Goldberg, Senior Judge:
This matter concerns the final determination issued by the Department of Commerce ("Commerce" or the "Department") in the antidumping duty investigation of certain cut-to-length steel products. Certain Carbon and Alloy Steel Cut-to-Length Plate from Austria , 82 Fed. Reg. 16,366 (Dep't Commerce Apr. 4, 2017) (final determ.) (" Final Determination "), and accompanying Issues & Decision Mem. ("I&D Mem."); Certain Carbon and Alloy Steel Cut-to-Length Plate from Austria , 81 Fed. Reg. 79,416 (Dep't Commerce Nov. 14, 2016) (" Preliminary Determination "), and accompanying Issues & Decision Mem. ("Prelim. Mem."). Plaintiffs Bohler Bleche GmbH & Co. KG, Bohler International GmbH, voestalpine Grobblech GmbH, and voestalpine Steel & Service Center GmbH (collectively, "Plaintiffs"), challenge the methodology used by Commerce to select foreign like products in connection with its calculation of antidumping duties. For the reasons below, the court remands the Final Determination for reconsideration in accordance with this opinion.
BACKGROUND
In order to determine whether certain products are being sold at less than fair value (LTFV) in the United States, Commerce compares the export price (EP), or constructed export price (CEP), with the normal value (NV). 19 U.S.C. § 1677b(a)(1)(A). EP/CEP is the price at which the subject merchandise is being *1347sold in the U.S. market, while NV is the price at which a "foreign like product" is sold in the producer's home market or in a comparable third-country market. Therefore, before calculating a dumping margin, Commerce must identify a suitable "foreign like product" with which to compare the exported subject merchandise. A "foreign like product," in order of preference, is:
(A) The subject merchandise and other merchandise which is identical in physical characteristics with, and was produced in the same country by the same person as, that merchandise.
(B) Merchandise (i) produced in the same country and by the same person as the subject merchandise, (ii) like that merchandise in component material or materials and in the purposes for which used, and (iii) approximately equal in commercial value to the subject merchandise.
(C) Merchandise (i) produced in the same country and by the same person and of the same general class or kind as the merchandise which is the subject of the investigation, (ii) like that merchandise in the purposes for which used, and (iii) which the administering authority determines may reasonably be compared with that merchandise.
19 U.S.C. § 1677(16) ; NSK Ltd. v. United States , 26 CIT 650, 657-58, 217 F.Supp.2d 1291, 1299-1300 (2002) ("Section [1677](16) establishes a descending hierarchy of preferential modes that Commerce must select for matching purposes."). To identify such merchandise, Commerce designs a "model-match" methodology consisting of a hierarchy of certain characteristics used to sort merchandise into groups. Each group is then assigned a control number ("CONNUM"), used to match home market sales with U.S. sales.
In the instant proceeding, the Department compared the weighted-average of export sales within each CONNUM to the weighted-average of home market sales in that same CONNUM, i.e., identical merchandise, where such sales exist. Prelim. Mem. 7. Otherwise, "[w]here there were no sales of identical merchandise in the home market made in the ordinary course of trade to compare to U.S. sales, [Commerce] compared U.S. sales to sales of the most similar foreign like product made in the ordinary course of trade." Id. (emphasis added). "When comparing U.S. sales with comparison-market sales of similar, but not identical, merchandise, [the Department] also made adjustments for physical differences in the merchandise in accordance with section [ 19 U.S.C. § 1677b(a)(6)(C)(ii) ] and 19 CFR 351.411." Id. at 14. This adjustment is called a DIFMER.
