Katzmann, Judge: At the center of this case is the challenge to the Department of Commerce's ("Commerce") final affirmative determination of sales at less-than-fair value in its anti-dumping investigation of certain carbon and alloy steel cut-to-length ("CTL") plate from France. Certain Carbon and Alloy Steel Cut-To-Length Plate from Austria, Belgium, France, the Federal Republic of Germany, Italy, Japan, the Republic of Korea, and Taiwan: Amended Final Affirmative Antidumping Determinations for France, the Federal Republic of Germany, the Republic of Korea and Taiwan, and Antidumping Duty Orders ("Final Determination"), 82 Fed. Reg. 24,096 (May 25, 2017), P.R. 456 and accompanying *1355Issues and Decision Memorandum ("IDM") (Mar. 29, 2017), P.R. 445. Plaintiff Dillinger France S.A. ("Dillinger") contests multiple aspects of Commerce's decision -- including the level of trade analysis, the application of partial adverse facts available, the use of the differential pricing methodology, and cost-shifting between products -- and asks the court to remand the Final Determination. Pl.'s Br., Dec. 18, 2017, ECF Nos. 26-28. Defendant the United States ("the Government") and Defendant-Intervenor Nucor Corporation ("Nucor") ask the court to sustain Commerce's decision in its entirety. Def.'s Br., Feb. 16, 2018, ECF No. 32; Def.-Inter.'s Br., Feb. 20, 2018, ECF Nos. 38-39. The court sustains the Final Determination in part and remands Commerce's application of partial adverse facts available for reconsideration.
BACKGROUND
I. Legal Background.
Pursuant to United States antidumping law, Commerce must impose antidumping duties on subject merchandise that "is being, or is likely to be, sold in the United States at less than fair value" and that causes material injury or threat of material injury to a domestic industry. 19 U.S.C. § 1673 (2012).1 "Sales at less than fair value are those sales for which the 'normal value' (the price a producer charges in its home market) exceeds the 'export price' (the price of the product in the United States)." Apex Frozen Foods Private Ltd. v. United States, 862 F.3d 1322, 1326 (Fed. Cir. 2017) (quoting Union Steel v. United States, 713 F.3d 1101, 1103 (Fed. Cir. 2013) ). Normal value is defined as "the price at which the foreign like product is first sold ... in the exporting country [i.e., the home market]." 19 U.S.C. § 1677b(a)(l)(B)(i). 19 U.S.C. § 1677b(a)(1)(B)(i) dictates that Commerce make a fair comparison between the export price and normal value, and specifically, that the comparison between export price and normal value be made at the same level of trade. See also Hyundai Steel Company v. United States, 41 CIT ----, ----, 279 F.Supp.3d 1349, 1356 (2017) ; Pasta Zara SpA v. United States, 34 CIT 355, 369, 703 F.Supp.2d 1317, 1329 (2010).
Section 1677f-1(d)(1) and 19 C.F.R. § 351.414(b) describe three methods by which Commerce may compare the normal value to the export price: (1) average-to-average ("A-to-A"), a comparison of weighted-average normal values to weighted-average export prices for comparable merchandise; (2) transaction-to-transaction ("T-to-T"), a comparison of normal values based on individual transactions to the export prices of individual transactions for comparable merchandise; and (3) average to transaction ("A-to-T"), a comparison of weighted-average normal values to the export prices of individual transactions for comparable merchandise. Commerce "will use the average-to-average method unless [Commerce] determines another method is appropriate in a particular case." 19 C.F.R. § 351.414(c)(1) ; see *1356Stanley Works (Langfang) Fastening Systems Co., Ltd. v. United States, 41 CIT ----, ----, 279 F.Supp.3d 1172, 1177-78 (2017). Commerce may use the A-to-T method if "(i) there is a pattern of export prices ... for comparable merchandise that differ significantly among purchasers, regions, or periods of time, and (ii) [Commerce] explains why such differences cannot be taken into account using a method described in paragraph (1)(A)(i) [A-to-A] or (ii) [T-to-T]." 19 U.S.C. § 1677f-1(d)(1)(B) ; see Stanley Works, 279 F.Supp.3d at 1176-77.
When either necessary information is not available on the record or a respondent (1) withholds information that has been requested by Commerce, (2) fails to provide such information by Commerce's deadlines for submission of the information or in the form and manner requested, (3) significantly impedes an antidumping proceeding, or (4) provides information that cannot be verified, then Commerce shall "use the facts otherwise available in reaching the applicable determination." 19 U.S.C. § 1677e(a). This subsection thus provides Commerce with a methodology to fill informational gaps when necessary or requested information is missing from the administrative record. See Nippon Steel Corp. v. United States, 337 F.3d 1373, 1381 (Fed. Cir. 2003) ; Hyundai Steel, 279 F.Supp.3d at 1355. Commerce "may use an inference that is adverse to the interests of that party in selecting from among the facts otherwise available" if it "finds that an interested party has failed to cooperate by not acting to the best of its ability to comply with a request for information[.]" 19 U.S.C. § 1677e(b)(1)(A). A respondent's failure to cooperate to "the best of its ability" is "determined by assessing whether [it] has put forth its maximum effort to provide Commerce with full and complete answers to all inquiries." Nippon Steel, 337 F.3d at 1382.
II. Factual and Procedural Background.
Commerce initiated a less than fair value investigation regarding certain carbon and alloy steel CTL plate from countries including France. Certain Carbon and Alloy Steel Cut-To-Length Plate From Austria, Belgium, Brazil, France, the Federal Republic of Germany, Italy, Japan, the Republic of Korea, the People's Republic of China, South Africa, Taiwan, and the Republic of Turkey: Initiation of Less-Than- Fair-Value Investigations, 81 Fed. Reg. 27,089 (Dep't Commerce May 5, 2016), P.R. 43. Dillinger was chosen as one of the mandatory respondents2 from France. Certain Carbon and Alloy Steel Cut-To- Length Plate From France: Preliminary Determination of Sales at Less Than Fair Value and Postponement of Final Determination, 81 Fed. Reg. 79,437 (Dep't Commerce Nov. 14, 2016), P.R. 372, and accompanying Preliminary Decision Memorandum ("PDM") at 2, P.R. 364. In *1357its initial questionnaire, Commerce requested affiliate sales information, including, if the sales were not at arm's length, the affiliate's sales to unaffiliated customers. Initial Antidumping Duty Questionnaire at B-5 and 6 (May 31, 2016), P.R. 84. In its Section A questionnaire response, Dillinger reported three channels of distribution for its home market sales: (1) direct sales from its factory, (2) sales through affiliated service centers, and (3) sales of non-prime merchandise. Letter from deKieffer & Horgan, PLLC to Sec'y Commerce, re: Certain Carbon and Alloy Steel Cut-To-Length Plate from France; Dillinger France S.A. Section A Resp. (June 29, 2016) at A-12 & 13, C.R. 70-83, P.R. 154-52. As part of its response to supplemental questionnaires issued by Commerce, Dillinger also submitted a "selling activities chart," in which it identified the selling activities undertaken for each channel of distribution and the degree of involvement in each activity. Letter from deKieffer & Horgan, PLLC to Sec'y Commerce, re: Certain Carbon and Alloy Steel Cut-To-Length Plate from France; Dillinger France S.A. and Berg Steel Pipe Corp; Second Suppl. Sections A, B & C Resp. (Oct. 21, 2016) at 2-7, C.R. 362-63, P.R. 343.
Dillinger initially informed Commerce that it had difficulties collecting and reporting certain information on downstream3 sales made by Dillinger's home market service center Eurodécoupe, SAS ("Eurodécoupe"). Letter from deKieffer & Horgan, PLLC to Sec'y Commerce (June 8, 2016) at 1-3, C.R. 35, P.R. 102 ("June 8 Letter"). However, upon further request by Commerce, Dillinger was able to identify all of Eurodécoupe's home market sales by searching Eurodécoupe's sales and inventory records, including sales prices for all transactions, but was unable to identify the manufacturer of plate sold in every transaction. Third Suppl. Sections B & C Resp. (Nov. 2, 2016) at 2-5, C.R. 374, P.R. 358.
