OPINION AND ORDER
WOLSKI, Judge.This post-award bid protest was before the Court on the plaintiffs motion for a preliminary injunction. The plaintiff, University Research Co. (“URC”), requested that the Court enjoin defendant United States (“Government”) from initiating the transition to, and from continuing the performance of, the contract awarded to IQ Solutions, Inc. (“IQS”) under RFP No. 277-04-6091 (“Solicitation”). University Research Co. argued that the Government’s award of the contract was unreasonable, challenging the assessment of past performance and the cost realism analysis as arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law. The Government and IQS opposed the granting of this motion. The Court issued a preliminary injunction on May 6, 2005. This opinion explains that decision.
I. BACKGROUND
The plaintiff is the unsuccessful bidder for a contract awarded under a request for proposals issued by the Department of Health and Human Services, Substance Abuse and Mental Health Services Administration (“SAMHSA”). The Solicitation was for the operation of a clearinghouse for the public to acquire information and referrals about substance abuse and mental health issues from the federal government. Admin. R. 473, 545-47. The successful contractor will be responsible for providing information by telephone, internet, and mail to members of the public. Id.; see also id. at 546-77 (Performance Work Statement). Currently, URC provides these clearinghouse services for the incumbent prime contractor, Social and Health Services Ltd. Id. at 2321. The Solicitation has been the subject of two bid protests before the Government Accountability Office (“GAO”).I. 2 University Research Co., B-294358.6, 2005 U.S. Comp. Gen. LEXIS 73, 2005 WL 1026055 (2005); University Research Co., B-294358, 2004 CPD 1Í 217, 2004 WL 2496439 (2004). Moreover, this Court, on August 6, 2004, issued a preliminary injunction enjoining the agency from overriding the automatic stay provision of the Competition in Contracting Act (“CICA”), 31 U.S.C. § 3553(d)(3)(A), pending a determination by the GAO of the merits of the plaintiffs protest.
A. The Solicitation
The Solicitation, issued on December 19, 2003, anticipated the award of a one year cost-plus-award-fee contract with four one-year options. Admin. R. 469,473, 490. The purpose of this procurement was to realign SAMHSA’s health information dissemination programs under what amounts to a single roof. Id. at 473, 546. Currently, the services covered by the Solicitation are dispersed across multiple contracts. Id. Included in the procurement is the operation of the National Clearinghouse for Alcohol and Drug Information (“NCADI”) and the National Mental Health Information Clearinghouse (“NMHIC”). Id. at 473, 545-46. The NCA-DI provides information services related to substance abuse and is the hub of the federal government’s efforts to gather and communicate information about effective prevention, intervention, and treatment policies, programs, and practices. Id. The NMHIC is a source of information and referrals, to users of mental health services and their families, service providers, policy makers, the media, and the general public. Id.
B. The Source Selection Determinations
The agency’s original Source Selection Determination was made on June 30, 2004, by the Contracting Officer (“CO”). Admin. R. 387-93. The CO determined that of the six proposals submitted, three were in the competitive range: the proposals from Dan-ya/JBS Joint Venture (“Danya”), IQS, and *502URC. Id. at 388. The offerors in the competitive range were given a technical review score on a 100 point scale, a past performance score on a 36 point scale, and a total score that combined the technical and past performance scores. Id. at 388-89. Danya received a technical score of 83.8, and a past performance score of 28.9, for a total score of 112.7. Id. at 389. IQS received a technical score of 93.3, and a past performance score of 31.2, for a total score of 124.5. Id. And, URC received a technical score of 88.7, and a past performance score of 29.8, for a total score of 118.5. Id. The proposals were also evaluated for cost. The final revised costs for the offerors were as follows: Danya, $85,130,845; IQS, $57,122,677; URC, $63,148,266. Id. After considering the evaluation scores and proposed costs, the CO determined that “IQ Solutions is technically excellent and technically superior to other offerors, offers reasonable and realistic estimated costs, has low performance risk, and offers the best overall value to the Government.” Id. at 392.
The second Source Selection Determination was made on December 22, 2004, by a new CO, after URC’s first bid protest before the GAO was sustained. Admin. R. 2756-70. In the second determination, only IQS and URC were considered. See id. at 2759. The second determination again considered the offerors’ technical and past performance evaluations and conducted a cost realism analysis of the offerors’ proposed costs. IQS received a technical score of 94.4 and a past performance score of 33.12, and its final evaluated cost was $61,130,093. Id. at 2769. URC received a technical score of 92.8 and a past performance score of 32.33, and its final evaluated cost was $67,274,293. Id. at 2769. The CO concluded that “[a]lthough technical considerations were paramount in this acquisition, since the technical scores were almost equal with IQ slightly higher, IQ is 10.05% lower in cost or $6,144,200 (using cost realism), and it is recommended that the award be made to IQ as the best value to the government.” Id.
C. Procedural History
On July 19, 2004, the plaintiff filed an application for a temporary restraining order and a motion for a preliminary injunction in this Court. At that time URC sought to enjoin the Government from continuing performance of the award of the contract to IQS pending the receipt of a mandatory debriefing from the contracting officer and an opportunity to file a bid protest in the GAO. Before the Court could rule on URC’s application and motion, URC received the debriefing and filed a bid protest in the GAO on July 21, 2004, triggering the CICA automatic stay provision.3 On August 2, 2004, however, the Government decided, pursuant to 31 U.S.C. § 3553(d)(3)(C)(i)(I), to override the automatic stay.4
Accordingly, URC again filed an application for a temporary restraining order and a motion for a preliminary injunction in this Court seeking to enjoin the Government, pursuant to RAMCOR Services Group, Inc. v. United States, 185 F.3d 1286, 1291 (Fed. Cir.1999),5 from overriding the automatic *503stay, so that the GAO could consider its bid protest. In opposition to the plaintiffs application and motion, the Government argued that: 1) the courts are without authority to review an agency’s determination that the override of an automatic stay is in the best interest of the government, see Tr. (Aug. 3, 2004) at 39:16-24; and 2) even if the courts have the power of review, the best interests of the United States were served by the override because the contract awarded to IQS was less costly and a better value to the government than the contract that would remain in effect during the pendency of the automatic stay, see id. at 47:3-17.