The investigation at issue covers certain steel cut-to-length (CTL) plate products from Austria. I&D Mem. 5. Commerce also conducted concurrent investigations of steel CTL plate from other countries.1 Early in the investigation, on May 19, *13482016, Commerce proposed a model-match methodology and sought comments from interested parties across the concurrent CTL plate investigations. Ltr. to All Interested Parties 5 (May 19, 2016), P.R. 83. Among other features of Commerce's proposed methodology, the third field in the hierarchy was QUALITY, which sorts merchandise based on various quality-related characteristics. Id. at 5. Plaintiffs requested a number of changes to the methodology, including that the QUALITY field be placed first in the hierarchy, and that a QUALITY subcategory be added to distinguish high alloy tool steel products. I&D Mem. 16. Plaintiffs explained that their suggested changes would help solve issues created by the allegedly broad scope of the investigation, namely, that the model-match methodology, as proposed, would likely fail to sufficiently differentiate products with distinct commercial characteristics and values. Pls. Cmts. on Product Characteristics 3 (June 2, 2016), P.R. 96; see also Pls. Case Br. 4-5 (February 16, 2017), P.R. 397-403 (arguing that "[t]he breadth of the scope of this investigation is unprecedented" and that "the Department's CONNUM methodology ... essentially mirrored the methodology used in past investigations that covered only carbon CTL plate.").
On June 10, 2016, Commerce issued its revised model-match methodology, adopting some of Plaintiffs' requests, over the objections of Petitioners, and rejecting other suggestions by Plaintiffs. See Product Characteristics 1, 6 (June 10, 2016), P.R. 115. Commerce's revised methodology:
Matched foreign like products, based on the physical characteristics reported by the respondents, in the following order of importance: quality, minimum specified carbon content, minimum specified chromium content, minimum specified nickel content, minimum specified yield strength, nominal thickness, heat treatment status, nominal width, form, painting, the existence of patterns in relief and descaling.
See Prelim. Mem. 7.
In its July 15, 2016 questionnaire responses, Plaintiffs requested additional changes to Commerce's model-match methodology, on the basis that the revised methodology still captured drastically dissimilar products within individual CONNUMs. Pls. Questionnaire Resp. B-13, C-10 (July 15, 2016), P.R. 163-174. Specifically, Plaintiffs argued that the methodology failed to account for the alloy content of Plaintiffs' specialized high alloy steel products, thereby failing to account for significant differences in physical characteristics, costs, and price. Id. Plaintiffs proposed a revised methodology that would begin by sorting products by GRADE. Id. at B-14, C-11. The GRADE field is essentially a function of the amount of alloy in a product and the costs of those alloys. See Pls. Supp. Questionnaire Sec. D & E Resp. Ex. SQ D-3 (Oct. 6, 2016), P.R. 268-89. Plaintiffs later suggested a PROCESS field to further account for what Plaintiffs insist are significant variations in cost of production resulting from different manufacturing processes. See Pls. Supp. Questionnaire Sec. B & C Resp. SBC-1 (Oct. 13, 2016), P.R. 296-304; see also Pls. Case Br. 11. Commerce rejected Plaintiffs' proposed revisions on two bases: Commerce considered the proposals to be untimely and Commerce also disagreed that the newly proposed methodologies would have the effect of creating closer matches between exported merchandise and home market merchandise. I&D Mem. 17-32.
*1349On November 14, 2016, Commerce published its Preliminary Determination . 82 Fed. Reg. 16,366. On April 7, 2017, Commerce published its Final Determination . 81 Fed. Reg. 79,416. Plaintiffs timely filed the instant action to contest the Final Determination , challenging Commerce's model-match methodology and the resulting antidumping duties. For the reasons discussed below, the court remands to Commerce for reconsideration of its model-match methodology.
JURISDICTION AND STANDARD OF REVIEW
The court has jurisdiction over this matter under 28 U.S.C. § 1581(c). The court will sustain Commerce's determinations regarding its model-match methodology if they are supported by substantial evidence and otherwise in accordance with law. SKF USA, Inc. v. United States , 537 F.3d 1373, 1378 (Fed. Cir. 2008).