Commerce published its preliminary determination on November 14, 2016, and amended this determination on December 2, 2016 in light of ministerial errors. PDM; Certain Carbon and Alloy Steel Cut-to-Length Plate From France: Amended Preliminary Determination of Sales at Less Than Fair Value, 81 Fed. Reg. 87,019 (Dec. 2, 2016), P.R. 390. In its preliminary determination, Commerce found that there was one level of trade for Dillinger's home market sales. PDM at 12. Commerce also applied facts available for the Eurodécoupe transactions in which the plate manufacturer remained unidentified by attributing all sales to Dillinger and consequently using these transactions in its dumping margin calculation. Id. at 10. Commerce applied its differential pricing method to calculate Dillinger's dumping margin, and found that the use of its A-to-T method was appropriate in this case. Id. at 4-7. Additionally, when evaluating Dillinger's cost of production, Commerce allocated costs between prime and non-prime plate according to values recorded in Dillinger's normal books and records. Id. at 15.
Following the submission of case and rebuttal briefs, Commerce published its final determination on April 4, 2017, and, in response to ministerial error comments, published an amended final determination on May 15, 2017 in which it calculated a final antidumping margin of 6.15 percent for Dillinger. Final Determination. In its *1358IDM, Commerce continued to find that there was one level of trade for Dillinger's home market sales, to apply the A-to-T differential pricing method, and to allocate costs between prime and non-prime plate according to values recorded in Dillinger's normal books and records. IDM at 12-13. Commerce also applied partial adverse facts available ("AFA") for the Eurodécoupe transactions in which the plate manufacturer remained unidentified. Rather than using the reported sales price information for these transactions, Commerce used "the highest non-aberrational net price among Dillinger France's downstream home market sales, and assigned that price to all of these sales where Dillinger France failed to report the manufacturer of the CTL plate in [Commerce's] margin calculations for the final determination." Id. at 47.
Dillinger initiated this action challenging Commerce's Final Determination on June 23, 2017, Summ., ECF No. 1, and filed its complaint on July 24, 2017, Compl., ECF No. 8. This court granted Nucor's consent motion to intervene as defendant-intervenor on August 24, 2017. Order, ECF No. 20. Dillinger filed its Motion for Judgment on the Agency Record on December 18, 2017. Pl.'s Br. The Government submitted its Response in Opposition to Plaintiff's Rule 56.2 Motion on February 16, 2018. Def.'s Br. Nucor filed its response brief, and subsequently its revised response brief, on February 16, 2018 and February 20, 2018, respectively. Def.-Inter.'s Br., ECF Nos. 32-33, 38-39. Dillinger submitted its reply on March 26, 2018. Pl.'s Reply, ECF Nos. 42-43. This court heard oral argument on October 2, 2018. ECF No. 49.
JURISDICTION AND STANDARD OF REVIEW
The court has jurisdiction over this action pursuant to 28 U.S.C. § 1581(c) and 19 U.S.C. § 1516a(a)(2)(A)(i)(I) and (a)(2)(B)(iii). The standard of review in this action is set forth in 19 U.S.C. § 1516a(b)(l)(B)(i) : "[t]he court shall hold unlawful any determination, finding or conclusion found ... to be unsupported by substantial evidence on the record, or otherwise not in accordance with law."
DISCUSSION
I. Commerce's Level of Trade Determination.
Noting that in the context of making a "fair comparison" between the export price and normal value, the Act specifically directs that the comparison be made at the same level of trade, 19 U.S.C. § 1677b(a)(1)(B)(i), see supra, p. 3, Dillinger argues that Commerce violated this provision when it collapsed all of Dillinger's reported home market and United States channels of distribution into one level of trade. In particular, Dillinger contends that Commerce's level of trade determination is unsupported by substantial evidence because it failed to recognize a separate level of trade of sales made through affiliated service centers. Pl.'s Br. at 4. In its view, the record established that Dillinger's "affiliated service centers perform the functions normally performed by distributors and therefore sales by these service centers are at a more remote level of trade than direct sales by Dillinger as a producer." Pl.'s Br. at 5. Dillinger notes that, under Commerce's regulations, different levels of trade represent separate marketing stages, such as direct sales by producers to customers and sales by resellers or distributors. Pl.'s Br. at 4-5 (citing 19 C.F.R. § 351.412(c)(2) ; Preamble to Antidumping Duties; Countervailing Duties, 62 Fed. Reg. 27,296, 27,371 (May 19, 1997) ("Preamble") ). Dillinger states that record evidence shows that the affiliated service centers function as distributors by warehousing plate and selling it to downstream *1359customers in smaller lot sizes or cut to order and that the affiliated centers have their own sales, logistics, and production personnel. Pl.'s Br. at 5-6 (citing Letter from Judith L. Holdsworth, deKieffer & Horgan, to Sec'y of Commerce (June 8, 2016) at 1-3, C.R. 35, P.R. 102; Second Suppl. Sections A, B, & C Resp. (Oct. 21, 2016) at App. SA-15, p. 7, C.R. 362. Therefore, according to Dillinger, the service centers exist at a different marketing stage than the direct sales, and Commerce's decision that Dillinger's selling activities were at only one level of trade was in "direct conflict" with the Preamble and unsupported by substantial evidence in the record.
Substantial evidence is "more than a mere scintilla," but "less than the weight of the evidence." Altx, Inc. v. United States, 370 F.3d 1108, 1116 (Fed. Cir. 2004). "A finding is supported by substantial evidence if a reasonable mind might accept the evidence as sufficient to support the finding." Maverick Tube Corp. v. United States, 857 F.3d 1353, 1359 (Fed. Cir. 2017) (citing Consol. Edison Co. of N.Y. v. NLRB, 305 U.S. 197, 229, 59 S.Ct. 206, 83 L.Ed. 126 (1938) ). "The substantiality of evidence must take into account whatever in the record fairly detracts from its weight." CS Wind Vietnam Co. v. United States, 832 F.3d 1367, 1373 (Fed. Cir. 2016). This includes "contradictory evidence or evidence from which conflicting inferences could be drawn." Suramerica de Aleaciones Laminadas, C.A. v. United States, 44 F.3d 978, 985 (Fed. Cir. 1994) (quoting Universal Camera Corp. v. NLRB, 340 U.S. 474, 487, 71 S.Ct. 456, 95 L.Ed. 456 (1951) ).
The court finds that Commerce's level of trade determination was supported by substantial evidence. Pursuant to Commerce's regulations, sales are made at different levels of trade if they are made at different marketing stages. 19 C.F.R. § 351.412(c)(2). "Substantial differences in selling activities are a necessary, but not sufficient, condition for determining that there is a difference in the stage of marketing." Id. Commerce typically divides selling activities into four categories: 1) sales and marketing; 2) freight and delivery; 3) inventory maintenance and warehousing; and 4) warrant and technical support. See Hyundai Steel, 279 F.Supp.3d at 1369 ; Certain Orange Juice from Brazil, 75 Fed. Reg. 50,999 (Dep't Commerce Aug. 18, 2010) and accompanying IDM at cmt. 7 (dividing selling functions into the four categories).
Based on the evidence in the record, Commerce reasonably concluded that the sales activities of the affiliated service centers did not differ substantially enough to merit a separate level of trade. IDM at 41-42. Dillinger reported two selling functions performed by factories that the affiliated service centers did not -- rebates and personnel training -- while the affiliated service centers performed one selling function -- inventory maintenance -- which Dillinger's factories did not. Id. at 40-42, 42 n.133; Selling Functions Chart. Commerce determined that these differences in selling activities were not substantial enough to warrant separate levels of trade. IDM at 42; Pasta Zara SpA v. United States, 781 F.Supp.2d 1297, 1301 ("Commerce permissibly determined on the record as a whole that Zara's selling activities directed to its local customers were not so separate from Zara's other selling activities."); Hyundai Steel, 279 F.Supp.3d at 1370 ("Commerce reasonably determined that the differences here were not substantial. According to evidence in the record, overall, only two out of the sixteen selling functions -- cash discounts and direct guarantees -- provided in the home market were not provided in the U.S. market.");
*1360Sucocitrico Cutrale Ltda. v. United States, 36 CIT ----, ----, 2012 WL 2317764, at *6 (2012) ("Although Commerce noted minor differences between the two markets, these differences to not rise to the level required by the statute.").