The Court granted the plaintiffs request for a temporary restraining order, which the Court then converted into a preliminary injunction on August 6, 2004. The Court gave the reasons for its ruling on the record at the conclusion of oral argument. Id. at 131-139. The Court held that it has jurisdiction to review an agency’s decision to override a CICA automatic stay under 28 U.S.C. 1491(b)(1). RAMCOR, 185 F.3d at 1291; Altos Fed. Group, Inc. v. United States, 60 Fed.Cl. 832, 833 (2004); Keeton Corrections, Inc. v. United States, 59 Fed.Cl. 753, 755 (2004); PGBA LLC v. United States, 57 Fed.Cl. 655 (2003). Further, the Court held, following PGBA 57 Fed.Cl. at 659, that override decisions under the “best interests” prong of section 3553 may be reviewed by this Court under 28 U.S.C. § 1491(b).6
This Court reviews an agency’s decision to override an automatic stay under the arbitrary, capricious, abuse of discretion, or otherwise not in accordance with law standard incorporated in 28 U.S.C. § 1491(b)(4). See, e.g., RAMCOR, 185 F.3d at 1290; PGBA 57 Fed.Cl. at 657. The Court reasoned that if a determination by a contracting officer that one bidder’s proposal represented the best value to the government were sufficient for overriding the automatic stay, as a practical matter, the automatic stay would be meaningless in virtually every single instance in which a GAO protest was filed. Tr. (Aug. 3, 2004) at 132-33. Almost invariably when a contract is awarded through a competitive process, the award will be based on the awardee offering the government the best value, and in many cases this also means the contract will cost the government less money. Accordingly, the reasons the agency, gave for overriding the automatic stay were unreasonable. See PGBA 57 Fed.CI. at 662 (“the court does not believe that the mere fact that the new contracts would be better than the old is enough to support the override decision under either prong of section 3553”).
After the Court enjoined the override of the automatic stay, the plaintiff filed two bid protests with the GAO. The GAO sustained URC’s first protest because “the selection official failed to state any basis for rejecting the award recommendation of agency project officers, and without such documentation, [there is] no basis for determining whether her actions were reasonable.” 2004 CPD H 217. The GAO decision stated that it “reached this conclusion after the selection official testified in a hearing before our Office that she knowingly mischaracterized ... the recommendation of agency project officers.” Id. The GAO did not, however, sustain the plaintiffs second protest, even though it found errors in the agency’s procurement decision, as the errors were found to be immaterial. 2005 U.S. Comp. Gen. LEXIS 73, 2005 WL 1026055. Although GAO decisions are persuasive on this Court, they are not binding. See, e.g., Overstreet Elec. Co. v. United States, 47 Fed.Cl. 728, 733 n. 7 (2000); CACI Field Servs. v. United States, 13 Cl.Ct. 718, 731 n. 28 (1987).
D. URC’s Grounds for Protest
The plaintiff asserts two grounds for protesting the award of the contract to IQS. *504First, URC asserts that the agency failed to conduct a proper past performance evaluation. Second, the plaintiff protests that the agency improperly normalized the cost of photocopying. With regard to both of its grounds for protest, URC argues that the agency acted in a manner that was arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law.
The plaintiff argues there were two problems with the agency’s past performance evaluation. University Research Co.’s first argument is that the agéncy used a flawed method to average past performance scores and, as a result, assigned arbitrary and irrational weights to the offerors’ past performance scores. PL’s Mem. at 4. The plaintiffs second claim is that the agency failed to distinguish between degrees of relevance within each offeror’s past performance. Id. In other words, URC contends that the agency should have given greater weight to those particular previous government contracts that are the most relevant to the services required by the Solicitation.
Additionally, URC argues that the agency acted in an unreasonable manner when it normalized the costs of making photocopies. PL’s Mem. at 20-35. The plaintiff contends that by normalizing reproduction costs the agency ignored real differences between the proposals. According to URC, the agency normalized production costs because the agency believed its initial estimate of the number of copies required by the Solicitation was too low and because IQS’s proposal was unclear as to copying costs. The plaintiff argues that both of these reasons for normalization are unreasonable, and asserts that but for the agency’s erroneous normalization, it would have had the lowest evaluated cost and thus would have been awarded the contract.
II. DISCUSSION
This Court has jurisdiction over post-award bid protests pursuant to the Tucker Act, 28 U.S.C. § 1491, as amended by the Administrative Dispute Resolution Act of 1996 (“ADRA”), Pub.L. No. 104-320, 110 Stat. 3870 (1996). The Tucker Act grants the Court jurisdiction over bid protests brought 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.” 28 U.S.C. § 1491(b)(1). The Court may issue a preliminary injunction pursuant to Rule 56 of the Rules of the Court of Federal Claims (“RCFC”) and 28 U.S.C.. § 1491(b)(2). In determining whether to grant a motion for a preliminary injunction the Court applies a four-part standard. Under this standard a plaintiff must show: 1) that it will suffer irreparable injury if the procurement is not enjoined; 2) a reasonable likelihood of success on the merits of its claim; 3) that the harm suffered by it, if the procurement is not enjoined, will outweigh the harm to the Government and third parties; and 4) that granting injunctive relief serves the public interest. FMC Corp. v. United States, 3 F.3d 424, 427 (Fed.Cir.1993); Chrysler Motors Corp. v. Auto Body Panels, Inc., 908 F.2d 951, 953 (Fed.Cir.1990); Zenith Radio Corp. v. United States, 710 F.2d 806, 809 (Fed.Cir. 1983). None of the four factors, standing alone, is dispositive; thus, “the weakness of the showing regarding one factor may be overborne by the strength of the others.” Chrysler Motors, 908 F.2d at 953; FMC Corp., 3 F.3d at 427. Conversely, the lack of an “adequate showing with regard to any one factor may be sufficient, given the weight or lack of it assigned the other factors,” to deny the injunction. Chrysler Motors, 908 F.2d at 953.
A. Is URC Reasonably Likely to Succeed on the Merits?
The Court will first consider the likelihood of success factor, which is primarily a question of law. Under the likelihood of success prong of the Court’s preliminary injunction standard, and in light of the other three prongs, the plaintiff must demonstrate a reasonable likelihood of success on at least one of its two grounds for protesting the award of the contract to IQS. Thus, the plaintiff must show a sufficient strength in either its past performance evaluation claim or its cost normalization claim, on the merits. When this Court reviews an agency’s decision in a bid protest, the Court must use the standards set forth in the Administrative Proce*505dure Act (“APA”), 5 U.S.C. § 706. 28 U.S.C. § 1491(b)(4). Accordingly, a protestor must show that the agency’s decision was arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law. See 5 U.S.C. § 706(2)(A). Moreover, if the protestor is able to show an error in the procurement process, it must also show that “the error prejudiced it.” Data General Corp. v. Johnson, 78 F.3d 1556, 1562 (Fed.Cir.1996). In order for a protestor to establish that it was prejudiced by an error, it must demonstrate “that there was a substantial chance it would have received the contract award but for that error.” Statistica Inc. v. Christopher, 102 F.3d 1577, 1581 (Fed.Cir.1996).