DISCUSSION
Plaintiffs challenge the model-match methodology applied by Commerce to determine whether Plaintiffs sold certain steel products at LTFV. Plaintiffs argue, as they did below, that the Department's methodology yielded erroneous comparisons of merchandise. In its questionnaire responses to the Department, Plaintiffs asserted that the Department's CONNUM system "does not provide an accurate basis for comparing [home market and export] sales ... because it unreasonably groups together high alloy, Special Steel CTL plate products that differ significantly ... resulting in CONNUMs ... with wildly divergent sales prices and costs of production" which could "lead to highly arbitrary dumping margin calculations." Pls. Questionnaire Resp. B-13-14. In support, Plaintiffs have identified individual CONNUMs that group together products that Plaintiffs describe as having commercially significant physical differences, which are reflected in the costs and prices of those products.
Plaintiffs explain that the variances in cost of manufacturing and price are primarily attributable to the alloy content of the products, a physical feature insufficiently unaccounted for under the Department's methodology. Plaintiffs argue that alloy content, or grade, is not only a commercially significant physical characteristic of its various products, but indeed one of their defining commercial features. Therefore, according to Plaintiffs, the Department's insistence that these various, distinct products can nevertheless form the basis of a single CONNUM is not supported by substantial evidence and not otherwise in accordance with the law.
Plaintiffs' proposed two alternative model-match methodologies that they argue would achieve greater differentiation of distinct products, leading to more accurate dumping margins. The first alternative proposed the addition of a GRADE field to sort products by alloy content. The second alternative proposed that, in addition to a GRADE field, a PROCESS field be included to account for variations among the products owing to certain manufacturing processes to which only some products are subjected. The court sustains the determination of the Department insofar as it refused to amend the methodology to account for the manufacturing processes identified by Plaintiffs. However, the court remands to Commerce for redetermination in light of its methodology's deficient treatment of commercially significant physical differences among Plaintiffs' products owing to alloy content.
*1350A. Foreign Like Product
As discussed, " '[f]oreign like product' is defined as either identical merchandise, § 1677(16)(A), or similar merchandise, § 1677(16)(B) and (C)." SKF USA, Inc. , 537 F.3d at 1375 (emphasis added). "Congress has granted Commerce considerable discretion to fashion the methodology used to determine what constitutes 'foreign like product' under the statute." Id. at 1379 (citing Pesquera Mares Australes Ltda. v. United States, 266 F.3d 1372, 1384 (Fed. Cir. 2001) ). In identifying "merchandise which is identical in physical characteristics" to subject merchandise, 19 U.S.C. § 1677(16)(A), Commerce may select products with "minor differences in physical characteristics, if those [minor] differences are not commercially significant." Pesquera Mares , 266 F.3d at 1384. What is considered a "commercially significant" factor is determined on a case-by-case basis, but at the very least it is a feature that is recognized in the broader industry of the subject merchandise. See id. at 1385. At the end of the day, the Department's methodology must be faithful to the statutory directive that a "fair comparison shall be made." 19 U.S.C. § 1677b(a).
B. The Department's Methodology Fails to Account for Commercially Significant Physical Differences Based on Alloy Content
Here, Commerce's methodology is not suited for such fair comparisons. The methodology clearly groups, within individual CONNUMs, goods that have differences in material composition. These differences are neither minor nor commercially insignificant. Plaintiffs explain that:
[T]he percentage differences between the highest grade and the lowest cost grade within each CONNUM are enormous. For example, CONNUM [ [ ] ] covers four grades ... ranging from a high VCOM of €[ [ ] ]/MT to a low VCOM of €[ [ ] ]/MT, which is a difference of €[ [ ] ]/MT or [ [ ] ]% of TCOM. On average across all listed CONNUMs, there is a difference of 96%, as a percentage of TOTCOM, between the highest-cost grade and the lowest-cost grade within each CONNUM ....