Dillinger argues that Commerce impermissibly did not account for differences in "sawing and cutting" functions between the factories and affiliated service centers, and cites Pasta Zara SpA, 703 F.Supp.2d 1317 (2010) for support. Pl.'s Br. at 9-10. That case is inapposite. There, Commerce failed to consider packing, advertising, and freight expenses in its level of trade analysis because those expenses had been accounted for elsewhere in its dumping calculation; however, as the court explained, because those are selling activities under Commerce's regulations, they must be considered as part of the level of trade analysis. Pasta Zara, 703 F.Supp.2d at 1327-28 ; see supra, p. 10 (discussing the four categories of selling activities which include marketing, freight, and delivery). In contrast, here Commerce reasonably considered "cutting and sawing" to be product processing, and not a selling activity, and thus did not include cutting and sawing as a factor in its level of trade analysis. IDM at 42.
Finally, Dillinger contends that Commerce's determination was not in accordance with law because Commerce's conclusion that Dillinger's affiliated service centers and factories operated at the same level of trade was inconsistent with Commerce's typical treatment of affiliated service centers. Pl.'s Br. at 7-8 (citing Notice of Final Determination of Sales at Less than Fair Value: Certain Cold-Rolled Steel Flat Products from the United Kingdom, 81 Fed. Reg. 49,929 (July 29, 2016) and accompanying IDM at cmt. 1, p. 8 & n.21; Notice of Final Determination of Sales at Less than Fair Value: Structural Steel Beams from Germany, 67 Fed. Reg. 35,497 (May 20, 2002) and accompanying IDM at cmt. 3; Stainless Steel Bar from Germany: Final Results of Anitdumping Duty Administrative Review, 69 Fed. Reg. 32,982 (June 14, 2004) and accompanying IDM at cmt. 1).
Commerce has also come to the opposite conclusion in prior cases. See, e.g., Non-Oriented Electrical Steel From the Republic of Korea: Final Determination of Sales at Less than Fair Value and Negative Final Determination of Critical Circumstances, 79 Fed. Reg. 61,612 (Dep't Commerce Oct. 14, 2014) and accompanying IDM at 8-15 ("We disagree ... that the selling activities associated with the reseller channel of distribution are significantly different to warrant a[ ] [level of trade] designation separate from POSCO's direct sales to unaffiliated customers."); Light-Walled Rectangular Pipe and Tube From Mexico: Preliminary Results and Partial Rescission of Antidumping Administrative Review 2011-2012, 78 Fed. Reg. 54,864 (Dep't Commerce Sept. 6, 2013) at 8-9 (finding one level of trade where the respondent sold sales either directly or through an affiliated reseller), unchanged in final results, 79 Fed. Reg. 5375 (Dep't Commerce Jan. 31, 2014) ; see also Corus Staal BV v. U.S. Dep't Commerce, 27 CIT 388, 401, 259 F.Supp.2d 1253, 1271 (2003) ("Commerce has not explained whether it generally finds that producers and service centers operate at the same LOT -- but there is no requirement that it do so. Rather than broadly differentiating between types of sellers, Commerce compares the observations of different selling functions, which is consistent with statute, regulation, and SAA.").
Moreover, the proceedings Dillinger cites in support of its position are distinct from this case. In Cold-Rolled Steel Flat Products from the United Kingdom, there were "significant differences in numerous selling functions," rather than the minor differences present here. See Cold-Rolled *1361Steel IDM at cmt. 1. The conclusions in Structural Steel Bar from Germany and Stainless Steel Bar from Germany were based on a level of trade analysis method no longer in use by Commerce in which it did consider further processing. See Steel Beams IDM at cmt. 3; Steel Bar IDM at cmt. 1. Those proceedings are thus not reflective of Commerce's current typical level of trade practice.4 Therefore, Commerce's level of trade determination is supported by substantial evidence and in accordance with law.
II. Application of Partial Adverse Facts Available to Service Center Downstream Sales.
Dillinger contends that Commerce's application of partial adverse facts available ("AFA") to the downstream sales of its affiliated service centers was not supported by substantial evidence because (1) Dillinger put forth its best efforts to provide the price and manufacturer data requested by Commerce, and (2) for transactions where the manufacturer data was unknown but the sales price was contained in the record, Commerce impermissibly replaced the record sales prices with the highest non-aberrational net price among Dillinger France's downstream home market sales. The court determines that substantial evidence supports Commerce's conclusion that Dillinger did not put forth best efforts to provide the manufacturer data for all downstream service center transactions, and thus permissibly resorted to partial AFA. However, the court concludes that Commerce did not adequately justify its decision to ignore existing record price data and replace this record evidence with the highest non-aberrational net price. Accordingly, the court remands to Commerce for reconsideration of this decision.
As previously discussed, when either necessary information is not available on the record, or a respondent (1) withholds information that has been requested by Commerce, (2) fails to provide such information by Commerce's deadlines for submission of the information or in the form and manner requested, (3) significantly impedes an antidumping proceeding, or (4) provides information that cannot be verified, then Commerce shall "use the facts otherwise available in reaching the applicable determination." 19 U.S.C. § 1677e(a). This subsection thus provides Commerce with a methodology to fill informational gaps when necessary or requested information is missing from the administrative record. See Nippon Steel, 337 F.3d at 1381. Commerce "may use an inference that is adverse to the interests of that party in selecting from among the facts otherwise available", if it "finds that an interested party has failed to cooperate by not acting to the best of its ability to comply with a request for information[.]" 19 U.S.C. § 1677e(b)(1)(A). A respondent's failure to cooperate to "the best of its ability" is "determined by assessing whether [it] has put forth its maximum effort to provide Commerce with full and complete answers to all inquiries." Nippon Steel, 337 F.3d at 1382.
Dillinger argues that it did indeed put forth its maximum effort to provide the information on Eurodécoupe's downstream sales. Pl.'s Br. at 10-19. Dillinger notes that it initially alerted Commerce that it would have difficulty fully reporting all the requested information; however, when Commerce continued to request that information, Dillinger asserts that it provided *1362as much as it could through a manual search of Eurodécoupe's records and through reviewing the mill certificates for individual transactions. Pl.'s Br. at 10-12; Third Suppl. Sections B & C Resp. at 2-5. According to Dillinger, it was able to report "the correct price for each and every transaction. The only information that was missing for a small number of transactions was the manufacturer of some of the plate." Pl.'s Br. at 12. In its submissions, Dillinger explained that plate from various manufacturers is co-mingled in Eurodécoupe's warehouse and that the manufacturer information on individual pieces that have been further processed and put back into inventory is not always recorded. See Service Center Reporting Letter at 1-3, C.R. 35, P.R. 102; Third Suppl. Sections B & C Resp. at 2-3. Therefore, according to Dillinger, evidence on the record suggests Dillinger put forth its best efforts and Commerce's decision to apply partial AFA was not supported by substantial evidence.
The Government argues that Commerce's decision to apply partial AFA was supported by substantial evidence. First, the Government asserts that, because Dillinger was able to provide mill certificates for some transactions, and "Dillinger presumably would not know whether its service center customer requires mill certificates until the time of sale," there is no evidence on the record that the identity of the plate sold to customers in a particular transaction was unavailable to Dillinger. Def.'s Br. at 23. Nucor elaborates, suggesting that because "[t]here is nothing on the record indicating that mill certificates are missing for any transaction for which no manufacturer was identified or that Dillinger undertook any effort to find mill certificates for those transactions," Commerce's determination is supported by substantial evidence. Def.-Inter.'s Br. at 20; see IDM at 47 (finding that Dillinger had demonstrated familiarity with Eurodécoupe's systems by providing requested information about some sales transactions, and that, therefore, Dillinger should have been able to report information for all sales transactions). Second, the Government contends that "[t]he identity of the manufacturers of CTL plate resold by Dillinger's affiliate, Eurodécoupe, is the type of information that an importer reasonably should 'anticipate being called upon to produce' and that a large steel producer, such as Dillinger, should be able to provide after 'conduct[ing] prompt, careful, and comprehensive investigations.' " Def.'s Br. at 22 (citing IDM at 47; Nippon Steel, 337 F.3d at 1382 ).
The court concludes that Commerce's reliance on the absence of mill certificates for the transactions with missing manufacturer data to support its determination that Dillinger could access the missing information is unreasonable and unsupported by substantial evidence on the record. As discussed supra, Dillinger did detail the thorough efforts it undertook to provide the mill certificates for transactions with missing manufacturer data and provided mill certificates available to it.5 It is unclear how the presence of mill certificates Dillinger indicates it could provide and the absence of mill certificates which Dillinger *1363claims were unavailable to it substantially support the conclusion that Dillinger did have access to the missing information.