1. Technical Evaluation
Before considering the two grounds for protest, the Court will address briefly one ground raised before the GAO, but not here, concerning the technical evaluation process. After an initial review, three of the six offer-ors fell within the competitive range — URC, IQS, and Danya. Admin. R. 445. The technical review committee had performed technical evaluations of all of the bidders, and Danya had the best score — 87 out of 100. Plaintiff URC was second, at 81.8, and inter-venor IQS third, at 79.86. Admin. R. 442. After negotiations were held with these three offerors, revised proposals were submitted and evaluated. This time, IQS earned the highest technical score, a 93.3. Plaintiff received an 88.7, and Danya an 83.8. Admin. R. 389. The difference in scores between URC and IQS was entirely attributable to the scores given by one of the six evaluators. This evaluator gave IQS 99 points, and URC only 68 points. Compare Admin. R. 1007 with Admin. R. 2121. If these scores are omitted, then URC would grade slightly higher than IQS in this category — 92.8 versus 92.2. No other reviewer’s scoring made a difference in the relative ranking of the two.
After the GAO’s initial decision, the agency’s December 22, 2004 Source Selection Determination addressed the alleged bias of this individual (identified therein as “Reviewer # 3,” Admin. R. 2766) by dropping the lowest technical score for both URC and IQS. This resulted in IQS’s margin over URC being reduced to 1.6 points — based on respective scores of 94.4 and 92.8. Admin. R. 2766. The source selection authority thus concluded that “the technical scores were almost equal with IQ slightly higher.” Id. 2769; see also id. 2781 (debriefing letter to URC) (contracting officer states “the technical scores as well as past performance scores were considered equal”).
In the second protest, the GAO determined that this adjustment was not rational, as both scores from the allegedly partial evaluator should have been dropped, rather than the low score for each offeror. 2005 U.S. Comp. Gen LEXIS 73 at *27, 2005 WL 1026055 at *9. This has not been raised as an issue in the current protest before the Court. In any event, the Court recognizes that the agency’s judgment was that the technical scores, whether they result in an edge to IQ Solutions of 1.6 points or to URC of 0.6 of a point, are basically equal for purposes of comparing the two. See Admin. R. 2769, 2781.
2. Past Performance
When the Court considers a bid protest challenge to the past performance evaluation conducted in the course of a negotiated procurement, “the greatest deference possible is given to the agency.” Gulf Group Inc. v. United States, 61 Fed.Cl. 338, 351 (2004); see, e.g., E.W. Bliss Co. v. United States, 77 F.3d 445, 449 (Fed.Cir.1996) (“Procurement officials have substantial discretion to determine which proposal represents the best value for the government.”) (citing Lockheed Missiles & Space Co. v. Bentsen, 4 F.3d 955, 958 (Fed.Cir.1993)); Overstreet Electric Co. v. United States, 59 Fed.Cl. 99, 117 (2003) (stating that “when a procurement involves performance standards ... a court must grant even more deference to the evaluator’s decision ... a triple whammy of deference”); id. at 102 (characterizing the standard of review as “near draconian”); Computer Sciences Corp. v. United States, 51 Fed.Cl. 297, 319 (2002) (holding that “an agency is accorded broad discretion when conducting its past performance evaluations”); SDS Int’l v. United States, 48 Fed.Cl. 759, 769 (2001) (same).
*506The deference the Court gives an agency in performing a past performance evaluation is, however, not without limit, as “it is settled law that past performance evaluations are subject to the same APA review as other agency actions challenged in this court on a bid protest.” CSE Constr. Co. v. United States, 58 Fed.Cl. 230, 252 (2003); Seattle Sec. Servs., Inc. v. United States, 45 Fed.Cl. 560, 567 (2000) (stating that “bound by the arbitrary and capricious standard of review, the exercise of this discretion obviously must be reasonable”). Accordingly, the Court’s review of an agency’s “evaluations of an offeror’s ... past performance ... should be limited to determining whether the evaluation was reasonable, consistent with the stated evaluation criteria and complied with relevant statutory and regulatory requirements.” JWK Int’l Corp. v. United States, 52 Fed.Cl. 650, 659 (2002).
The plaintiff has advanced two arguments regarding the agency’s past performance evaluation. The first of the plaintiffs arguments, as the Government concedes7 and the GAO recognized,8 is correct — that is, the agency acted in an irrational manner when it averaged the scores in the manner it did. The plaintiffs second argument — that the most relevant scores should have received more weight — is, however, without basis. There is nothing in the Solicitation or the regulations cited by the plaintiff that requires more relevant past performance receive greater weight in the past performance evaluation.
a. The methodology for averaging past performance scores.
Turning to URC’s first argument, it is clear that the averaging method used by the agency was irrational. The Solicitation called for past performance to be rated on a thirty-six point scale. The past performance scores that the agency had for the offerors were on twenty and twenty-five point scales, and thus needed to be converted to the Solicitation’s required thirty-six point scale. But the spreadsheet the agency used to make the conversions led to an error in the final past performance score each offeror was given. The spreadsheet contained two rows for each offeror: one row for those contracts that were scored on a twenty-five point scale and one row for those contracts that were scored on a twenty point scale.9 In order to place the scores on a thirty-six point scale, the agency multiplied those scores that were on a twenty-five point scale10 by 1.44, and those scores that were on a twenty-point scale by 1.8. Admin. R. 2761.
Instead of first adjusting each individual score, adding these together and dividing by the total number of scores to determine an offeror’s average score, the agency did something rather odd: It determined the average score for each row based on the unadjusted scores, adjusted these averages to fit a thirty-six point scale, and then took the average of these averages to determine each offeror’s score. Id. Thus, for instance, IQS, which only had one score weighted on a twenty-five point scale, had that score count for half of their past performance evaluation score, with the five scores it received on a twenty-point scale counting for the other half. Id. Calculating past performance scores in such a manner has no basis in logic or reason and, therefore, this ground for protest would assuredly be sustained on the merits were it to be shown prejudicial to the plaintiff. The *507GAO’s correction of these averaging errors, in which the parties apparently acquiesced, resulted in an increase in URC’s past performance score from 32.33 to 33.12, and a reduction in IQS’s score from 33.12 to 32.16. See 2005 U.S. Comp. Gen. LEXIS 73 at *11, *30, 2005 WL 1026055 at *4, *10. Instead of IQS having an edge of 0.79 of a point, URC should have had a 0.96 of a point advantage in this category. The Court observes that, as was the ease with the “paramount” technical scores, either difference was so close to zero that the scores were essentially equal, and this adjustment alone would not be the basis for changing the outcome of the award.
b. The failure to give extra weight based on greater relevance.
