F.3d 1374, 1381 (Fed. Cir. 2009) (citing Impresa Construzioni Geom. Domenico Garufi v. United States, 238 F.3d 1324, 1332 (Fed. Cir. 2001)). Agency procurement actions may be set aside if they are “arbitrary, capricious, an abuse
6 of discretion, or otherwise not in accordance with the law.” 5 U.S.C. § 706(2)(A). To prevail, the protestor who must show that the agency’s decision either lacks a rational basis or involves a “a clear and prejudicial violation of applicable statutes or regulations.” Impresa, 238 F.3d at 1333 (quoting Kentron Hawaii, Ltd. v. Warner, 480 F.2d 1166, 1169 (D.C. Cir. 1973)).
Rational basis review is “highly deferential” and contracting officers are “entitled to exercise discretion upon a broad range of issues confronting them in the procurement process.” See Savantage Fin. Servs., Inc. v. United States, 595 F.3d 1282, 1286 (Fed. Cir. 2010) (first quoting CHE Consulting, Inc. v. United States, 552 F.3d 1351, 1354 (Fed. Cir. 2008) then quoting Impresa, 238 F.3d at 1332) (internal quotation marks omitted). The Court cannot substitute its judgment for that of the agency, even if reasonable minds could reach differing conclusions. Bowman Transp., Inc. v. Ark.-Best Freight Sys., Inc., 419 U.S. 281, 285–86 (1974); Honeywell, Inc. v. United States, 870 F.2d 644, 648 (Fed. Cir. 1989) (quoting M. Steinthal & Co. v. Seamans, 455 F.2d 1289, 1301 (D.C. Cir. 1971)) (“If the court finds a reasonable basis for [an] agency’s action, the court should stay its hand even though it might, as an original proposition, have reached a different conclusion as to the proper administration and application of the procurement regulations.”). Accordingly, the Court will “interfere with the government procurement process ‘only in extremely limited circumstances.’” EP Productions, Inc. v. United States, 63 Fed. Cl. 220, 223 (2005) (quoting CACI, Inc.-Fed. v. United States, 719 F.2d 1567, 1581 (Fed. Cir. 1983)).
III. Discussion
The protestors have collectively raised a bevy of objections to the award decision. The Court considers each in turn.
A. Census’s Evaluation of IgniteAction’s Demonstrated Prior Experience (Factor 1)
Two protestors, SES and Intellect, take issue with Census’s evaluation of IgniteAction’s quote under Factor 1, “demonstrated prior experience.” IgniteAction—a joint venture between mentor ActioNet, Inc. (“ActioNet”) and protégé Ignite IT, LLC (“Ignite”)—submitted three contracts for evaluation: two contracts performed by ActioNet for the Department of Energy (“DOE”) and the Department of Transportation (“DOT”), respectively; and one contract with the Department of Homeland Security (“DHS”) for which Ignite was the subcontractor. AR 824. IgniteAction stated that, as against the BPA’s estimated value of $557 million, the value of these contracts was $1.3 billion, $305.1 million, and $111 million, respectively. AR 824. Census assigned a “high confidence” rating on Factor 1 to IgniteAction based on its determinations that the submission demonstrated ample experience along all five Experience Elements, and there were “no notable aspects which decreased the Government’s confidence.” AR 1707; 1505–10.
a. Past Performance Evaluation: Recency of IgniteAction’s Prior Work Experience.
SES and Intellect argue, firstly, that Census should have disregarded most of the work performed under the first contract cited by IgniteAction—ActioNet’s DOE contract—because the period of performance of that contract (February 2019 through July 2019) overlapped only
7 slightly with the five-year recency window specified in the Solicitation (which began sometime in June 2019). 5 SES MJAR at 20; Intellect MJAR at 26. The protestors present two distinct theories. First, they suggest that Census should have treated the DOE contract—an omnibus indefinite-delivery, indefinite-quantity (“IDIQ”) contract—as a series of delivery orders under which the work was performed, and considered only those delivery orders that overlapped with the five-year recency period. SES Reply at 10. According to protestors, publicly available data shows that the DOE contract is associated with six separate delivery orders, only three of which overlap with the recency window. Intellect MJAR at 27. The protestors would have Census ignore, on recency grounds, the work performed under the other three orders. Id. at 27; SES MJAR at 15.