However, Commerce's other reason for applying partial AFA -- that an importer like Dillinger should have been aware of the need to record and provide manufacturer information on downstream sales -- is supported by substantial evidence. As the Federal Circuit opined in NipponSteel, which Commerce cited as the basis for its decision:
Before making an adverse inference, Commerce must examine respondent's actions and assess the extent of respondent's abilities, efforts, and cooperation in responding to Commerce's requests for information. Compliance with the "best of its ability" standard is determined by assessing whether respondent has put forth its maximum effort to provide Commerce with full and complete answers to all inquiries in an investigation. While the standard does not require perfection and recognizes that mistakes sometimes occur, it does not condone inattentiveness, carelessness, or inadequate record keeping. It assumes that importers are familiar with the rules and regulations that apply to the import activities undertaken and requires that importers, to avoid a risk of an adverse inference determination in responding to Commerce's inquiries: (a) take reasonable steps to keep and maintain full and complete records documenting the information that a reasonable importer should anticipate being called upon to produce; (b) have familiarity with all of the records it maintains in its possession, custody, or control; and (c) conduct prompt, careful, and comprehensive investigations of all relevant records that refer or relate to the imports in question to the full extent of the importers' ability to do so.
Nippon Steel, 337 F.3d at 1382 (emphasis added). In light of Nippon Steel' s requirements, Commerce's determination that "[t]he information in question (i.e., the identity of the manufacturers of the CTL plate at issue that was resold by Dillinger France's affiliate, [Eurodécoupe] ) is the type of information that a large steel manufacturer such as Dillinger France should reasonably be able to provide," and that Dillinger's failure to have this information available merited the application of partial AFA, was supported by substantial evidence. IDM at 46-47.
Dillinger argues that "when, as in this case, a respondent notifies the Department of its difficulties in providing requested information pursuant to 19 U.S.C. § 1677m(c), the Department may not impose adverse facts available unless it has first responded to the 'overtures of cooperation from the exporter/producer.' " Pl.'s Br. at 18 (quoting World Finer Foods v. United States, 24 CIT 541, 544-45, 2000 WL 897752 (2000) ). According to Dillinger, because Commerce did not respond to Dillinger's letter notifying Commerce of Dillinger's difficulty identifying the manufacturers for all transactions, Commerce cannot apply AFA.
However, Dillinger here did not meet all of the criteria under § 1677m(c).6 Before *1364Commerce is required to respond, interested parties must explain why they have difficulty providing the requested information and offer an alternative form of the information. 19 U.S.C. § 1677m(c)(1) ; World Finer Foods, 24 CIT at 543-45 (holding that because the plaintiff notified Commerce of its difficulties reporting the information and "offer[ed] to supply any 'limited information that Commerce felt might be worthwhile or helpful,' " Commerce was required to provide the plaintiff with guidance). While Dillinger did notify Commerce of its difficulties supplying the requested manufacturer information, it did not offer any alternative form of information that Commerce could use instead pursuant to § 1677m(c)(1). See June 8 Letter. Therefore, § 1677m(c) does not apply here.
The court next considers how Commerce applied partial AFA. The Government contends that, "[a]s a partial adverse facts available rate, Commerce reasonably assigned the highest non-aberrational net price among Dillinger's downstream home market sales to the sales where Dillinger failed to report the manufacturer of CTL plate in its margin calculations." Def.'s Br. at 23-24; IDM at 47. The court is unpersuaded by this argument. Although, as discussed above, 19 U.S.C. § 1677e permits Commerce to use adverse facts available to fill in information gaps, that is not what Commerce did in this case. Here, Commerce replaced known, unchallenged record information -- the sales price recorded in transactions where the CTL plate's manufacturer was unknown -- with adverse facts available. IDM at 47. In the IDM, Commerce did not explain what authority permitted it to replace known information with adverse facts available. Furthermore, Commerce did not explain how it determined the "highest non-aberrational price."
In its brief, the Government cites to Acciai Speciali Terni S.P.A. v. United States, 25 CIT 245, 142 F.Supp.2d 969, 994 (2001), in support of Commerce's decision to utilize the highest non-aberrational net price among Dillinger's downstream home market sales. In that case, the court found that Commerce reasonably determined that the price data on the record was unreliable due to numerous errors and inconsistencies in the price data provided, and thus application of AFA to fill the gap in reliable price data was appropriate. Id. at 987-89. In the instant case, however, the reliability of the reported sales prices has not been called into question and there is no informational gap in the sale prices for Commerce to fill. Therefore, Commerce's decision to apply partial AFA to replace information in the record was not supported by substantial evidence and was contrary to law, and the court remands this issue to Commerce for reconsideration of its decision.
III. Commerce's Application of the Differential Pricing Methodology.
Dillinger contends that Commerce's use of its differential pricing methodology ("DPM") was not supported by substantial evidence and was contrary to law. Specifically, Dillinger argues that: (1) the World Trade Organization's ("WTO") decision in United States - Antidumping and Countervailing *1365Measures on Large Residential Washers from Korea (WT/DS464/AB/R) adopted Sept. 26, 2016 ("Korean Washers") should prompt this court to reconsider whether Commerce's DPM is a reasonable interpretation of the statute; (2) Commerce's DPM fails to show that Dillinger's prices differed significantly among purchasers, regions, or periods of time; and (3) Commerce's use of zeroing as part of its DPM is impermissible under the statute.7 The court concludes that Korean Washers is not controlling and that Commerce's application of its DPM here is supported by substantial evidence and in accordance with law.
A. Legal and Statutory Framework of Differential Pricing.
When calculating a dumping margin, Commerce may compare the weighted average of the normal values8 to the weighted average of the export prices9 (or constructed export prices10 ) for comparable merchandise or compare the normal values of individual transactions to the export prices (or constructed export prices) of individual transactions for comparable merchandise. 19 U.S.C. § 1677f-1(d)(1) ; see also Stanley Works, 279 F.Supp.3d at 1176. As discussed above, Section 1677f-1(d)(1) and 19 C.F.R. § 351.414(b) describe three methods by which Commerce may compare the normal value to the export price: (1) average-to-average ("A-to-A"), a comparison of weighted-average normal values to weighted-average export prices for comparable merchandise; (2) transaction-to-transaction ("T-to-T"), a comparison of normal values based on individual transactions to the export prices of individual transactions for comparable merchandise; and (3) average to transaction ("A-to-T"), a comparison of weighted-average normal values to the export prices of individual transactions for comparable merchandise. Commerce "will use the average-to-average method unless [Commerce] determines another method is appropriate in a particular case." 19 C.F.R. § 351.414(c)(1) ; see *1366Stanley Works, 279 F.Supp.3d at 1177-78. Commerce may use the A-to-T method if "(i) there is a pattern of export prices ... for comparable merchandise that differ significantly among purchasers, regions, or periods of time, and (ii) [Commerce] explains why such differences cannot be taken into account using a method described in paragraph (1)(A)(i) [A-to-A] or (ii) [T-to-T]." 19 U.S.C. § 1677f-1(d)(1)(B) ; see Stanley Works, 279 F.Supp.3d at 1176-77. Section 1677f-1(d) does not prescribe how Commerce should determine whether a pattern of prices that differ significantly exists.
Commerce's "differential pricing analysis consists of three tests, segregated into two stages." Stanley Works, 279 F.Supp.3d at 1178 (citing Apex Frozen Foods Private Ltd. v. United States, 862 F.3d 1337, 1342 n.2 (Fed. Cir. 2017) ). "In the first stage, Commerce utilizes two tests to determine whether there exists a pattern of prices that differ significantly, such that an alternative comparison method should be considered, pursuant to 19 U.S.C. § 1677f-1(d)(1)(B)(i)." Id. The first test is a "statistical measure of the extent of the difference between the mean of a test group and the mean of a comparison group" known as the Cohen's d test. Id. (internal quotations omitted). The Cohen's d test quantifies "the extent to which the net prices to a particular purchaser, region or time period differ significantly from the net prices of all other sales of comparable merchandise." Id. (internal citations omitted). If the Cohen's d coefficient is .8 or greater, the net price to a particular purchaser, region, or time period "passes" the test. Id. at 1179.