Pls. Pre-Prelim. Determ. Cmts. 11 (Oct, 19, 2016), P.R. 312-17. Similarly, Plaintiffs explain that CONNUM [ [ ] ] includes products that cost anywhere from €[ [ ] ] to €[ [ ] ] to make and are priced at anywhere from €[ [ ] ] to €[ [ ] ]. See Pls. Case Br. 16. Moreover, Plaintiffs insist, and the Department does not contest, that these differences in cost and price are driven by alloy content which, in turn, is driven primarily by the intended end use. See, e.g. , Pls. Case Br. 3-4, 7-8. As is supported by the record, end use is plainly a commercially significant factor. See Pls. Supp. Questionnaire Resp. Sec. D & E 7-10 (discussing the various manufacturing applications that require particular grades of steel products, including "automobiles, home appliances, electronic devices, medical consumables (e.g., syringes), [and] consumer goods."). Accordingly, because a single set of CONNUMs are used to identify and compare home market and U.S. sales, Commerce's methodology essentially treats certain "foreign like products" as "identical" to certain exported products, even though customers would view those products as commercially distinct in both utility and value.2
*1351It appears that the Department, too, would typically view many of these products as dissimilar. As the Department has explained in prior proceedings, the statute, at 19 U.S.C. § 1677b(a)(6)(C)(ii) :
Requires that we account for and adjust for any differences attributable to physical differences between subject merchandise and foreign like product if similar products are compared. For this purpose, [ 19 C.F.R. § 351.411(b) ] directs us to consider differences in variable costs associated with the physical differences in the merchandise, i.e., the difference-in-merchandise adjustment [or DIFMER].
Thai Plastic Bags Indus. Co. v. United States , 34 CIT 1389, 1394, 752 F.Supp.2d 1316, 1322 (2010). Further, Commerce's own practice would have called into question whether some of the products within individual CONNUMs were even "similar," given the magnitude of the DIFMER adjustments that would be applied. See, e.g. , Department of Commerce Import Administration Policy Bulletin, Number 92.2, "Differences in Merchandise; 20% Rule" (Jul. 29, 1992) (available at < https://enforcement.trade.gov/policy/bull92-2.txt>). Specifically, under the Department's "20% Rule," a DIFMER greater than 20% creates a presumption that two products are not "similar." And as Plaintiffs have demonstrated, under Commerce's methodology, there are a number of product pairs within individual CONNUMs with differences in variable costs well above 20%. See, e.g. , Pls. Case Br. 16.
Nevertheless, as discussed, where there are export sales and home market sales within the same CONNUM, and the Department has therefore deemed those respective sales to be of "identical," rather than "similar," merchandise, the Department has not applied a DIFMER to account for any differences in material composition. Prelim. Mem. 7; 19 C.F.R. § 351.411(b) ("The Secretary will not consider differences in cost of production when compared merchandise has identical physical characteristics."). As Plaintiffs correctly note, it is not reasonable to interpret the statute as permitting a greater gap in variable costs for "identical" merchandise than would be allowed for "similar" merchandise. That anomalous result is explained here by the fact that Commerce's methodology is not in fact designed to reasonably ensure fair comparisons of "identical" merchandise within each CONNUM.3
Plaintiffs identified such erroneous comparisons in their Case Brief below. Specifically, *1352Plaintiffs demonstrated that certain products that were "sold at prices above COP" but nevertheless "appear to have been sold below cost because the weighted average COP used for the Department's CONNUM is skewed by extremely high-cost" products. Pls. Case Br 15-17 & Attach. 2. Neither the Department below nor the Defendant in this action has called this analysis, or the record evidence on which it relies, into question.
In sum, the Department's methodology is unreasonable in failing to sufficiently account for alloy content. Commerce's methodology cannot be sustained because it allows subject merchandise to be cast as "identical" to dubiously similar foreign like products, when the statute plainly requires a different approach. Moreover, the record lacks substantial evidence that Commerce's flawed methodology nevertheless yielded "fair comparisons." See 19 U.S.C. § 1677b(a). On this basis, the court remands to Commerce for reconsideration of its model-match methodology.
Throughout the investigation, the Department largely ignored Plaintiffs' central argument: that the Department's methodology allows comparisons of products with commercially distinct physical characteristics, without applying a DIFMER, to determine whether Plaintiffs are dumping. Unfortunately, the Department dedicates significant energy to explaining why it believes it did not have to address Plaintiffs' arguments. Specifically, the Department insists that the Plaintiffs' challenges to the model-match methodology were untimely. They were not.