The Court disagrees that Census was required, or even permitted, to break down the DOE contract into individual delivery orders for purposes of determining recency. Start with the plain language of the Solicitation, which must be gleaned using “principles governing interpretation of [g]overnment contracts.” Banknote Corp. of Am., Inc. v. United States, 365 F.3d 1345, 1353 n.4 (Fed. Cir. 2004); see also FAR 8.405-3(b)(2)(vi) (requiring the contracting officer to “ensure [that] all quotes received are fairly considered and award is made in accordance with the basis for selection in the [Solicitation]”). As the protestors themselves acknowledge, the DOE contract is a government contract. Intellect Reply at 17 (citing PAE-Parsons Glob. Logistics Servs., LLC v. United States, 147 Fed. Cl. 294, 303 (2020)). Thus, it is also a contract within the meaning of the Solicitation, and Census has no discretion to replace the contract proffered by IgniteAction with some other contract while evaluating IgniteAction’s proposal. Indeed, if Census did so, it would be imposing unstated evaluation criteria, rendering the decision arbitrary and capricious. Samsara Inc. v. United States, 166 Fed. Cl. 457, 471 (2023). To be sure, protestors are right that the six individual delivery orders are also contracts. But the fact that individual delivery orders are also contracts vests discretion in IgniteAction, not Census. IgniteAction could have chosen to cite to just one delivery order, but given IgniteAction’s choice to rely on the master contract, Census properly evaluated that contract for recency rather than any other.
Intellect argues that the reading just adopted by the Court would lead to disparate treatment, because Census treated two directly related task orders cited by Intellect in its own past performance submission as two distinct contracts. Intellect MJAR at 27 n.5 (citing AR 1527); Intellect Reply at 18. That objection is misplaced. As a preliminary matter, Census followed the terms of the Solicitation. Unlike IgniteAction, Intellect rested on its task orders, not a master contract. 6 AR 859. Therefore, Census had no legal basis to treat Intellect’s task orders
5 The parties disagree about when the five-year look-back period begins. The Solicitation describes the time limitation as “within the last five (5) years of the date of RFQ release.” AR 765. The RFQ (i.e., the Solicitation) was initially released on June 10, 2024, but amended on June 26, 2024. The five-year recency window therefore began either on June 10, 2019, or on June 26, 2019. Either way, the period of performance of the DOE contract overlaps with the recency window. 6 In its reply brief, Intellect clarifies that its task orders were issued pursuant to the same IDIQ contract, and that its submission mentioned that contract by number. Intellect Reply at 23. Intellect misses the point. Unlike IgniteAction, Intellect chose to present its task orders for consideration, not the overall IDIQ contract. Census properly took the submissions on face value. Just as Census was not required, for purposes of determining recency,
8 as a single contract. And while that does mean that the Solicitation permits a master contract to count for more than the sum of its parts, that is sensible. To prevail on a disparate evaluation claim, “a protestor must show that the agency unreasonably downgraded its proposal for deficiencies that were substantively indistinguishable [from] or nearly identical [to] those contained in other proposals.” Off. Design Grp. v. United States, 951 F.3d 1366, 1372 (Fed. Cir. 2020) (internal quotation marks omitted). Where, as here, an agency is searching for a contractor with a comprehensive approach to fulfilling a BPA—itself a master contract—it may reasonably conclude that a master contract is more probative of the offeror’s fitness for the award than constituent task orders. In other words, it is proper for an agency to treat a master contract as relevantly different from task orders if the agency cares about the offeror’s ability to be a long- term partner in a complex federal initiative, as opposed to simply executing discrete deliveries. See Hanford Tank Disposition All., LLC v. United States, 173 Fed. Cl. 269, 289 (2024) (offeror’s “comprehensive” solution strategy pivotal to agency’s evaluation of quotes for a master contract). That is the case here. See AR 766–67 (Solicitation emphasizing under each Experience Element, by using terms such as “general approach,” “strategies,” and “methodologies,” that offerors should describe their big picture approach). Thus, the Solicitation reasonably puts IgniteAction’s master contract on a different footing than Intellect’s task orders for the purpose of counting the number of contracts submitted for consideration.