Commerce next applies the ratio test, which assesses the extent of the significant price differences for all sales as measured by the Cohen's d test. Id.; PDM at 6. "If the value of sales to purchasers, regions, and time periods that pass the Cohen's d test account for 66 percent or more of the value of total sales, then the identified pattern of prices that differ significantly supports the consideration of the application of the A-T method to all sales as an alternative to the A-A method." Stanley Works, 279 F.Supp.3d at 1179 ; see PDM at 6.
"If Commerce determines that both of these tests demonstrate the existence of a pattern of prices that differ significantly enough to warrant consideration of an alternative comparison method, then Commerce proceeds to the second stage of the differential pricing analysis, in which it examines whether using only the A-A method can appropriately account for those differences, pursuant to 19 U.S.C. § 1677f-1(d)(1)(B)(ii)." Stanley Works, 279 F.Supp.3d at 1179 ; see PDM at 6. At this second stage, the meaningful difference analysis, "[a] difference in the weighted-average dumping margins is considered meaningful if (1) there is a 25 percent relative change in the weighted-average dumping margin between the A-A method and the appropriate alternative method where both rates are above the de minimis threshold, or (2) the resulting weighted-average dumping margin moves across the de minimis threshold." Stanley Works, 279 F.Supp.3d at 1179.
B. Korean Washers.
As Dillinger acknowledges, this court and the Federal Circuit have upheld various aspects of Commerce's DPM.11
*1367Pl.'s Br. at 22. Nonetheless, Dillinger encourages this court to "review these decisions in light of the persuasive authority of Korean Washers." Id. In that case, the Appellate Body of the WTO found that the DPM was "inconsistent 'as such' with Article 2.4.2 of the Anti-Dumping Agreement,"12 which closely resembles 19 U.S.C. § 1677f-1(d)(1). Korean Washers . Dillinger notes that at a meeting of the WTO Dispute Settlement Body, the United States stated its intention to implement the Korean Washers decision, and that the deadline for the United States to bring its antidumping methodology into conformity with its WTO obligations was December 26, 2017. Korean Washers, Status Report, WT/DS464/17 (Nov. 10, 2017). According to Dillinger, once Commerce changes its practice to comply with WTO obligations, it must adhere to the new practice consistently, regardless of any existing U.S. court decisions upholding Commerce's prior practice. Pl.'s Br. at 21 (citing Dongbu Steel Co. v. United States, 635 F.3d 1363, 1371-73 (Fed. Cir. 2011) ). Dillinger further suggests that, because the Charming Betsy Doctrine indicates that U.S. statutes should never be interpreted to conflict with international obligations absent express Congressional language to the contrary, this court should find Korean Washers particularly persuasive. Id. (citing Murray v. Schooner Charming Betsy, 6 U.S. (2 Cranch) 64, 81, 2 L.Ed. 208 (1804) ).
However, as Dillinger acknowledges, id. at 22, WTO decisions do not directly bind Commerce or this court. Corus Staal BV v. Dep't of Commerce, 395 F.3d 1343, 1348 (Fed. Cir. 2005) (citing Timken v. United States, 354 F.3d 1334, 1344 (Fed. Cir. 2004) ); The Stanley Works (Langfang) Fastening Sys. Co. v. United States, 42 CIT ----, ----, Slip Op. 18-99 (Aug. 13, 2018) at 50 ("WTO decisions are irrelevant to the interpretation of domestic U.S. law."). Indeed, WTO decisions are not self-executing under U.S. law; they can only be implemented through a prescribed statutory procedure. See, e.g., SAA, H.R. Rep. No. 103-316 at 656, 659 ("WTO dispute settlement panels will have no power to change U.S. law or order such a change. Only Congress and the Administration can decide whether to implement a WTO panel recommendation and, if so, how to implement it."); 19 U.S.C. § 3538 ; see also Stanley Works, Slip. Op. 18-99 at 50-51. Indeed, Congress "expressly designed [its implementation procedure] so as to preserve the independence of U.S. law from adverse decisions of the [WTO] until such time as the political branches decide that, of the options available to the United States under the WTO Agreements, a change in U.S. law and/or policy or methodology is most appropriate." Andaman Seafood Co. v. United States, 34 CIT 129, 140, 675 F.Supp.2d 1363, 1373 (2010) (citing 19 U.S.C. § 3538 ). Put another way, "[i]ssues brought before WTO panels and the Appellate Body deal with whether a country is complying with the terms of the WTO Agreement," whereas "[c]ases brought before the Court of International Trade present questions dealing with domestic U.S. law." Stanley Works, Slip Op. 18-99 at 51 (internal citations omitted). Because "Commerce's interpretation of a statute might well be a perfectly reasonable interpretation of U.S. law and nonetheless be found to violate the WTO Agreement ... plaintiff's argument that *1368the Appellate Body's decision in Washers from Korea somehow shows that Commerce's interpretation and implementation of the targeted dumping statute is unreasonable under U.S. law is far wide of the mark." Id. Thus, as Korean Washers has not yet been implemented under U.S. law, "[w]e therefore accord no deference to the cited WTO case[ ]." Corus Staal, 395 F.3d at 1349.
Additionally, Dillinger's citation to Dongbu Steel is inapposite. In that case, Commerce had changed its practice because an adverse WTO decision had been implemented, and Commerce's application of zeroing became inconsistent between investigations and administrative reviews. Dongbu Steel, 635 F.3d at 1371-72. The Federal Circuit found that Commerce had inadequately explained its inconsistent application. Id. at 1372-73 ; see also Apex, 862 F.3d at 1336. In contrast, Korean Washers has not been implemented, and this court here is "not faced with a conflicting statutory interpretation demanding Commerce's explanation." Apex, 862 F.3d at 1337. For these reasons, the court does not find Korean Washers persuasive here.
C. Pattern of Significant Price Differences Among Purchasers, Regions, and Time.
Dillinger alleges that Commerce failed to find a pattern of significant price differences among purchasers, regions, and time, pursuant to 19 U.S.C. § 1677f-1(d)(1)(B). Specifically, Dillinger contends that Commerce (i) failed to demonstrate a "pattern" of differences; (ii) improperly aggregated price differences across the categories of purchasers, region, and time; and (iii) improperly considered the significance of price differences. The court concludes that Commerce's decision was supported by substantial evidence and in accordance with law.
When evaluating whether Commerce's interpretation and application of a statute is in accordance with law, courts consider "whether Congress has directly spoken to the precise question at issue," and, if not, whether the agency's interpretation of the statute is reasonable. Apex, 862 F.3d at 1344 (quoting Chevron U.S.A, Inc. v. Natural Res. Def. Council, Inc., 467 U.S. 837, 842-43, 104 S.Ct. 2778, 81 L.Ed.2d 694 (1984) ); Stanley Works, 279 F.Supp.3d at 1184. If the statute is silent or ambiguous regarding the relevant issue, the traditional second prong of the Chevron analysis requires courts to "defer to the agency's interpretation of its own statute as long as that interpretation is reasonable." Koyo Seiko Co., Ltd. v. United States, 36 F.3d 1565, 1573 (Fed. Cir. 1994) ; see Kyocera Solar, Inc. v. United States Int'l Trade Comm'n, 844 F.3d 1334 (Fed. Cir. 2016).
19 U.S.C. § 1677f-1 does not prescribe how Commerce is to conduct its targeted dumping analysis, and therefore Commerce's discretionary choice to employ DPM is entitled to deference so long as that methodological choice is reasonable. Stanley Works, 279 F.Supp.3d at 1185 (citing JBF RAK LLC v. U.S., 790 F.3d 1358, 1362 (Fed. Cir. 2015) and Chevron, 467 U.S. at 842-43, 104 S.Ct. 2778 ). "Antidumping ... duty determinations involve complex economic and accounting decisions of a technical nature, for which agencies possess far greater expertise than courts," PSC VSMPO-Avisma Corp. v. United States, 688 F.3d 751, 764 (Fed. Cir. 2012), cited in Apex, 862 F.3d at 1347, and so courts afford Commerce significant deference in those determinations, see id.; Fujitsu Gen. Ltd. v. United States, 88 F.3d 1034, 1039 (Fed. Cir. 1996). See also Stanley Works, 279 F.Supp.3d at 1185. Nonetheless, Commerce "must cogently explain why it has exercised its discretion in a *1369given manner." Motor Vehicle Mfrs. Ass'n of U.S. v. State Farm Mut. Auto. Ins. Co., 463 U.S. 29, 48, 103 S.Ct. 2856, 77 L.Ed.2d 443 (1983) ; see also CS Wind Vietnam Co. v. United States, 832 F.3d 1367, 1377 (Fed. Cir. 2016) ("The requirement of explanation presumes the expertise and experience of the agency and still demands an adequate explanation in the particular matter." (citing Burlington Truck Lines, Inc. v. United States, 371 U.S. 156, 167-68, 83 S.Ct. 239, 9 L.Ed.2d 207 (1962) ) ).