For URC’s past performance challenge to make a difference, it must result in a much larger difference in final, average scores than the correction of the averaging error produces. Prejudice to the plaintiff would thus turn on its second argument for this ground. But there is no basis for URC’s argument that both the Solicitation and regulations require that more relevant past performance receive greater weight in the evaluation. Nothing in either the Solicitation or the regulations cited by the plaintiff requires this. An agency could, of course, choose to give greater weight to contracts it found to be more relevant than others, but the agency here did not so choose. See Forestry Surveys & Data v. United States, 44 Fed.Cl. 493, 499 (1999) (“An agency, in evaluating past performance, can give more weight to one contract over another if it is more relevant to an offeror’s future performance on the solicited contract.”).
The Solicitation contained two sections related to the past performance evaluation: section L and section M. Section L contains instructions to offerors, who were required to submit a “list of the last 5 contracts completed during the past THREE years and ALL CONTRACTS currently in process that are similar in nature to the solicitation work scope.” Admin. R. 532. It was then explained that:
Each offeror will be evaluated on its performance under existing and prior contracts for similar products or services. Performance information will be used for both responsibility determinations and as an evaluation factor against which offeror’s relative ranking will be compared to assure the best value to the Government. The Government will focus on information that demonstrates quality of performance relative to the size and complexity of the acquisition under consideration.
Id. (emphasis added).
Section M provides the evaluation factors for the contract award. It begins with the pronouncement that “[sjelection of an offeror for contract award will be based on an evaluation of proposals against four [sic] factors. The factors are as follows: technical, past performance and cost.” Admin. R. 540. The past performance subsection of Section M states that “[i]n evaluating past performance, the Government will consider the offer- or’s effectiveness in quality or product or services; timeliness or [sic] performance; cost control, business practices; customer (end user) satisfaction; and key personnel past performance.” Admin. R. 543. It further noted that “[p]ast performance will receive 36 percent of the non-cost/price ratings.” Admin. R. 543.
The argument of URC, boiled down to its essence, is that in reviewing past performance for purposes of awarding a clearinghouse contract, the government must give the greatest weight to evaluations relating to clearinghouse contracts. See Pl.’s Mem. at 12-16. This is exactly the sort of decision that is best left to the discretion of the agency procurement officials, however. As a general proposition, it cannot be said that the best way to award a contract under a solicitation for services ranging from A to Z, is to place particular emphasis on past performance of contracts for services that also ranged fully from A to Z. The incumbent performing such a contract might think that way is the best, but an agency could rationally decide that contracts combining various of the services to be hired — perhaps A to G in one, H to K in another, L to W in a third, etc. — could also demonstrate the ability to do the work at issue, with the added virtue of expanding the pool of potential offerors. Of *508course, the agency could instruct the offerors that it will weight past performance based on degrees of relevance, and would be bound by such instructions.
But here, all the agency asked for was information on contracts “for similar products or services,” not identical ones. Admin. R. 532 (emphasis added). In its brief and at oral argument URC asserted vigorously that the language from Section L supported its argument. It focused on the statement that “[t]he Government will focus on information that demonstrates quality of performance relative to the size and complexity of the acquisition under consideration.” See, e.g., Pl.’s Mem. at 8; Tr. (May 5, 2005) at 58:11-14 (emphasis added). But this statement does not establish a methodology in which the similarity of work performed under past contracts will be the agency’s focus. Rather, for contracts meeting the threshold of being “for similar products or services,” the agency stated that its focus would be on other factors contributing to relevance: size and complexity. Complexity deals with how intricate the services provided in a contract were, or how difficult the task called for by a contract was to complete. And size obviously refers to how large previous contracts were. Neither term is a proxy for the functional similarity of products or services. The sections of the Solicitation that describe the past performance evaluation process do not require the agency to rank past contracts based on relative degrees of relevant products or services — nor, for that matter, do they require such a ranking even concerning the size or complexity of contracts.
The same can be said of the section of the FAR cited by the plaintiff. The plaintiff cites two clauses in FAR § 15.305 for the proposition that more relevant past performance must be given greater weight in the past performance evaluation: 48 C.F.R. § 15.305(a)(2)(i) and § 15.305(a)(2)(ii). Section 15.305(a)(2)(i) states:
Past performance information is one indicator of an offeror’s ability to perform the contract successfully. The currency and relevance of the information, source of the information, context of the data, and general trends in contractor’s performance shall be considered. This comparative assessment of past performance information is separate from the responsibility determination required under subpart 9.1.
48 C.F.R. § 15.305(a)(2)(i). Section 15.305(a)(2)(ii) states, in pertinent part, that “[t]he source selection authority shall determine the relevance of similar past performance information.” 48 C.F.R. § 15.305(a)(2)(ii).
Although the clauses upon which plaintiff relies deal with relevance of past performance, the language cited by the plaintiff the does not support the expansive interpretation plaintiff urges upon the Court. Quite simply, the FAR does not require the agency to distinguish between relative degrees of relevance. The agency is required to, among other things, consider “the relevance of similar past performance information”; however, this obligation was met through the agency’s determination that the past performance that was analyzed met a threshold level of relevance. All that is required is the binary choice of yes/no regarding relevance. The approach urged by URC is, to be sure, one way the agency could have met its FAR obligation to consider the relevance of similar past performance, but it is certainly not the only way. The agency’s decision on how to conduct the past performance analysis was in accordance with the FAR; therefore, the Court will defer to the judgment of the agency. See Gulf Group, 61 Fed.Cl. at 353 (“As long as these individuals considered the relevant factors and information and articulated a reason for their decision that is not clearly erroneous, the Court will defer to the judgment of the Executive branch.”).
In sum, despite plaintiffs insistence that both the Solicitation and the FAR require that past performance be weighted based on the degree of relevance of contracts, nothing in either supports this proposition. The provisions in the Solicitation and the FAR instruct that relevant as opposed to irrelevant past performance is to be considered, but no ranking of relative relevance is mandated. The plaintiff may still argue that the agency included irrelevant, or excluded relevant, past performance in the offerors’ past performance evaluations, and has identified one *509of its past contracts that it believes is irrelevant. See Tr. (May 5, 2005) at 45-47, 63-64; Admin. R. 2598-2602. But the erroneous inclusion of just one past performance score is a rather thin reed upon which to rest an argument of prejudice, and does not appear to have made a difference to the Source Selection Determination.11 At least for purposes of this motion, the agency’s past performance evaluation appears to have been “reasonable, consistent with the stated evaluation criteria and [in] compliance] with relevant statutory and regulatory requirements.” JWK Int'l Corp., 52 Fed.Cl. at 659. Accordingly, it is unlikely that the plaintiff would succeed on the merits of its claim that the past performance evaluation was, to its prejudice, improper.