Second, and in the alternative, SES and Intellect argue that Census should have sliced the DOE contract (or the individual delivery orders) into work done before and after the commencement of the recency window, and considered only the work “actually performed . . . during the relevant period.” SES MJAR at 14, 2020; Intellect MJAR at 26. The Court acknowledges that dictum in a recent case, albeit with readily distinguishable facts, could lend some support to protestors’ theory. Fluor Fed. Servs., Inc. v. United States, 169 Fed. Cl. 70, 80 (2023) (“[A]ny consideration of past performance information . . . outside the stated recency window would be unreasonable.”). Nonetheless, the Court need not address that argument because protestors lack standing to make it. SES and Intellect themselves submitted contracts that overlap only partially with the recency window, and Census evaluated those contracts in their entirety. AR 858 ( the performance period outside recency window) (Intellect); AR 969 (same) (SES); AR 1511–14, 1524–27 (Census’s evaluation). “There has been no prejudice when a bid protestor benefited from the same potentially unlawful discretion from which the awardee benefited.” DigiFlight, Inc. v. United States, 165 Fed. Cl. 588, 606 (2023) (citation omitted).
Thus, the Court rejects Intellect’s and SES’s arguments that Census should have excluded on recency grounds work done pursuant to three delivery orders under the DOE contract, or work done prior to the commencement of the five-year look-back period. Census properly considered the entire contract. 7
to unbundle the contract proffered by IgniteAction, it was also not required to bundle the contracts proffered by Intellect. 7 To be clear, the Court only holds that Census was prohibited from excluding from consideration any work done under the master contract on recency grounds. Nothing in the opinion should be understood to require Census to treat work performed at different times as equally probative; the agency was free to treat “more recent experience [as] more relevant than old experience.” Intellect Reply at 20.
9 scope of [Census’s] intended effort” and stressed that it is “rare” to find experience “relative in size” to Census’s planned initiative. AR 1508.
Intellect argues that the scale of the effort under the DOT contract was “much smaller in scope and complexity,” because the DOT contract involved 126 applications to be exact, which is a much smaller number than the “250 applications” contemplated in the Solicitation. Intellect MJAR at 30 (citing AR 271, 832). But comparability is determined in context, and Intellect supplies no reason to think that a contract involving 126 applications is “much smaller in scope and complexity” that one involving 250 applications. Especially because experience at the larger end of the scale was “rare,” Census had reasonable grounds to decide that 126 applications and 250 applications are in the same ballpark.
Third, for Experience Element 1005—providing detailed cost analysis and enforcing compliance with prescribed financial design factors—Census favorably rated ActioNet’s “strategy for Tagging/Billing Governance and Control” under the DOE contract because Census was impressed by the “strategy to track costs by ” and the “Cost Monitoring Tool utilized” by ActioNet. AR 1509. Intellect objects on the grounds that IgniteAction did no more than “restate the Solicitation’s minimum requirements,” and therefore should not have been assigned a strength. Intellect Reply at 31; Intellect MJAR at 25–26; see also, e.g., Quest Diagnostics, Inc. v. United States, 110 Fed. Cl. 716, 726–27 (2013) (upholding agency determination that “meeting solicitation requirements should not be considered a strength”).