Plaintiffs raised their concerns at every turn. Plaintiffs proposed addition of a GRADE field to account for alloy content was submitted with their questionnaire responses on July 15, 2016, just 35 days after the Department had issued its revised model-match methodology, four months prior to the Department's Preliminary Determination , and four months before the Department issued its final ruling on the scope of the investigation. Commerce then reviewed Plaintiffs' GRADE-field proposal and sought additional clarifying information on this issue in its September 14, 2016 supplemental questionnaire, which Plaintiff then provided. See Pls. Supp. Questionnaire Resp. Sec. D & E 7. The court will not now entertain the Government's argument that the model-match methodology was a closed issue prior to July 15, 2016. But even if the Department did consider the issue closed, this is a case where "the interests in fairness and accuracy outweigh the burden upon Commerce" presented by having to consider Plaintiffs' concerns about the model-match methodology. See Grobest & I-Mei Indus. (Vietnam) Co. v. United States , 36 CIT ----, ----, 815 F. Supp. 2d 1342, 1367 (2012).
The Government also feigns confusion at Plaintiffs' supposed change of position over time. Specifically, Commerce insists that Plaintiffs provide no explanation for suddenly providing an alternative set of CONNUMs based on GRADE, and then PROCESS, when previously they had provided comments only on the QUALITY-based CONNUMs. On the contrary, the record makes clear that Plaintiffs' overarching concern-that the Department's proposed model-match methodology would insufficiently differentiate high alloy specialty steel products-was consistently communicated.
*13534 This concern permeated Plaintiffs' comments on scope and on the QUALITY-based CONNUMs, as well as their later proposals of alternative CONNUMs. That Plaintiffs' proposed solution for this problem evolved over a relatively short period does not make their conduct contradictory. Frankly, Commerce's alleged confusion as to why Plaintiffs "suddenly" supported a GRADE-based CONNUM system strains credulity.
In response to Plaintiffs' claim that its products have commercially significant differences in physical characteristics that impact costs and price, the Department stated that this argument improperly focuses on "differences in costs or prices that may coincide with some type of variation in physical differences." I&D Mem. 23, 31. The Department cited prior proceedings to support the idea that differences in costs, in and of themselves, are not probative of relevant physical differences among products. I&D Mem. 23 n.86. As one prior proceeding explained, differences in costs are sometimes attributable to factors other than the physical characteristics of the products, such as "differences in production quantities [or] differences in the timing of production." See id. (citing Certain Cold-Rolled Steel Flat Products from the United Kingdom , 81 Fed. Reg. 49,929 (Dep't Commerce July 29, 2016) (final determ.), and accompanying I&D Mem. cmt. 5). It is not immediately clear why Commerce relies on these particular authorities here. Plaintiffs consistently point to material composition-a bona fide physical characteristic-that results in significant variances in costs and prices. By no means do Plaintiffs rely on costs "in and of themselves."
The Department's remaining responses essentially call into question the merits of Plaintiffs' proposed alternative methodologies. The Department expresses concern that Plaintiff's proposals would amount to a proliferation of respondent- and product-specific CONNUMS. I&D Mem. 21. As a consequence, Commerce insists, the product distinctions Plaintiffs are requesting would not be relevant to the Department's other concurrent steel CTL plate investigations.
First, the reasonableness of the methodology in those other investigations is not the subject of this action. And while it would certainly be more convenient for the Department, Commerce cites no authority supporting a right to apply a single methodology across multiple investigations, notwithstanding potentially serious issues with that methodology. Indeed, perhaps the Department should anticipate the issues raised here when it applies a single methodology across 15 concurrent investigations of broad scope. Second, it may be that this particular investigation, or Plaintiffs' particular range of product offerings, call for more granular distinctions between products. See, e.g. , Pastificio Lucio Garofalo, S.p.A. v. United States , 35 CIT ----, ----, 783 F.Supp.2d 1230, 1242 (2011) (the Department defended its application of company-specific CONNUMs on the basis that "substantial evidence regarding physical differences" were "reflected in ... costs and prices."); SKF USA , 537 F.3d at 1379 (explaining that a methodology with "increased number of price comparisons ... yields more accurate results because it matches the most similar product rather than merely pooling several models that *1354matched as to eight characteristics but could vary significantly in price or cost, due to differences in materials for certain components or added features.").