3. Cost Normalization
The contract at issue in this ease is a “cost reimbursement” contract, under which the government is required to pay the actual and allowable costs incurred regardless of the proposed or estimated costs set forth in the contractor’s bid. Because this was a cost reimbursement contract, the agency performed a cost realism analysis “to determine what the probable cost of performance [would] be for each offeror.” JWK Intl Corp. v. United States, 49 Fed.Cl. 371, 393 (2001) (citing Day & Zimmermann Sens. v. United States, 38 Fed.Cl. 591, 597-98 (1997)).12 Under certain circumstances it is appropriate, following a cost realism analysis, to normalize particular costs. Normalization is the practice by which the agency substitutes a set or “baseline” price that the agency has determined is appropriate, in place of the prices proposed by the offerors; in other words the agency substitutes a “common should have bid estimate.” Computer Sciences Corp. v. United States, 51 Fed.Cl. 297, 316 (2002) (internal citations omitted). This Court and the GAO have recognized that normalization is appropriate when there is no logical basis for differences in approach, or there is insufficient information available in the proposals. See, e.g., id. at 316; General Research Corp., B-241569, 70 Comp. Gen. 279, 1991 WL 72853 (1991). Thus, in general, the purpose of the cost realism analysis “is to segregate cost factors which are ‘company unique’ — depending on variables resulting from dissimilar company policies — from those which are generally applicable to all offerors and therefore subject to normalization.” Computer Sciences Corp., 51 Fed.Cl. at 316 (citing SGT, Inc., B-281773,1999 U.S. Comp. Gen. LEXIS 138, *15 (1999)).
In this case the agency decided to normalize several costs, including copying costs. The plaintiff seeks to overturn only the normalization of reproduction costs. As this Court has recognized, “[i]n order to overturn a cost realism decision plaintiff must demonstrate that the choice made by the agency was irrational.” JWK Intl, 49 Fed.Cl. at 393; CTA Inc. v. United States, 44 Fed.Cl. 684, 693 (1999) (“Decisions on cost realism are within the agency’s sound discretion and expertise, and the judgment will not be overturned absent any rational basis.”). While an agency’s cost realism analysis does not have to be performed with “impeccable rig- or” in order to be rational, OMV Medical, Inc. v. United States, 219 F.3d 1337, 1344 (Fed.Cir.2000), the analysis must nonetheless reflect that the agency took into account the information available and did not make “irrational assumptions or critical miscalculations.” Id.
*510 a. The government’s treatment of reproduction costs.
While the RFP identified two areas of costs that were to be considered the same for all offerors, see Admin. R. 569 ($1.5 million per year for special initiative projects); id. 572 ($1.8 million to $2 million per year in shipping costs), reproduction costs were not one of these. The latter were to be estimated by offerors based on the requirement that they “[rjeproduce (black and white-60 percent and color-40 percent) approximately 1.6 million pieces (single sheets) each month during the first year of operation, with a 15-20 percent increase for each year thereafter.” Admin. R. 573.
At the time the initial offers were being evaluated, the Government was working from an Independent Government Cost Estimate (“IGCE”) that calculated annual reproduction costs to be $[XXXXX]. See Admin. R. 428-29, 436, 440. Over a five-year period, this would amount to $[XXXXX] in reproduction costs. Intervenor had initially proposed these to be $[XXXXX], see Admin. R. 1128, but somehow the Government’s cost analyst totaled these up as $[XXXXX], see id. 429. Using this higher figure, the Government told IQS to “[pjlease address these costs in accordance with the SOW (p.29) and correct accordingly. Please reduce by 50%.” Admin. R. 1015. Reducing the Government’s version of IQS’s reproduction costs by fifty percent would have lowered this figure to $[XXXXX], But IQS went even lower, incorporating copying costs of $[XXXXX] in its revised final proposal — cutting in half its own numbers, and not the Government’s version of them. Admin. R. 931.
Intervenor responded to the Government’s request that it cut copying costs in half by quoting from the referenced page of the RFP and adding the following:
IQ Solutions prepared our initial estimate based on the quantities stipulated in the underlying budget assumptions for reproduction. We are assuming that your guidance to reduce our estimate by 50 percent implies that a portion of the stipulated copies will be produced by GPO in accordance with the requirements mandated by the Government Printing and Binding Regulations (Section H. 16 of the RFP on p. 29 of 74) and the historical data of the incumbent contractor. We have therefore reduced our estimate accordingly. IQ Solutions is prepared to make further adjustments should the Government direct us to do so. [H] As you directed, we have reduced our costs by 50%.
Admin. R. 823.13 IQ Solutions cut each of its annual copying costs estimates in half, but even more dramatically changed its per page cost estimates. Black-and-white copies went from $[XXXXX] per page to $[XXXXX] per page, almost six times the initial estimate. Compare Admin. R. 1128 with id. 931. The per unit cost of color copies tripled, from $[XXXXX] per page to $[XXXXX] per page. Id.
The reproduction costs initially proposed by URC, $[XXXXX], see Admin. R. 2344, were slightly higher than both the IGCE the Government had been working with, and the figure that IQS was asked to propose. But the Government nevertheless told URC to raise its reproduction costs by sixteen percent. Id. 2128. This appears to be based on another error by the Government, as the “Pre-Negotiation Plan” prepared by the cost analyst mistakenly calculated the five-year reproduction cost target to be $[XXXXX], despite the one-year IGCE being $[XXXXX]. Compare Admin. R. 438 with id. 436.14 In its revised final proposal, URC stuck to its proposed reproduction cost figure, Admin. R.2007, and explained this decision:
Through our close association with Xerox Corporation, we are able to provide great cost savings in reproduction and copying *511that will translate into cost savings for the contract. [XXXXX] [H] We believe our reproduction costs are consistent with the quantities in the SOW and represent a cost savings that we are passing on to the Government.
Admin. R.2001; see also id. 2333-34.
The Government’s treatment of the third offeror in the competitive range, Danya, demonstrates that it could not have believed at the time that copying costs were expected to be the same for all potential contractors. This offeror initially proposed $[XXXXX] in reproduction costs, was told to come down to $[XXXXX], and obliged. Admin. R. 421. The Government was apparently content to allow Danya to propose copying costs that were over $[XXXXX] more than the IGCE it was using in May, 2004.