But IgniteAction did more than just rephrase the question in the form of an answer. Under Experience Element 1005, offerors were to “describe your specific experience” with, among other things, “tagging.” AR 360. IgniteAction addressed tagging briefly, but it did state that its tagging strategy “track[ed] costs by .” AR 833. That is plainly more information than the question prompt provided, and Intellect does not explain why Census was wrong to have been impressed. Similarly, the Solicitation asked for a “description of the tools.” AR 360. If IgniteAction had responded by saying, for example, that ‘We used tools,’ Intellect’s objection would have been justified. In fact, however, IgniteAction specified which tools it used: and . AR 833. Intellect’s contention is, therefore, factually incorrect. And neither the record nor Intellect supplies a reason to think that Census’s preference for those specific tools is in some way irrational.
Finally, returning to the argument from recency, SES and Intellect argue that Census assigned too much weight to work done by ActioNet under the DOE contract prior to the commencement of the recency window, in June 2019. Intellect mounts the most forceful case for this position, arguing that Census failed to consider that much of the work done under the contract is “stale and outdated,” “particularly in the constantly changing area of cloud services.” Intellect Reply at 21, 28. Although plausible in theory, this argument lacks any factual substantiation. Intellect merely argues that work done several years ago is ipso facto stale: “[F]ast-paced innovation [in the IT industry] means that practices, knowledge, and experience from more [than] five years ago [are] already outdated.” Intellect MJAR at 28. But that sweeping assertion is hardly amenable to meaningful fact-finding. IgniteAction devoted several pages to describing ActioNet’s activities under the DOE contract in detail. AR 824–36, passim.
11 The protestors should have identified precisely what aspect of the work, in their opinion, was “stale” yet blindly relied on by Census. Since they did not, the argument fails.
Intellect
In Census’s technical evaluation of Intellect’s past performance, the agency noted three strengths and two weaknesses. AR 1524–26. Broken down by Experience Elements, the agency assigned neither strengths nor weaknesses as to two elements—1001 and 1002; a mixed picture with two strengths and one weakness for element 1003; a weakness but no strength for element 1004; and a strength but no weakness for element 1005. AR 1524–26. Overall, Census assigned Intellect a “some confidence” rating on Factor 1. AR 1524.
Intellect argues that as to both weaknesses, Census’s explanation runs contrary to the record. As a preliminary matter, the Court notes that Intellect was assigned no strengths for three out of the five Experience Elements: 1001, 1002, and 1004. Thus, Intellect may well not have been upgraded to “high confidence” even if the weaknesses were expunged. In any case, the Court is not persuaded that Census ventured outside its zone of discretion in assigning either weakness rating.
For Experience Element 1003, Census assigned two strengths but also faulted Intellect for “not provid[ing] sufficient detail on the employed cloud security strategies to demonstrate success strengthening the security posture” and instead “focus[ing] on the tools utilized.” AR 1525. Intellect counters that it described its strategies in detail. Intellect MJAR at 15–16. Setting aside the fact that even a lot of detail is not necessarily sufficient detail, the very passages highlighted by Intellect for support, see id. at 16 (quoting AR 848, 852), seem to back up Census’s finding. To take but one example, Intellect picks out the following sentence from its submission: “We implement continuous monitoring and logging using , , , , and .” Id. (quoting AR 852) (emphasis in Intellect’s brief). Respectfully, if that is not focusing on the tools utilized without describing the strategy, the Court does not know what is.
For Experience Element 1004, Census assigned Intellect a weakness on the grounds that Intellect only has experience supporting cloud migration in relation to “[s]ites [that] service approximately internal users and of external users and have a size of [terabytes],” whereas Census’s systems “can include up to 22,000 [internal] users per Sections C.1 and C.4.1 of RFQ and multiple Petabytes of information transferred.” AR 1526.