The court concludes that Commerce's application of its DPM in this case is a reasonable exercise of its discretion and that Commerce adequately explained in the IDM its choice of methodology. At the administrative stage, Dillinger expressed its concerns that its made-to-order sales and various other economic considerations meant that Commerce's DPM only detected random variations, rather than a significant pattern of price differences. Dillinger provides no support for its assertion that its made-to-order sales would cause distortions in Commerce's calculations. However, Commerce explained that companies' economic goals were reflected through their pricing behavior, IDM at 21, and that DPM was designed to reveal when companies resorted to targeted dumping, id. at 22. Commerce also detailed why its DPM took into account both higher and lower priced sales:
The SAA states that "targeted dumping" is where "an exporter may sell at a dumped price to particular customers or regions, while selling at higher prices to other customers or regions." For "targeted" or masked dumping to exist, there must be both lower-priced U.S. sales which evidence dumping as well as higher-priced, non-dumped U.S. sales which "conceal," mask, hide this evidence of dumping. Therefore, since the purpose of section 777A(d)(1)(B) is to provide a remedy for "targeted dumping," pursuant to which the Department must satisfy the pattern requirement to demonstrate that the respondent's pricing behavior in the U.S. market exhibits characteristics "where targeted dumping may be occurring," the Department continues to find reasonable and logical its approach of including both lower-priced and higher-priced U.S. sales as part of a potential pattern of prices that differ significantly.
Id. at 22 (internal citations omitted); see Stanley Works, 279 F.Supp.3d at 1191.
Moreover, Commerce specifically addressed Dillinger's concerns about the made-to-order nature of its products and explained that the CONNUMs13 Commerce used in its analysis took into account variation between Dillinger's products.
Dillinger France further asserts that its made-to-order products are inferably so unique and embrace such a wide range of grades within a given product control number (CONNUM) that any comparison of U.S. prices on a CONNUM basis must take into account these inter-CONNUM variations. The Department disagrees. The CONNUM and its constituent physical characteristics are all subject to notification and comment during this investigation. Dillinger France provided comments, and Dillinger France's arguments have been fully considered. The established CONNUMs are the foundation for reporting not only comparison and U.S. market sales, but also Dillinger France's costs of production, and are the basis for comparison of U.S. prices with normal value.
*1370Since the purpose of the differential pricing analysis is to consider whether the A-to-A method is appropriate to calculate Dillinger France's weighted-average dumping margin, and the comparisons on which this calculation is based is defined by CONNUMs, the Department therefore finds that it is appropriate, and reasonable, to use these same CONNUMs as the basis for the comparisons of U.S. prices in the differential pricing analysis.
IDM at 21-22.14
Dillinger further asserts that Commerce's DPM improperly aggregates price differences across the categories of purchasers, region, and time, and that consequently, Commerce "completely failed to establish any pattern of price differences with respect to each category on its own" in contravention of the statute. Pl.'s Br. at 26. In its IDM, Commerce addressed in detail how its DPM evaluated such differences, consistent with the statute:
The Cohen's d test compares the U.S. sale prices sequentially to each purchaser, region, and time period, with all other U.S. sale prices (i.e., the U.S. sales to all other purchasers, regions, or time periods, respectively) of comparable merchandise. What appears to be the concern of Dillinger France, for example with purchasers, is that the U.S. sales to each purchaser may not be evenly distributed across the other two types of groups, regions, and time periods. Thus, Dillinger France posits that the "Department therefore cannot determine whether a price difference is actually due to real differences between purchasers or simply due to the fact that the sales are to purchasers in different regions or during different time periods."
The Department finds that this is neither a flaw in the Cohen's d test and nor a distortion of the results, nor that there are flaws related to the other two groups (i.e., U.S. sales to a particular region that are equally distributed across all purchasers and time periods or U.S. sales in a particular time period that are equally distributed across all purchasers and regions). The one possible distortion that could arise, for example that each purchaser is located in a specific region, is that similar results would occur when comparing prices by purchaser and by region. However, the ratio test does not double-count the sales value when a given U.S. sale price is found to be significantly different by purchaser and region. There is no assumption about correlated distribution of sales between purchasers, regions, or time periods, and indeed a given U.S. sale price may be found to be significantly different by all three categories. Yet the ratio test ensures that any such correlation between purchasers, regions and/or time periods does not distort the results of the test and result in a finding that a larger proportion of the U.S. sale value is at prices which differ significantly.
IDM at 23.
Finally, Dillinger asserts that Commerce improperly considered the significance of price differences because, according to Dillinger, "[t]he significance of differences in export prices can only be seen with respect to their relationship to the development of the normal value and cost of production of comparable merchandise." Pl.'s Br. at 28. Dillinger provides no support for this requirement, and the term "significance" is not defined in the statute. "Commerce is entitled to interpret the statutory language, and the court must defer to that interpretation, so long as it is reasonable." Stanley Works, 279 F.Supp.3d at 1186 (rejecting plaintiff's *1371assertion that Commerce must interpret "significant" as "statistically significant") (internal citations omitted). In its IDM, Commerce adequately explained why its interpretation of "significant" did not take into account the factors suggested by Dillinger: the pattern requirement relates to differences among U.S. prices, and thus normal value and cost of production are not relevant considerations. IDM at 21. For these reasons, Commerce's interpretation of the statute is entitled to deference.15
D. Zeroing.
Dillinger argues that Commerce's use of zeroing "distorts both of the requirements provided in section 1677f-1(d)(1)(B)" because it compares "a non-zeroed margin" with a "zeroed margin" and the difference in result between the A-to-A and A-to-T method is due solely to this asymmetrical use of zeroing. Pl.'s Br. at 31. However, in the Apex line of cases, the Federal Circuit upheld Commerce's use of zeroing in its differential pricing methodology: "[w]e hold that Commerce's meaningful difference analysis -- comparing the ultimate antidumping rates resulting from the A-A methodology, without zeroing; and the A-T methodology, with zeroing -- was reasonable." Apex, 862 F.3d at 1348. The Federal Circuit rejected the very same argument Dillinger raises here, stating:
[W]e find it immaterial whether the A-A and A-T margins would be nearly identical if zeroing were applied evenly or not at all ... The notion that Commerce's chosen methodology is unreasonable because it only measures the effects of zeroing is misplaced ... [D]ifferences revealed by zeroing are not inconsequential or to be ignored ... In other words, the effects of zeroing are precisely what 19 U.S.C. § 1677f-1(d)(1)(B) seeks to address.
Id. at 1349 (citations omitted).
Dillinger also alleges that the exception in § 1677f-1(d)(1)(B) applies only to significant differences for the same product among purchasers, regions, or time periods and does not relate to significant price differences between different products. Pl.'s Br. at 31. The Government argues that, because Dillinger did not raise the issue of applying the A-to-T method between different products at the administrative level, Dillinger failed to exhaust its administrative remedies with respect to this argument. Def.'s Br. at 37. In response, Dillinger contends that its inter-product argument is "part and parcel" of its argument that zeroing distorts the requirements under § 1677f-1(d)(1)(B)(i)-(ii) for application of the A-to-T method. Pl.'s Reply at 17.