The Government’s confusion in evaluating reproduction costs was further evidenced in the “Summary of Negotiations” document signed by the cost analyst and the first Contracting Officer on June 25, 2004. Attached to this document are cost analysis sheets for each of three bidders in the competitive range. The table for IQS has the reproduction cost figure initially proposed ($[XXXXX], as calculated by the Government), the Government’s pre-negotiation target (the first figure cut in half), the ultimate proposal from the Intervenor, and the following note: “The offerors had the correct amount in the first proposal.” Admin. R. 410. The text following this table explains: The Government cost objective of $[XXXXX] was successfully met and surpassed by $[XXXXX]. The Government erroneously negotiated for a lower number of copies then [sic] projected. The offeror proposed these costs based on $[XXXXX] for B & W copies and $[XXXXX] for Colored copies. This cost, it [sic] will be normalized in the .cost realism analysis to $ [XXXXX], an amount which will be used for all of the offerors in the competitive range.
Admin. R. 413.
For URC, the cost analysis table correctly identifies its initial proposal and final proposal for reproduction costs to be the same ($[XXXXX]), but mistakenly lists $9 million as its “pre-negotiation objectiv[e],” and adds the note: “Contractor refused SOW recommendations.” Admin. R. 418. The Government apparently confused the postage and shipping requirement with the copying requirement, stating in the text that “URC/SHS proposed $[XXXXX] [sic] the of-feror was advised to increase this to the recommended amount of $9,000,000 in the SOW; the offeror refused.” Id. 419.15 This $9 million figure was not required in the SOW for reproduction costs, [XXXXX]. See id. 421.
Four days later, the contracting officer was sent a “note” from three agency employees which informed her of the following:
Wanted to give you a heads-up in light of the pending award for the SAMHSA Health Information Network contract. During negotiations, there was a mathematical error made with regard to guidance provided to all three of the offerors on the cost of reprographics. The CO of record was advised of this problem and he advised that the correction should be made at a later time rather than reopen negotiations. [11] Regardless of which offeror receives the contract award, the successful offeror must increase their reprographics costs to a minimum of $[XXXXX]. [11] Please advise as to whether you would like to correct this at the time of award or in the future? Thank you.
Admin. R. 396.
The initial source selection determination explained:
Cost realism adjustments were made for all three offerors in the areas of ... reproduction ____These areas will be the same for any winning offeror; therefore, it is important to make sure that there is no ‘buying in’ or that other cost errors occur in these areas, which would effect [sic] the eost/price analysis of the offerors. [ ] The Reproduction amount was provided during *512negotiations, but due to a government miscalculation — an incorrect number was provided. This number is corrected in the cost realism analysis.
Admin. R. 391.
But the same document stated that:
The Cost Analyst, the Project Officer, and the Contracting Officer have reviewed every cost element proposed by the offerors and contemplated by the IGCE. All elements proposed by the offerors have been supported by reasonable, realistic, estimation assumptions, which have been explained in their business proposals. They are based on actual costs, and experiences of the contractors.
Id. 390. Concerning IQS specifically, it was said “[a]ll elements proposed by the offeror have been supported by reasonable, realistic assumptions, which have been explained in their business proposal. They are based on actual experiences of the contractor.” Id.
After the initial GAO decision, a second Contracting Officer made the Source Selection Determination, dated December 22, 2004. In this document, reproduction costs are merely explained as follows: “There are four cost elements within ODC that are adjusted for all offerors since the costs are considered to be the same for all.” Admin. R. 2763. The familiar figure of $[XXXXX] is used as the total reproduction costs for all offerors. Id. There is no explanation for the derivation of this number. The record contains a document that is identified in the index as “Independent Government Cost Estimate. Dated December 10, 2004.” This document does not itself contain any date, see Admin. R. 600-609. Moreover, while it lists as “Duplication and Photocopying” costs of $[XXXXX] for year one, based on 28,200,-000 copies at $[XXXXX] per copy, Admin. R. 601, this number appears to increase by 3.5% per year for years two through five. Admin. R. 603, 605, 607, 609.16
b. Was the normalization of reproduction costs rational?
As recounted above, at the time the initial proposals were reviewed by the agency, it had in its possession a government estimate that reproduction costs could be $[XXXXX] annually. The three offerors in the competitive range proposed costs of $[XXXXX], $[XXXXX], and $[XXXXX] over five years. Although it had an IGCE at the time, the agency did not propose that these costs be normalized, and in negotiations with the of-ferors it asked for bids ranging from $DXXXXX] to $[XXXXX] for this service.
IQS proposed to cut reproduction costs in half despite revising upward its per-page costs — tripling them for color copies and increasing to nearly six times the initial black- and-white estimate. After the second IGCE came in at close to twice the original one, the government decided to normalize this cost. But there is no explanation as to why reproduction costs were determined to be the same for all offerors. The record contains no contemporary rationale for this decision. It could not be merely based on the existence of an IGCE, as the agency had an IGCE when the initial proposals were reviewed. See Admin. R. 428-29, 436, 440. The record demonstrates that URC had explained why its copying costs might be lower than that of other offerors, Admin. R.2001; that the of-ferors’ initially proposed costs varied widely, see Admin. R. 421, 1128, 2344; and that during negotiations the agency was content for these to vary widely, see id. 421, 1015, 2128.
Nor would it be rational, given that the proposed copying costs varied widely, for the government to have normalized these costs to correct an underestimate of the number of copies required by the RFP. See Admin. R. 391, 396, 413. In the initial proposals, URC had a cost advantage in this area over the other two bidders in the competitive range— lower than Danya’s by $[XXXXX], and lower than IQS’s by $[XXXXX]. Compare Admin. R. 2344 with id. 421, 1128. Moreover, IQS concedes that its initially proposed cost of $[XXXXX] per black-and-white page was a mistake, Tr. (May 5, 2005) at 142. Thus, the difference in proposed costs between URC and IQS in this area might be much larger than $[XXXXX]. The marginal cost of re*513producing a page was initially $[XXXXX] (blending the black-and-white and color estimates) and rose to $[XXXXX] in IQS’s final proposal. This suggests that URC’s advantage over IQS in this area might have been greater than $[XXXXX] — and thus could have far exceeded IQS’s overall edge of $6.1 million, upon which the Source Selection Determination was based.
Under these circumstances, it appears it. was clearly arbitrary for the agency to have substituted the normalized reproduction costs figure of $[XXXXX] for URC’s proposed costs of $[XXXXX], It is also highly questionable for the agency to have adjusted intervenor’s costs by merely substituting the IGCE number.17 The latter was based on an average cost of $[XXXXX] per page, which was less than one-third of the marginal cost estimated by IQS in its final proposal. See Admin. R. 931. The per-page costs submitted by IQS constituted sufficient information upon which to calculate the costs of reproducing the number of pages required by the government. Computer Sciences Corp., 51 Fed.CI. at 316. In light of IQS’s apparent confusion concerning the requirements of the contract, the government might have been justified to seek further clarifications from IQS, but it does not appear rational to reward a confused offeror by eliminating a major cost advantage of one of its rivals. The plaintiff is highly likely to succeed on this claim. The Court thus turns to the other three prongs of the preliminary injunction standard, which primarily present issues of fact.