Intellect argues that Census deviated from the Solicitation because the question prompt for element 1004 asks about the total “user base size,” whereas Census’s demerit focuses on the internal user base size. Intellect MJAR at 17 (quoting AR 767); Intellect Reply at 17–18. Intellect also argues that Census’s reliance on data volumes— terabytes versus multiple petabytes—introduced an unstated evaluation criterion into the agency’s analysis. Intellect MJAR at 17–18. The Court does not agree. The Solicitation clearly states that, under Factor 1, Census will consider which offeror has experience “most similar in scope” to the work contemplated under the BPA. AR 767. Contrary to Intellect’s suggestion, the Solicitation did not—and did not need to—provide an exhaustive list of all facts that Census could find useful
12 when comparing the scope of a past migration effort with Census’s planned migration effort. Indeed, it would be surprising if the Solicitation had stated that Census would disregard the breakdown of the user base or the data volume: Common sense suggests that the largest statistical agency in the federal government would find those factors relevant when determining whether an offeror can support Census’s own migration effort. Thus, Census did not err by relying on the internal user base size and data volume of Intellect’s past migration experience. 8
SES
Census assigned one strength and three weaknesses to SES’s past performance submission and assigned it a “some confidence” rating on Factor 1. AR 1511–13. SES contests all three weaknesses, and further argues that Census should have assigned it more strengths. SES MJAR at 17–26.
SES first takes issue with Census’s assignment of a weakness under Experience Element 1001, which pertains to on-site services. Under that element, Census assigned SES a strength for “experience managing . . . data centers and facilities,” which “far exceeds the footprint of [Census] HQ as well as the multiple regional offices.” AR 1511. However, Census also assigned a weakness for lack of “sufficient detail on methodologies and strategies.” AR 1511. Census then elaborated on the methodologies and strategies to which it was referring. AR 1511.
SES’s counterargument is a legal one. It contends that because the question prompt for element 1001 invited offerors to describe “strategies, methodologies and/or architecture,” AR 1511 (emphasis added), the Solicitation prohibited Census from penalizing SES for describing architecture but not methodologies and strategies, SES MJAR at 18.
That argument is unavailing. This Court must interpret the Solicitation in accordance with its “plain and ordinary meaning.” Banknote, 365 F.3d at 1353. An invitation to speak on one or more of three subjects does not negate the possibility that the listener would ideally like to hear about all three. It was no error for Census to treat nonresponsiveness to a portion of the prompt as a drawback. At best, the word ‘or’ promises that Census could assign a favorable rating to a submission that does not describe strategies, methodologies and architecture. For example, if Intellect had made such a strong case under the architecture prong as to overcome the
8 Intellect also argues that Census disparately evaluated its experience under element 1004, because Census concluded that a contract cited by IgniteAction was “comparable to the size and scope of [Census’s] intended effort” even though it had a small number of users and involved a low data volume. AR1737; Intellect MJAR at 18–19. But Census specifically stated that its assessment traded off on another factor relevant to the scope determination: “Number of Applications.” AR1737. Intellect does not dispute that its own migration experience involved number of applications. Thus, Intellect has failed to show that its migration experience was “substantively indistinguishable or nearly identical” to that of IgniteAction. Off. Design Grp., 951 F.3d at 1372 (internal quotation marks omitted).
Intellect also makes another disparate evaluation argument based on a comparison of its evaluation under element 1004, which addresses cloud services, and IgniteAction’s evaluation under element 1001, which addresses on-site services. That argument falters for the same reason: A comparison across elements, by definition, finds Intellect and IgniteAction in non-identical postures.
13 The first weakness pertains to Functional Area 0001, on-premises operational support. Aside from assigning two strengths, Census assigned a weakness on the grounds that Intellect’s “proposed and management services approach lacked sufficient information to instill confidence in the methodologies or the ‘how’ to the determine likelihood of success.” AR 1548. Census further explained that Intellect’s proposed approach “is currently in place at Census and has proven to be time consuming and cumbersome,” and “inefficient or ineffective.” AR 1548.
Intellect argues that Census assigned the weakness for a lack of detail, and then points the Court to the details it presented. Intellect MJAR at 20–21. But the crux of Census’s evaluation is that Census did not like the approach proposed by Intellect because of past adverse experiences. In the Court’s judgment, even if Intellect were correct that its presentation was “sufficient,” Census would still have assigned it a weakness. Therefore, Intellect’s argument goes nowhere.
