This case requires the Court to decide the foreign reach of 15 U. S. C. §1114(1)(a) and §1125(a)(1), two provisions of the Lanham Act that prohibit trademark infringement. The case concerns a trademark dispute between Hetronic (a U. S. company) and six foreign parties (collectively Abitron). Hetronic manufactures remote controls for construction equipment. Abitron, once a licensed distributor for Hetronic, claimed ownership of the rights to much of Hetronic’s intellectual property and began employing Hetronic’s marks on products it sold.

   Hetronic sued Abitron in the Western District of Oklahoma for trademark violations under two related provisions of the Lanham Act, both of which prohibit the unauthorized use in commerce of protected marks when, inter alia, that use is likely to cause confusion. See §§1114(1)(a), 1125(a)(1). Hetronic sought damages for Abitron’s infringing acts worldwide. Abitron argued that Hetronic sought an impermissible extraterritorial application of the Lanham Act. The District Court rejected Abitron’s argument, and a jury later awarded Hetronic approximately $96 million in damages related to Abitron’s global employment of Hetronic’s marks. The District Court also entered a permanent injunction preventing Abitron from using Hetronic’s marks anywhere in the world. On appeal, the Tenth Circuit narrowed the injunction, but otherwise affirmed the judgment, concluding that the Lanham Act extended to “all of [Abitron’s] foreign infringing conduct.”

Held: Applying the presumption against extraterritoriality, §1114(1)(a) and §1125(a)(1) of the Lanham Act are not extraterritorial and extend only to claims where the infringing use in commerce is domestic. Pp. 3–15.

  (a) The presumption against extraterritoriality reflects the longstanding principle “that legislation of Congress, unless a contrary intent appears, is meant to apply only within the territorial jurisdiction of the United States.” Morrison v. National Australia Bank Ltd., 561 U. S. 247, 255. The presumption “serves to avoid the international discord that can result when U. S. law is applied to conduct in foreign countries” and reflects the “commonsense notion that Congress generally legislates with domestic concerns in mind.” RJR Nabisco, Inc. v. European Community, 579 U. S. 325, 335–336.

  Applying the presumption involves a two-step framework, which asks at step one whether the statute is extraterritorial. This step turns on whether “Congress has affirmatively and unmistakably instructed that” the provision at issue should “apply to foreign conduct.” Id., at 335. If Congress has provided such an instruction, then the provision is extraterritorial. If not, then the provision is not extraterritorial and step two applies. That step resolves whether a suit seeks a (permissible) domestic or (impermissible) foreign application of the provision. That determination requires courts to identify the “focus” of congressional concern underlying the provision at issue, id., at 336, and then “as[k] whether the conduct relevant to that focus occurred in United States territory,” WesternGeco LLC v. ION Geophysical Corp., 585 U. S. ___, ___. Thus, to prove that a claim involves a domestic application of a statute, “plaintiffs must establish that ‘the conduct relevant to the statute’s focus occurred in the United States.’ ” Nestlé USA, Inc. v. Doe, 593 U. S. ___, ___ (emphasis added). Step two is designed to apply the presumption to claims that involve both domestic and foreign conduct, separating the activity that matters from the activity that does not. After all, the Court has long recognized that the presumption would be meaningless if any domestic conduct could defeat it. See Morrison, 561 U. S., at 266. Pp. 3–5.

   (b) Neither provision at issue provides an express statement of extraterritorial application or any other clear indication that it is one of the “rare” provisions that nonetheless applies abroad. Both simply prohibit the use “in commerce” of protected trademarks when that use “is likely to cause confusion.” §§1114(1)(a), 1125(a)(1). Hetronic maintains that the Lanham Act’s definition of “commerce”—“all commerce which may lawfully be regulated by Congress,” §1127—rebuts the presumption against extraterritoriality. But this Court’s repeated holding that “ ‘even statutes . . . that expressly refer to “foreign commerce’ ” when defining “commerce” are not extraterritorial, Morrison, 561 U. S., at 262–263, dooms Hetronic’s arguments. Pp. 5–7.

  (c) Because §1114(1)(a) and §1125(a)(1) are not extraterritorial, the Court must consider at step two when claims involve “domestic” appli cations of these provisions. Under the proper test, the ultimate question regarding permissible domestic application turns on the location of the conduct relevant to the focus of the statutory provisions. But much of the parties’ dispute in this case misses this critical point and centers on the “focus” of the relevant provisions without regard to the “conduct relevant to that focus.” WesternGeco, 585 U. S., at ___. Abitron contends that §1114(1)(a) and §1125(a)(1) focus on preventing infringing use of trademarks, while Hetronic argues that they focus both on protecting the goodwill of mark owners and on preventing consumer confusion. The United States as amicus curiae argues that the provisions focus only on likely consumer confusion. The parties all seek support for their positions in Steele v. Bulova Watch Co., 344 U. S. 280, but because Steele implicated both domestic conduct and a likelihood of domestic confusion, Steele does not answer which one determines the domestic applications of the provisions here.