The court concludes that Dillinger failed to exhaust its administrative remedies for advancing its inter-product argument. This court "shall, where appropriate, require the exhaustion of administrative remedies." 28 U.S.C. § 2637(d). "The doctrine of exhaustion provides 'that no one is entitled to judicial relief for a supposed or threatened injury until the prescribed administrative remedy has been exhausted.' " Essar Steel, Ltd. v. United States, 753 F.3d 1368, 1374 (Fed. Cir. 2014) (quoting Sandvik Steel Co. v. United States, 164 F.3d 596, 599 (Fed. Cir. 1998) ). "Simple fairness to those who are engaged in the tasks of administration, and to litigants, requires as a general rule that courts should not topple over administrative decisions unless the administrative body not only has erred but has erred against objection made at the time appropriate *1372under its practice." Mittal Steel Point Lisas Ltd. v. United States, 548 F.3d 1375, 1383-84 (Fed. Cir. 2008) (quoting United States v. L.A. Tucker Truck Lines, Inc., 344 U.S. 33, 37, 73 S.Ct. 67, 97 L.Ed. 54 (1952) ). "Exhaustion serves two main purposes: 'to allow an administrative agency to perform functions within its special competence-to make a factual record, to apply its expertise, and to correct its own errors,' and to 'promot[e] judicial efficiency by enabling an agency to correct its own errors so as to moot judicial controversies.' " Stanley Works, 279 F.Supp.3d at 1189 (quoting Sandvik Steel, 164 F.3d at 600 ). The issue here -- whether § 1677f-1(d)(1)(B) ever permits Commerce to use the A-to-T method when evaluating significant price differences between products -- implicates both of these concerns; had Dillinger raised its inter-product argument at the administrative level, "Commerce would have had the opportunity to better develop the record and apply its expertise to assess its [practice]." Id. Dillinger's contention that its inter-product argument is merely "part and parcel" of its zeroing allegations does not excuse its failure to raise the inter-product argument at the administrative level. See id. A challenge to one aspect of § 1677f-1(d)(1)(B)'s application "do[es] not incorporate any conceivable challenge to elements of that analysis." Id.
Nor do any of the exceptions to administrative exhaustion apply. Dillinger contends that "[b]ecause Dillinger's argument is based upon a pure question of interpretation of the statute, it would qualify for the exception related to pure legal questions." Pl.'s Reply at 18. "Statutory construction alone is not sufficient to resolve this case." Consol. Bearings Co. v. United States, 348 F.3d 997, 1003 (Fed. Cir. 2003). "Rather, the question is whether the methodology is justifiable, and to resolve that issue, a factual record needs to be developed." Stanley Works, 279 F.Supp.3d at 1190 (citing Consol Bearings, 348 F.3d at 1003 (determining that the pure legal question exception could not apply when the court would have to assess Commerce's justifications for its practice) ); Mittal Steel Point Lisas, 548 F.3d at 1384 (finding the pure question of law exception not applicable when argument relies on unique facts of the case); Fuwei Films (Shandong) Co. v. United States, 35 CIT ----, ----, 791 F.Supp.2d 1381, 1384-85 (2011) (concluding that the pure legal question exception could not apply when the statute at issue did not speak to the required methodology and Commerce's interpretation was needed to fill the statutory gap) ).
Dillinger also argues that the exception for futility applies because, "[i]n all the decades that the Department and the courts have been dealing with the issue of zeroing, the Department has not once found that anything in the antidumping statute prevented it from using zeroing or restricted the application of zeroing in any respect." Pl.'s Reply at 18. "The futility exception to the exhaustion requirement has been applied in situations in which enforcing the exhaustion requirement would mean that parties would be required to go through obviously useless motions in order to preserve their rights." Corus Staal BV v. United States, 502 F.3d 1370, 1378-81 (Fed. Cir. 2007) (internal quotations omitted). However, this court generally takes a strict view of administrative exhaustion, and the futility exception is quite narrow: "The mere fact that an adverse decision may have been likely does not excuse a party from a statutory or regulatory requirement that it exhaust administrative remedies." Id. (internal citations omitted).
Here, "even if it is likely that Commerce would have rejected [Dillinger's] legal and factual showings, it would *1373still have been preferable, for purposes of administrative regularity and judicial efficiency, for [Dillinger] to make its arguments in its case brief and for Commerce to give its full and final administrative response in the final results," id. at 1380, because, as discussed above, Commerce would have had the opportunity to better develop the record and apply its expertise to address Dillinger's concerns. "[A] litigant must diligently protect its rights in order to be entitled to relief." JBF RAK, 790 F.3d at 1367 (quoting Mukand Int'l, Ltd. v. United States, 502 F.3d 1366, 1370 (Fed. Cir. 2007) ).
Moreover, Dillinger's inter-product argument fails on the merits. Dillinger provides no reasoning or authority to support its assertion that it is "unlawful for the Department to apply inter-product zeroing under the guise of applying the exception in section 1677f-1(d)(1)(B)." See Pl.'s Br. at 27, 29, 31. In contrast, as discussed above, Commerce adequately explained how its application of zeroing identifies dumping and is consistent with the statute. As Commerce's interpretation of the statute is reasonable and nothing in the statute dictates how Commerce should conduct its targeted dumping analysis, this court defers to Commerce's methodological choice. See JBF RAK, 790 F.3d at 1362 ; Chevron, 467 U.S. at 842-43, 104 S.Ct. 2778 ; supra, pp. 25-26 (discussing the deference owed to Commerce when its statutory interpretation is reasonable).16
IV. Cost Shifting.
Dillinger challenges Commerce's production cost determination of its prime and sub-prime plates. Pl.'s Br. at 32. Dillinger alleges that Commerce's use of Dillinger's balance sheets to determine production costs, in lieu of the production costs Dillinger calculated, is not in accordance with law. Id. Dillinger further contends that Commerce's finding that prime and sub-prime plates have different applications is unsupported by substantial evidence. Id. at 37. The court holds Commerce's use of Dillinger's balance sheet to be a reasonable methodological choice permitted by the statute, precedent, and Commerce's practice and is thus in accordance with law. The court further finds Commerce's determination that prime and sub-prime plates have different applications to be supported by substantial evidence.
A. Legal Background.
19 U.S.C. § 1677b governs the calculation of costs in determining antidumping duties. The statute defines the constructed value of imported merchandise as "an amount equal to the sum of- - (1) the cost of materials and fabrication or other processing of any kind employed in producing the merchandise, during a period which would ordinarily permit the production of the merchandise in the ordinary course of trade." 19 U.S.C. § 1677b(e). Subsection (f) applies to subsection (e) and outlines the ways in which Commerce shall calculate cost:
(A) In general
Costs shall normally be calculated based on the records of the exporter or producer of the merchandise, if such records are kept in accordance with the generally accepted accounting principles of the exporting country (or the producing country, where appropriate) and reasonably reflect the costs associated with the production and sale of the merchandise. The administering authority shall consider all available evidence on the proper allocation of costs, including that which is made available by the exporter or producer on a timely basis, if such allocations have been historically used *1374by the exporter or producer, in particular for establishing appropriate amortization and depreciation periods, and allowances for capital expenditures and other development costs.
19 U.S.C. § 1677b(f)(1)(A).
B. Use of Balance Sheet.
According to Dillinger, Commerce's use of Dillinger's balance sheet -- which bases the production cost of sub-prime plates on their net recovery price -- to reallocate cost between prime and sub-prime plates contravened the statute and is inconsistent this court's precedent. Pl.'s Br. at 32-33, (citing 19 U.S.C. § 1677b(e) and IPSCO v. United States 965 F.2d 1056 (Fed. Cir. 1999) ). Dillinger argues that section 19 U.S.C. § 1677b(b)(3) requires that "the reported cost of production for a product should correspond to the full costs of producing that product including 'the cost of materials and of fabrication or other processing of any kind employed in producing' the product." Pl.'s Br. at 32 (citing IPSCO, 965 F.2d at 1059 (stating that, "[b]y its terms, the statute expressly covers actual production costs") ). Because the materials and production process are the same, and sub-prime plates can only be differentiated from prime plates after production, Dillinger argues that the production costs of both types of plate are equal under the antidumping statute. Id. at 34. In Dillinger's view, Commerce's use of Dillinger's balance sheet in its cost determination "has the effect of reporting the cost of non-prime plate at less than the actual full cost incurred in producing that merchandise and reporting the cost of production for prime plate at an amount greater than the actual cost incurred in producing that merchandise." Id. at 32. Therefore, because this reallocation resulted in Commerce increasing the cost of prime plates and decreasing the cost of sub-prime plates to reflect their lower market value, Dillinger contends that Commerce violated the statute's requirement to use the actual costs of production. Id. at 32-33. Dillinger claims that Commerce instead should have used the actual production costs Dillinger calculated, which reflect an equal cost allocation between prime and subprime plates. Id. at 32.