Intellect’s second weakness was assigned in Functional Area 0003, integration and migration engineering. Census stated its concern that using containers can be “counterproductive,” and “costly” if done “incorrectly,” and although Census credited Intellect for explaining how it would “deploy[] containers,” it also faulted Intellect for providing “insufficient detail” on “methodologies to instill confidence in the effective of the containers after deployment.” AR 1549–50.
Intellect argues that it provided sufficient detail, pointing the Court to several paragraphs in its proposal. Intellect MJAR at 21 (quoting AR 1178–79). There is no need to reproduce that language; suffice to say that they primarily address deployment rather than steps “after deployment.” In the relevant sections as a whole, Intellect mentioned decommissioning only in passing. AR 1180, 1184. Census’s conclusion about post- deployment methodologies was therefore rational.
Census assigned a rating of “some confidence” to SES’s technical approach after finding nine strengths and seven weaknesses. AR 1544–47. SES correctly notes that all seven weaknesses are based on findings that SES had failed to provide “sufficient detail about particular Functional Areas and subtasks.” SES MJAR at 26. SES argues that, in doing so, Census misread the submission, applied unstated evaluation criteria, failed to document its reasoning, and failed to explain how the lack of detail increased the risk of successful performance. Id.
SES first alleges that Census applied unstated evaluation criteria as to Functional Areas 0001 and 0004. The Court is not persuaded. For example, for Functional Area 0001, on- premises operational support, Census assigned SES a weakness for proposing a security management approach that “focused primarily on maintaining the status quo” but did “offer[] detailed strategic approaches . . . to reduce effort.” AR 1544–45. SES argues that Census applied unstated evaluation criteria while assigning this weakness because security management is not a subtask listed in Functional Area 0001. SES MJAR at 26. This is a befuddling
15 argument. SES elected to discuss security management, highlighting it as a separately titled subpart, in its proposal for that functional area. AR 1279. It cannot now turn around and argue that Census should have disregarded the information. 10 See, e.g., Weston Sols., Inc. v. United States, 95 Fed. Cl. 311, 325 (2010) (consideration of additional information volunteered by offeror not reliance on unstated evaluation criteria). The same is true for SES’s objection as to Functional Area 0004.
Next, SES alleges that for Functional Areas 0001 through 0003, Census erred by faulting SES for failing to provide sufficient detail as to particular subtasks. Rather than go through each objection in detail, the Court simply notes that it has compared the level of detail provided by SES with that by other offerors and with SES’s own answers to other subparts, and concludes that Census’s conclusions are well-justified. The Court also disagrees with SES’s objection that Census failed to document the reasons why the details were not sufficient. Although Census did not explain “precisely what detail was missing” as to each weakness, SES MJAR at 38, the agency did briefly summarize its reasons, AR 1544–47. No more was required. CAN Softtech, Inc. v. United States, No. 24-670, 2024 WL 4434253, at *8 (Fed. Cl. Oct. 4, 2024) (“The streamlined or minimum documentation required by FAR Subpart 8.4 [means that] the Agency did not have to explain in exhaustive detail when it awarded a[n adjectival rating].”) (internal quotation marks and citation omitted). 11
Red Cedar
Red Cedar received a “low confidence” rating for its technical approach after Census found three strengths and sixteen weaknesses. AR 1552–57. Red Cedar takes issue with twelve findings of weakness. The Court will not go into them in detail, because Red Cedar’s arguments are the quintessential example of protestors inviting courts to “micro[]manage the minutiae of a procurement to ferret out technical deficiencies.” Info. Scis. Corp., 73 Fed. Cl. at 104.