B. Will Plaintiff Suffer Immediate and Irreparable Harm?
The plaintiff has asserted that without a preliminary injunction, it will suffer immediate and irreparable harm. First, URC states that it will likely lose its staff, “whose skills range from professional public health experts to warehouse and distribution workers, specially trained to perform critical health information services to more than one million Americans each year,” if the award of the contract is not enjoined. Declaration of Dr. Lewis D. Eigen 114 (July 19, 2004) (“Eigen Decl. I”); Declaration of Dr. Lewis D. Eigen 1119 (Aug. 3, 2004) (“Eigen Deck II”). The plaintiff claims this loss of employees would make it difficult for the plaintiffs team to perform the contract were the Court to ultimately rule in its favor on the merits. Id. In addition, the plaintiff asserts its subcontractor will lose leases on several facilities critical for the plaintiffs performance of the contract. Id. The plaintiff asserts that the loss of the leases would also cause the plaintiff to forfeit specialized improvements it made to the leased property. Id. Finally, URC states that it will lose its lease and services contract [XXXXX] and lose its contracts with internet service providers if preliminary injunctive relief is not granted. Id.; see also Ex. B to Pl.’s Mem. (Deck of Charles C. Pecarro).
Conversely, the Government and the inter-venor maintain that any harm the plaintiff will suffer is not immediate and irreparable. Def.’s Opp. at 25; Int.’s Opp. at 23-26. They argue that to the extent that IQS hires employees of the plaintiff, if the award of the contract to IQS were overturned, the plaintiff could rehire those employees. In terms of leases and related contracts, the intervenor asserts that URC should have anticipated the award of the contract to IQS and planned accordingly; that certain of those contracts and leases serve other operations of the plaintiffs subcontractor; and that some contracts could be, assumed by IQS and then re-assumed by URC in the event that the agency award determination were overturned. Int.’s Opp. at 23-24, 26. Thus, the Government and the intervenor argue that URC’s purported harm is not likely to arise or is speculative at best.
It appears to the Court that any improvements to facilities made by URC or its proposed subcontractor that would be forfeited upon the termination of leases would not be compensable were the contract to be deter*514mined to have been improperly awarded to IQS. Nor is it certain that increases in the cost of leased space and equipment would be reimbursable under the contract at issue. In addition, URC’s proposal was predicated on its ability to maintain the quality staff that were performing the incumbent contract, and this technical capability could be seriously impaired were the transition to continue and performance under the new contract to begin. Taken as a whole, this potential injury amounts to sufficiently immediate and irreparable harm for purposes of a preliminary injunction.
C. Will the Plaintiffs Harm Outweigh Harm to the Government and Third Parties?
The plaintiff asserts that the Government’s harm is actually minimized under a preliminary injunction, as URC’s team is currently providing the services required by the contract and there will be no loss of the services during the pendency this litigation. PL’s Mem. at 37-38. In addition, URC argues that IQS bid on this contract with the knowledge that the award or performance of the contract, like other government contracts, can be suspended as the result of a challenge by a disappointed bidder. Id. at 38. The Government counters by arguing that the delay of the beginning of the transition to a new contractor interferes with the efficient and effective provision of services by SAMH-SA. Def.’s Opp. at 26. The intervenor asserts that the delay in the transition from URC’s team to the intervenor has put strains on relationships it has with the potential subcontractors it would use if the award were upheld, and that it has lost money on resources it has been forced to maintain in the event it receives the contract. Int.’s Opp. at 27-29.
The Government and IQS’s asserted harm — delay in transition to the new contract — exists in virtually all bid protests where preliminary relief is sought. Our Court has on occasion been skeptical as to whether delay in transition, and associated costs and inconveniences, can outweigh a plaintiffs harms. See PGBA, 57 Fed.Cl. at 663; Overstreet, 47 Fed.Cl. at 744; Ellsworth Assocs., Inc. v. United States, 45 Fed.Cl. 388, 399 (1999). As has been observed, “only in an exceptional case would [delay] alone warrant a denial of injunctive relief, or the courts would never grant injunctive relief in bid protests.” Ellsworth Assocs., 45 Fed.Cl. at 399. This is not an exceptional case. The costs and inconvenience associated with the delay in transition caused by the issuance of a preliminary injunction in this case are typical of the costs and inconveniences that face almost every new contractor seeking to unseat an incumbent contractor, and do not outweigh the imminent harm to plaintiff.
Much of the harm asserted by the interve-nor would have its natural counterpart in harm to the plaintiff were a preliminary injunction not granted, and could constitute harm to the Government if injunctive relief is improvidently denied (such as transition costs otherwise avoided). This factor, as a practical matter, is for this reason usually dominated by the likelihood of success factor. And an incumbent is likely to suffer potentially greater injury than a new bidder for a contract, based on the difference between the actual and the prospective: an incumbent’s bid is based on keeping together a team whose existence is eliminated in the absence of injunctive relief, while a new bidder is usually projecting the sort of team it could assemble if awarded the contract. This factor favors URC.
D. Is the Granting of Injunctive Relief in the Public Interest?
Finally, the plaintiff claims that an injunction is in the public interest, as “by upholding the integrity of the procurement process, public confidence in the federal procurement process will be preserved.” SAI Indus. Corp. v. United States, 60 Fed.Cl. 731, 747 (2004). The Government and the intervenor maintain that the public interest is also served by the timely completion of the government procurement process. In support of their argument, the Government and the in-tervenor point to the amount of time that has already been spent litigating this matter before the GAO. Although the two GAO protests have delayed the award of the contract, it bears noting that the GAO did sustain the *515plaintiffs first protest due to an intentional misrepresentation of the agency during the initial procurement. University Research, 2004 CPD If 217. And even though the agency did not sustain the plaintiffs protest as to cost normalization in the second protest, the agency did note errors in the agency’s evaluation of the proposals. Given the misrepresentation and errors addressed by the GAO and errors this Court has noted from its initial review of the Administrative Record,18 the public’s interest in the integrity of the procurement process outweighs the public’s interest in the timely completion of the government procurement process.