In any event, Commerce's critiques concerning Plaintiffs' preferred methodologies largely lack legal significance. Certainly, Plaintiffs' case would be stronger if it had devised a perfect alternative to Commerce's methodology. However, if there are two reasonable conclusions-or, here, methodologies-supported by the record, the Department has the discretion to choose either. Thus, the only question before this court is whether the Department's chosen methodology is reasonable, supported by substantial evidence on the record, and otherwise in accordance with the law. Insofar as that methodology insufficiently accounts for alloy content in Plaintiffs' products, the court finds that it is none of the above.
Accordingly, on remand, the Department is ordered to amend its model-match methodology in this investigation, in accordance with this opinion, to better account for the commercially significant differences in physical characteristics among Plaintiffs' products owing to alloy content. The Department is strongly encouraged to ensure that products it would not consider "similar," or to which it would apply a meaningful DIFMER, are not matched as "identical" under its revised methodology.
C. The Court Sustains the Department's Determination that a Process is not a Commercially Significant Physical Characteristic
The court comes to a different conclusion concerning the arguments raised by Plaintiffs with respect to a PROCESS field. Substantial evidence supports the Department's determination that its methodology need not further account for process.
Plaintiffs explain the process issue as follows: "different production processes are required depending on the alloy content of the plate. As the alloy content increases, more control is required in the melting, solidification, and production processes. To achieve such control, [Plaintiffs] use different production processes, with significantly different costs." Pls. Case Br. 8. Plaintiffs' own explanation betrays that "process" is not itself a physical characteristic of the products, but is rather a function of a physical characteristic, alloy content. Moreover, while the record supports Plaintiffs' assertion that various processes can result in alterations in physical characteristics and lead to variances in costs and price, those variances should be accounted for when the Department gives sufficient consideration to grade in its redetermination. Ultimately, substantial evidence, as well as Plaintiffs' own arguments, support the Department's determination that a PROCESS field would largely account for "variations in cost," I&D Mem. 22, and that "the PROCESS field would have no impact" if the GRADE field were applied, I&D Mem. 28.
CONCLUSION
For the foregoing reasons, the court sustains Commerce's determination in part and remands to Commerce for reconsideration of its model-match methodology. It is hereby:
ORDERED that the final determination of the United States Department of Commerce, published as *1355Certain Carbon and Alloy Steel Cut-to-Length Plate from Austria , 82 Fed. Reg. 16,366 (Dep't Commerce Apr. 4, 2017) (final determ.) is hereby REMANDED to Commerce for redetermination; it is further
ORDERED that Plaintiffs' Motion for Judgment on the Agency Record Under USCIT Rule 56.2 is GRANTED in part as provided in this Opinion and Order; it is further
ORDERED that Commerce shall issue a redetermination ("Remand Redetermination") in accordance with this Opinion and Order that is in all respects supported by substantial evidence and in accordance with law; it is further
ORDERED that Commerce shall design a model-match methodology in this investigation that accounts for all commercially significant physical differences; it is further
ORDERED that Commerce apply recalculate dumping margins consistent with its redetermination of its model-match methodology; it is further
ORDERED that Commerce shall have ninety (90) days from the date of this Opinion and Order in which to file its Remand Redetermination, which shall comply with all directives in this Opinion and Order; that the Plaintiffs and Defendant-Intervenors shall have thirty (30) days from the filing of the Remand Redetermination in which to file comments thereon; and that the Defendant shall have thirty (30) days from the filing of Plaintiffs' and Defendant-Intervenors' comments to file comments.