Take one example. Red Cedar proposed to use a project management framework called Scrumban. In one finding of weakness, Census commended Red Cedar’s “bold thinking,” but demurred on the grounds that Scrumban “is not well understood.” AR 1786. In its motion, Red Cedar asks this Court to read the Wikipedia article on Scrumban and conclude that Scrumban “is far from not well understood.” Red Cedar MJAR at 16. Red Cedar insists that the Court should participate in the debate because such matters “do not take a degree in computer science to understand.” Red Cedar Reply at 1. Red Cedar has it backwards. Our Court refrains from wading into technical arguments not because they are too difficult to understand, but because our power of judicial review does not include the power to “second-guess the agency's discretionary determinations underlying its technical ratings.” Off. Design Grp., 951 F.3d at 1373. Since all
10 Besides, other offerors’ proposals for the same functional area also discuss security management throughout, see, e.g., AR 1054–57, passim (IgniteAction), indicating that offerors understood that security management was integral to the functional area.
11 SES also complains that Census treated SES and IgniteAction disparately while assigning strengths for Functional Area 0004, cloud security management. SES MJAR at 38–39. SES’s argument is that its discussion of the topic is similar to IgniteAction’s. The Court has gone through the proposals and sees no basis for second- guessing Census’s judgment.
16 - The Quoter shall enter the proposed GSA MAS Labor Category in Column C that corresponds to the BPA Role listed in Column A. The Quoter shall consider the BPA Role Description as a basis for labor category mapping.
- The Quoter shall enter the description of the proposed labor category from their GSA MAS Contract in Column D that corresponds to the GSA Labor Category in Column C.
AR 783 (emphasis added). The protestors further cite the evaluation criteria stating that Census would determine whether offerors “properly understood the price quote instructions” and “properly completed and provided all the required components” outlined in the BPA role- mapping materials. AR 1565. Relying on this language, protestors argue that the Solicitation “committed” Census to “investigat[ing] the qualifications for [all proposed] LCATs [and] whether those personnel could accomplish the Agency’s requirements.” Intellect MJAR at 11.
The Court is not persuaded. The excerpts cited from the Solicitation are clearly a requirement to ensure that offerors dot their i’s and cross their t’s. Census did that thoroughly, checking IgniteAction’s price submission for “completeness and accuracy” by “conduct[ing] a[] . . . line-by-line review of,” among other things, the “BPA Rate Card by resource (BPA Role and Labor Category).” AR 1567. The Court can see that the review was not perfunctory, because at one point Census was “unable to map” some proposed LCATs to the publicly available GSA MAS Price List and reached out to the GSA for clarification. AR 1567. The GSA then acknowledged that the data on its website (the GSA eLibrary) was “not the most current.” AR 1567. That is competent, thorough work.
As for whether IgniteAction “consider[ed]” role descriptions while mapping BPA roles to LCATs, protestors have not pointed to even one instance where IgniteAction’s mapping is such a poor fit as to reflect a lack of consideration of role descriptions. See Consider, Merriam- Webster's Collegiate Dictionary (11th ed. 2003) (“to think of especially with regard to taking some action; to take into account”). If a recruiter recommended a Level 2 IT systems administrator for the role of a senior-level backup/recovery engineer, she could truthfully say that she had “considered” the posted role description.
Protestors’ contention that more was needed by the FAR itself, proves too much. In the context of a price evaluation, requiring a heightened correlation 12 between BPA role descriptions and LCAT job descriptions amounts to a price realism analysis. See FAR 2.101 (defining a realism analysis to include whether the proposed costs are “realistic for the work to be performed” and “[r]eflect a clear understanding of the requirements.”). But a price realism analysis is optional, see Harmonia Holdings Grp., LLC v. United States, 167 Fed. Cl. 448, 473 (2023) (“A price realism analysis is performed only if the solicitation requires one.”), and Census explicitly stated that it “will not consider Price Realism,” AR 648. Given the clarification by
12 To be clear, an agency must always “consider[] . . . the mix of labor” proposed by an offeror. FAR 8.405- 3(b)(2)(vi). In this case, Census’s thorough line-by-line analysis meets that standard. To wit, the Court is convinced that Census would have caught any egregious or embarrassing discrepancies between BPA roles and proposed LCATs job descriptions.