The public interest is not served when a government contract that is subject to competitive bidding is arbitrarily awarded. The likelihood that this would result absent an injunction strongly militates in favor of in-junctive relief. And thus the desire for a timely completion of the procurement process may thereby be enlisted in the cause of ensuring that the procurement process is properly conducted — by giving the Government an incentive to take the time needed to get the award right the first (or at least the second) time. The public interest is not well-served when contracting officials rush to save a few weeks and end up delaying contracts by many months.
III. CONCLUSION
For the foregoing reasons, on May 6, 2005, the plaintiffs motion for a preliminary injunction was GRANTED. Accordingly, pending further order of the Court, IT IS ORDERED that the United States, including the Department of Health and Human Services, Substance Abuse and Mental Health Services Administration, its contracting officer, and its other officers, agents, servants, employees, and representatives, and all persons acting in concert and participating with them respecting the subject procurement, be and they are hereby RESTRAINED AND ENJOINED from initiating the transition to and continuing the performance of the contract awarded to IQ Solutions, Inc. pursuant to RFP No. 277-04-6091.
Under Rule 65(c) of the Rules of the Court of Federal Claims, the plaintiff must give security “in such sum as the court deems proper.” The Court has determined that the proper amount of security in this case is $0.00.
IT IS SO ORDERED.
. Protests can be brought before the GAO under 31 U.S.C. §§ 3551-3556. This Court does not sit in appellate review of GAO decisions. Arch Chemicals, Inc. v. United States, 64 Fed.Cl. 380, 383 n. 4 (2005); Cubic Applications, Inc. v. United States, 37 Fed.Cl. 339, 343 (1997).
. Section 3553 provides that if the agency receives notice of a bid protest in the GAO, "the contracting officer may not authorize performance of the contract to begin while the protest is pending." 31 U.S.C. § 3553(d)(3)(A)(i).
. The stay override provision in section 3553 provides that the "head of procuring activity” may override the stay by asserting in writing that either: 1) "performance of the contract is in the best interests of the United States,” or 2) "urgent and compelling circumstances that significantly affect interests of the United States will not permit waiting for the decision of the Comptroller General concerning the protest.” 31 U.S.C. § 3553(d)(3)(C)(i).
. In RAMCOR, the Federal Circuit held that this Court could review the merits of an agency’s override decision independent of any consideration of the merits of the underlying contract award. RAMCOR, 185 F.3d at 1289. This conclusion rested on the pertinent language of the Tucker Act, which states that the Court of Federal Claims has jurisdiction over, among other things, a "violation of a statute or regulation in connection with a procurement or a proposed procurement.” 28 U.S.C. § 1401(b)(1) (emphasis added); RAMCOR, 185 F.3d at 1289. The override provision of 31 U.S.C. § 3553(c)(2), according to the Circuit, "fits comfortably” within the "in connection with” language of the Tucker Act. RAMCOR, 185 F.3d at 1289. In accordance with RAMCOR, our Court has concluded that the override provision of 31 U.S.C. § 3553(d)(3)(C) similarly "fits comfortably” within the "in connection with” language of the Tucker Act. See *503PGBA, LLC V. United States, 57 Fed.Cl. 655, 659 (2004).
. As the Court noted in PGBA, section 3553 "bears little resemblance to the types of statutes that generally have been held to produce non-reviewable decisions, most of which specifically refer to the agency’s discretion and many of which have involved national security concerns.” PGBA, 57 Fed.Cl. at 659. Furthermore, as the PGBA Court explained, "there are standards by which to review the 'best interest’ finding, among them whether the assumptions or factual assertions upon which that finding is based are borne out by the record; and whether the factors relied upon by the contracting official were relevant or, conversely, whether factors relevant to the determination were ignored.” Id.
. The Government’s opposition to the plaintiffs motion implicitly acknowledges that the averaging of past performance scores was flawed. Def.'s Opp. 14-15. The Government's argument on this point is that the error is not prejudicial and, therefore, would not likely lead to success on the merits.
. The GAO found that the averaging method was irrational, 2005 U.S. Comp. Gen. LEXIS 73 at *30, 2005 WL 1026055 at *10, but held that the errors were non-prejudicial. 2005 U.S. Comp. Gen. LEXIS 73 at *63, 2005 WL 1026055 at *23.
. The plaintiffs portion of the spreadsheet had in total five rows: two for itself, two for one proposed subcontractor, and one for another proposed subcontractor. Admin. R. 2761.
. It appears that agency initially incorrectly multiplied the score on the twenty-five point scale by 1.5, see Admin. R. 450; however, this was corrected in the source selection determination. See Admin. R. 2761.
. It appears from the record that the agency added a fifth grade, for customer satisfaction, to this score, increasing it by four points. Admin. R. 3336 (explaining adjustment to, among others, contract PCE-I-00-97-00042-00). The adjusted score of seventeen does not appear to be included in the past performance evaluation table used in the Source Selection Determination. Compare Admin. R. 2761 (the twenty-five point scale row has no score of seventeen) with id. 3339.
. The applicable regulation states in relevant part:
(i) The probable cost may differ from the proposed cost and should reflect the Government's best estimate of the cost of any contract that is most likely to result from the offeror’s proposal. The probable cost shall be used for purposes of evaluation to determine the best value.
(ii) The probable cost is determined by adjusting each offeror's proposed cost, and fee when appropriate, to reflect any additions or reductions in cost elements to realistic levels based on the results of the cost realism analysis.
48 C.F.R. § 15.404-1 (d)(2)(i),(ii).
. It is hard to see why IQS thought that the Government meant to draw attention to paragraph H. 16, as opposed to the SOW page that was clearly referenced and discussed the printing requirements — particularly since the former begins "[ujnless otherwise specified in this contract, the Contractor shall not engage in, nor subcontract for, any printing....” Admin. R. 499 (emphasis added). The RFP obviously called for printing by the Contractor. See Admin. R. 573.
. This $[XXXXX] figure cannot be explained by increasing the IGCE number by 15-20% each of the succeeding four years, for this would exceed $[XXXXX] at the low range.
. Plaintiff proposed $[XXXXX] for shipping and postage, rather than the $9-10 million required by the SOW. Compare AR 2007 with id. 572.
. The five year total would thus be $[XXXXX].
. This is all the more questionable given the possibility that the IGCE number might have been based on URC’s historic costs and thus might reflect URC’s particular cost advantage. The record contains no support for the IGCE calculation. Moreover, the cost analyst’s many errors concerning this cost item, detailed above, provide even more reason for doubt. See Admin. R. 418-19, 429, 436, 438, 1128 (using inconsistent IGCE and target numbers).
. The errors the Court has noticed in its initial review of the Administrative Record are not necessarily dispositive on the merits; however, they do weigh against the Government as to this factor of the Court's injunctive relief analysis.