  The ultimate question regarding permissible domestic application turns on the location of the conduct relevant to the focus. See, e.g., RJR Nabisco, 579 U. S., at 337. And the conduct relevant to any focus the parties have proffered is infringing use in commerce, as defined by the Act. This conclusion follows from the text and context of both provisions. Both provisions prohibit the unauthorized “use in commerce” of a protected trademark when that use “is likely to cause confusion.” In other words, Congress proscribed the use of a mark in commerce under certain conditions. This conduct, to be sure, must create a sufficient risk of confusion, but confusion is not a separate requirement; rather, it is simply a necessary characteristic of an offending use. Because Congress has premised liability on a specific action (a particular sort of use in commerce), that specific action would be the conduct relevant to any focus on offer today. WesternGeco, 585 U. S., at ___–___.

  In sum, §1114(1)(a) and §1125(a)(1) are not extraterritorial, and “use in commerce” provides the dividing line between foreign and domestic applications of these provisions. The proceedings below were not in accord with this understanding of extraterritoriality. Pp. 7–10, 14–15.

10 F. 4th 1016, vacated and remanded.

 Alito, J., delivered the opinion of the Court, in which Thomas, Gorsuch, Kavanaugh, and Jackson, JJ., joined. Jackson, J., filed a concurring opinion. Sotomayor, J., filed an opinion concurring in the judgment, in which Roberts, C. J., and Kagan and Barrett, JJ., joined.

Harvard College and the University of North Carolina (UNC) are two of the oldest institutions of higher learning in the United States. Every year, tens of thousands of students apply to each school; many fewer are admitted. Both Harvard and UNC employ a highly selective admissions process to make their decisions. Admission to each school can depend on a student’s grades, recommendation letters, or extracurricular involvement. It can also depend on their race. The question presented is whether the admissions systems used by Harvard College and UNC are lawful under the Equal Protection Clause of the Fourteenth Amendment.

  At Harvard, each application for admission is initially screened by a “first reader,” who assigns a numerical score in each of six categories: academic, extracurricular, athletic, school support, personal, and overall. For the “overall” category—a composite of the five other ratings—a first reader can and does consider the applicant’s race. Harvard’s admissions subcommittees then review all applications from a particular geographic area. These regional subcommittees make recommendations to the full admissions committee, and they take an applicant’s race into account. When the 40-member full admissions committee begins its deliberations, it discusses the relative breakdown of applicants by race. The goal of the process, according to Harvard’s director of admissions, is ensuring there is no “dramatic drop-off” in minority admissions from the prior class. An applicant receiving a majority of the full committee’s votes is tentatively accepted for admission. At the end of this process, the racial composition of the tentative applicant pool is disclosed to the committee. The last stage of Harvard’s admissions process, called the “lop,” winnows the list of tentatively admitted students to arrive at the final class. Applicants that Harvard considers cutting at this stage are placed on the “lop list,” which contains only four pieces of information: legacy status, recruited athlete status, financial aid eligibility, and race. In the Harvard admissions process, “race is a determinative tip for” a significant percentage “of all admitted African American and Hispanic applicants.”

  UNC has a similar admissions process. Every application is reviewed first by an admissions office reader, who assigns a numerical rating to each of several categories. Readers are required to consider the applicant’s race as a factor in their review. Readers then make a written recommendation on each assigned application, and they may provide an applicant a substantial “plus” depending on the applicant’s race. At this stage, most recommendations are provisionally final. A committee of experienced staff members then conducts a “school group review” of every initial decision made by a reader and either approves or rejects the recommendation. In making those decisions, the committee may consider the applicant’s race.

  Petitioner, Students for Fair Admissions (SFFA), is a nonprofit organization whose stated purpose is “to defend human and civil rights secured by law, including the right of individuals to equal protection under the law.” SFFA filed separate lawsuits against Harvard and UNC, arguing that their race-based admissions programs violate, respectively, Title VI of the Civil Rights Act of 1964 and the Equal Protection Clause of the Fourteenth Amendment. After separate bench trials, both admissions programs were found permissible under the Equal Protection Clause and this Court’s precedents. In the Harvard case, the First Circuit affirmed, and this Court granted certiorari. In the UNC case, this Court granted certiorari before judgment.

Held: Harvard’s and UNC’s admissions programs violate the Equal Protection Clause of the Fourteenth Amendment. Pp. 6–40.

 (a) Because SFFA complies with the standing requirements for organizational plaintiffs articulated by this Court in Hunt v. Washington State Apple Advertising Comm’n, 432 U. S. 333, SFFA’s obligations under Article III are satisfied, and this Court has jurisdiction to consider the merits of SFFA’s claims.