18 Census, protestors faced at best a patent ambiguity, which should have been litigated pre-award. Blue & Gold Fleet, L.P. v. United States, 492 F.3d 1308, 1313 (Fed. Cir. 2007); Kearney & Co., P.C. v. United States, No. 24-162, 2024 WL 2209767, at *6 (Fed. Cl. Apr. 30, 2024) (“To the extent that any party had understood the RFQ to require a heightened correlation, or an exact match, between the PWS work and the offeror's LCATs . . . that would have been a patent ambiguity given [that the Solicitation] did not limit the ability to map as normal.”).
In sum, neither the Solicitation nor the FAR require a price realism analysis, yet that is precisely what protestors desire. That ends the matter. 13
D. Census’s Best-Value Determination
Finally, protestors argue that Census did not conduct an adequate best-value determination, because Census’s Source Selection Authority (“SSA”) blindly relied on the adjectival ratings and failed to exercise her independent judgment. Intellect MJAR at 29–30.
The Court agrees that the SSA must independently assess the evidence in every case, but disagrees that she failed to do so here. The SSA may not rely solely on adjectival ratings. Mil- Mar Century, 111 Fed. Cl. at 553 (interpreting FAR 15.308). That said, the purpose of looking beyond adjectival ratings is not paperwork for paperwork’s sake, but to ensure that proposals “with the same adjectival ratings” but “[un]equal quality” are not treated as identical. Femme Comp Inc. v. United States, 83 Fed. Cl. 704, 758 (2008) (internal quotation marks and alterations omitted). Therefore, where (as here) there is no real tradeoff based on the adjectival ratings, the Court’s only job is to check if the SSA made the mistake of treating two parties with the same adjectival ratings as perfectly equivalent.
The SSA did in fact look beyond the adjectival ratings. To be sure, she did state the following:
IgniteAction’s total evaluated price to the Government was the lowest in comparison to the three (3) competing submissions. Looking at both technical factors, IgniteAction is the highest technically rated quoter whom instilled the greatest degree of confidence at the lowest overall price. There are no superior non- price benefits to be offered by another Quoter to warrant an award at a significantly higher overall price. Following the basis of award in the RFQ, there is nothing additional to compare in my best value determination and no tradeoffs to be made.
AR1709 (emphasis added). However, the SSA also compared offerors with the same adjectival rating, explicitly noting strengths and weaknesses associated with each offeror. For example,
13 Protestors also argue that IgniteAction did not propose its own LCATs, but that of the mentor partner. The Court finds no basis for that assertion. All the LCATs proposed by IgniteAction are on its own GSA MAS contract; that they might also be on the mentor partner’s GSA MAS contract is irrelevant. To the extent that protestors argue that IgniteAction intends to violate the requirement that “the protégé partner [perform] at least 40% of the total [work] done by the partners,” 13 C.F.R. § 125.8(c)(3), protestors have failed to provide any supporting evidence.
19 recall that even though Census had “high confidence” in both IgniteAction’s and Intellect’s technical approach, Census identified weaknesses in Intellect’s approach but not in IgniteAction’s. Far from glossing over that fact, the SSA’s decisional document explicitly notes those weaknesses. Compare AR 1708 (“no notable aspects which decreased the Government’s confidence” in IgniteAction’s technical approach) with AR 1709 (“two noted aspects which reduced the Government’s confidence” in Intellect’s technical approach). Thus, the SSA assured herself that the adjectival ratings did not hide relevant nuances. Her conclusion is sound and it must stand.
IV. Conclusion
As indicated in a prior Order, and for the reasons explained in this opinion, the protestors’ MOTIONS for Judgment on the Administrative Record, ECF Nos. 41, 44 & 45, are DENIED. Defendant and defendant-intervenor’s CROSS-MOTIONS for Judgment on the Administrative Record, ECF Nos. 48 & 49, are GRANTED.
IT IS SO ORDERED.
s/ Loren A. Smith Loren A. Smith, Senior Judge