The court is unpersuaded by Dillinger's arguments and finds that Commerce's use of Dillinger's balance sheets was in accordance with law. As discussed above, 19 U.S.C. § 1677b(f)(1)(A) provides that "[c]osts shall normally be calculated based on the records of the exporter or producer of the merchandise, if such records are kept in accordance with the generally accepted accounting principles of the exporting country (or the producing country, where appropriate) and reasonably reflect the costs associated with the production and sale of the merchandise." Dillinger's books were kept in accordance with generally accepted accounting principles ("GAAP"), and Commerce adequately explained its reasoning for not deviating from using these records in its IDM:
Further, we note that section 773(f)(1)(A) of the Act mandates that a respondent's costs should be based on the company's normal books and records, if such records are kept in accordance with the GAAP of the exporting country and reasonably reflect the costs associated with the production of the merchandise. In its normal books and records, Dillinger France values non-prime products at their likely selling price and uses this value as an offset to its prime production. This value is significantly lower than the average value of prime production that Dillinger France assigned for purposes of reporting costs to the Department. This lower cost reflects the lower market value and different end uses for non-prime products compared to prime products. As such, we find Dillinger France's normal treatment of non-prime products to be reasonable *1375and consistent with GAAP and the lower of cost or market principle. Therefore, we find no basis for departing from Dillinger France's normal treatment of these products in its books and records, and, consistent with the Preliminary Determination, we continue to adjust Dillinger France's reported costs to reflect Dillinger France's normal treatment of non-prime and prime CTL plate.
IDM at 60.
Tension Steel Indust. Co. v. United States and PSC VSMO-Avisma Corp. v. United States also support Commerce's use of Dillinger's GAAP-consistent books. See Tension Steel Indus. Co. v. United States, 40 CIT ----, 179 F.Supp.3d 1185 (2016) ; PSC VSMPO-Avisma Corp. v. United States, 688 F.3d 751 (Fed. Cir. 2012). In Tension Steel, this court found that "[v]aluing the cost of non-prime pipe at the net recovery price of the product is consistent with GAAP" and thus Commerce "reasonably used the net recovery price to value the non-prime pipe." 179 F.Supp.3d at 1195-96. In PSC VSMPO-Avisma Corp., the Federal Circuit gave Commerce's cost methodology for joint products "tremendous deference," 688 F.3d at 764 (quoting Fujitsu Gen. Ltd. v. United States, 88 F.3d 1034, 1039 (Fed. Cir. 1996) ), and found reasonable Commerce's approach, whereby it tied the value of one joint product to its "real world price," id. at 765.
Dillinger argues that the Federal Circuit's holding in IPSCO precludes this "shifting of costs from non-prime to prime merchandise based upon the market value of the non-prime merchandise." Pl.'s Br. at 33. Dillinger contends that the Federal Circuit held that a methodology that shifted production costs from one non-prime co-product to another prime co-product, below the actual cost of production, was "unreasonable." Id. at 34 (citing IPSCO, 965 F.2d at 1058-59. Dillinger alleges that the situation here is analogous, and "[a]ll products subject to the investigation, whether prime or not, must be reported at their actual cost of production." Id.
IPSCO, however, was decided before Congress amended 19 U.S.C. § 1677b to include subsection (f), which, as discussed supra, specifically allows for Commerce to use GAAP-compliant "books and records" to determine cost of production, so long as they "reasonably reflect costs associated with the production and sale of the merchandise." 19 U.S.C. § 1677b(f)(1)(A). The Federal Circuit, therefore, did not have subsection (f) on which to rely in deciding IPSCO.
IPSCO, moreover, is factually distinguishable from the present case. In IPSCO, the Federal Circuit rejected cost reallocation between prime and limited-use certain oil country tubular goods (OCTG) from Taiwan because the products at issue there had the same applications. IPSCO, 965 F.2d at 1059-61. In the case of both prime and limited-use OCTG, "buyers purchase the separate grades for the same purpose - 'down hole' use in oil and gas wells." Id. at 1058. Therefore, the Federal Circuit concluded that this court's order for Commerce to use a methodology which would decrease the cost of limited-grade OCTG and increase the cost of prime OTCG, despite their same production process and applications, was "unreasonable." Id. at 1061. Here, however, Commerce determined, consistent with its practice,17 *1376that Dillinger's prime and non-prime plate had different applications. IDM at 59-60.
On point is the Federal Circuit's subsequent discussion of IPSCO in Thai Pineapple Pub. Co. v. United States, which stated that IPSCO did not control a case involving "...parts of the pineapple, i.e., the cylinder, core, shells, and ends" with "significantly different uses and values," meaning "they are not interchangeable when it comes to canned pineapple fruit versus juice production." Thai Pineapple Pub. Co. v. United States, 187 F.3d 1362, 1368 (Fed. Cir. 1999). In Thai Pineapple, Thai producers argued that an "alternative methodology was necessary [to determine cost of production] because their financial accounting cost allocations were based on certain managerial and tax goals, and thus, were not reflective of actual production costs." Id. at 1364. Commerce disagreed, rejecting "weight-based raw material fruit cost allocation for both [cost of production] and [constructed value]." Id. Thus, "Commerce relied upon the methodologies reflected in their financial accounting records." Id. This court, however, found that Commerce had erred in relying on the financial accounting records and remanded the case for Commerce to apply "either the Thai producers' weight-based allocation methodologies or a non-output price-based cost allocation methodology." Id. at 1363. The Federal Circuit overturned this decision, finding that "[a]ntidumping investigations are complex and complicated matters in which Commerce has particular expertise and thus, Commerce's determinations are entitled to deference." Id. at 1367 (citing Chevron, 467 U.S. at 844, 104 S.Ct. 2778 ). The Federal Circuit held that "Commerce's reliance on the books and records of [the Thai producers] was reasonable and is supported by substantial evidence." Id. Although Thai Pineapple involved both products included and excluded from the investigation -- unlike the prime and sub-prime pipe both covered by the investigation in IPSCO-- it nonetheless deferred to Commerce in its reliance on books and records in its constructed cost methodology and squarely addressed IPSCO's limited application to co-products with the same post-production application. Because prime and subprime pipes have different post-production applications, IPSCO does not control here. The court, therefore, defers to Commerce's methodological choice to rely on Dillinger's books and records, as such choice was not an unreasonable interpretation of 19 U.S.C. § 1677b, and thus was in accordance with law.
Dillinger alleges that Commerce's finding that its prime and sub-prime plates have different applications was not supported by substantial evidence. Dillinger argues that because both prime and non-prime plates have structural applications, "[t]he fact that the specific uses of non-prime plate are limited to applications that do not require a warranty in no way changes the actual cost of production of non-prime plate or permits a shifting of *1377actual costs from non-prime to prime plate." Pl.'s Br. at 37.
The court is not persuaded by this argument and finds that Commerce's determination was supported by substantial evidence. Commerce considered whether sub-prime plates were a minor downgrade of prime plates so as to remain within the same product group, or whether their application was sufficiently different so as to belong to a different product group. IDM at 60. Commerce reviewed the record of the investigation, as well as Dillinger France's Supplemental Response, and found "non-prime products are sold without certification as to grade, type, or chemistry ... Prime merchandise is used in applications that require these types of certifications," id., such as "as agricultural and construction equipment, bridges, machine parts, ships, and buildings," Def.'s Br. at 39; see International Trade Commission Preliminary Report (May 27, 2016) at I-22, P.R. 318 at 75; Dillinger Suppl. Section D Resp. (Aug. 17, 2016) at 4, P.R. 202; IDM at 60; Dillinger Verification Report, C.R. 678, P.R. 413, "while non-prime merchandise is used in applications such as counterweights for cranes and steel road planks," IDM at 60 (citing Dillinger France's Suppl. Section D Resp. (Aug. 17, 2016), at 4). Based on this record evidence, Commerce determined that prime and sub-prime steel plates have sufficiently different applications to warrant cost reallocation to reflect that the sub-prime plate's value is "significantly impaired to a point where its full cost cannot be recovered." IDM at 60. For these reasons, Commerce's decision was supported by substantial evidence and in accordance with law.
CONCLUSION
With the exception of Commerce's application of partial AFA, the court concludes that the Final Determination is in accordance with law and supported by substantial evidence. The court remands Commerce's partial AFA application for reconsideration consistent with this opinion. Commerce shall file with this court and provide to the parties its remand results within 90 days of the date of this order; thereafter, the parties shall have 30 days to submit briefs addressing the revised final determination to the court and the parties shall have 15 days thereafter to file reply briefs with the court.
SO ORDERED.