 The Court rejects UNC’s argument that SFFA lacks standing because it is not a “genuine” membership organization. An organizational plaintiff can satisfy Article III jurisdiction in two ways, one of which is to assert “standing solely as the representative of its mem bers,” Warth v. Seldin, 422 U. S. 490, 511, an approach known as representational or organizational standing. To invoke it, an organization must satisfy the three-part test in Hunt. Respondents do not suggest that SFFA fails Hunt’s test for organizational standing. They argue instead that SFFA cannot invoke organizational standing at all because SFFA was not a genuine membership organization at the time it filed suit. Respondents maintain that, under Hunt, a group qualifies as a genuine membership organization only if it is controlled and funded by its members. In Hunt, this Court determined that a state agency with no traditional members could still qualify as a genuine membership organization in substance because the agency represented the interests of individuals and otherwise satisfied Hunt’s three-part test for organizational standing. See 432 U. S., at 342. Hunt’s “indicia of membership” analysis, however, has no applicability here. As the courts below found, SFFA is indisputably a voluntary membership organization with identifiable members who support its mission and whom SFFA represents in good faith. SFFA is thus entitled to rely on the organizational standing doctrine as articulated in Hunt. Pp. 6–9.

 (b) Proposed by Congress and ratified by the States in the wake of the Civil War, the Fourteenth Amendment provides that no State shall “deny to any person . . . the equal protection of the laws.” Proponents of the Equal Protection Clause described its “foundation[al] principle” as “not permit[ing] any distinctions of law based on race or color.” Any “law which operates upon one man,” they maintained, should “operate equally upon all.” Accordingly, as this Court’s early decisions interpreting the Equal Protection Clause explained, the Fourteenth Amendment guaranteed “that the law in the States shall be the same for the black as for the white; that all persons, whether colored or white, shall stand equal before the laws of the States.”

 Despite the early recognition of the broad sweep of the Equal Protection Clause, the Court—alongside the country—quickly failed to live up to the Clause’s core commitments. For almost a century after the Civil War, state-mandated segregation was in many parts of the Nation a regrettable norm. This Court played its own role in that ignoble history, allowing in Plessy v. Ferguson the separate but equal regime that would come to deface much of America. 163 U. S. 537.

 After Plessy, “American courts . . . labored with the doctrine [of separate but equal] for over half a century.” Brown v. Board of Education, 347 U. S. 483, 491. Some cases in this period attempted to curtail the perniciousness of the doctrine by emphasizing that it required States to provide black students educational opportunities equal to—even if formally separate from—those enjoyed by white students. See, e.g., Missouri ex rel. Gaines v. Canada, 305 U. S. 337, 349–350. But the inherent folly of that approach—of trying to derive equality from inequality—soon became apparent. As the Court subsequently recognized, even racial distinctions that were argued to have no palpable effect worked to subordinate the afflicted students. See, e.g., McLaurin v. Oklahoma State Regents for Higher Ed., 339 U. S. 637, 640–642. By 1950, the inevitable truth of the Fourteenth Amendment had thus begun to reemerge: Separate cannot be equal.

 The culmination of this approach came finally in Brown v. Board of Education, 347 U. S. 483. There, the Court overturned the separate but equal regime established in Plessy and began on the path of invalidating all de jure racial discrimination by the States and Federal Government. The conclusion reached by the Brown Court was unmistakably clear: the right to a public education “must be made available to all on equal terms.” 347 U. S., at 493. The Court reiterated that rule just one year later, holding that “full compliance” with Brown required schools to admit students “on a racially nondiscriminatory basis.” Brown v. Board of Education, 349 U. S. 294, 300–301.

 In the years that followed, Brown’s “fundamental principle that racial discrimination in public education is unconstitutional,” id., at 298, reached other areas of life—for example, state and local laws requiring segregation in busing, Gayle v. Browder, 352 U. S. 903 (per curiam); racial segregation in the enjoyment of public beaches and bathhouses Mayor and City Council of Baltimore v. Dawson, 350 U. S. 877 (per curiam); and antimiscegenation laws, Loving v. Virginia, 388 U. S. 1. These decisions, and others like them, reflect the “core purpose” of the Equal Protection Clause: “do[ing] away with all governmentally imposed discrimination based on race.” Palmore v. Sidoti, 466 U. S. 429, 432.

 Eliminating racial discrimination means eliminating all of it. Accordingly, the Court has held that the Equal Protection Clause applies “without regard to any differences of race, of color, or of nationality”—it is “universal in [its] application.” Yick Wo v. Hopkins, 118 U. S. 356, 369. For “[t]he guarantee of equal protection cannot mean one thing when applied to one individual and something else when applied to a person of another color.” Regents of Univ. of Cal. v. Bakke, 438 U. S. 265, 289–290.

 Any exceptions to the Equal Protection Clause’s guarantee must survive a daunting two-step examination known as “strict scrutiny,” Adarand Constructors, Inc. v. Peña, 515 U. S. 200, 227, which asks first whether the racial classification is used to “further compelling governmental interests,” Grutter v. Bollinger, 539 U. S. 306, 326, and second whether the government’s use of race is “narrowly tailored,” i.e., “necessary,” to achieve that interest, Fisher v. University of Tex. at Austin, 570 U. S. 297, 311–312. Acceptance of race-based state action is rare for a reason: “[d]istinctions between citizens solely because of their ancestry are by their very nature odious to a free people whose institutions are founded upon the doctrine of equality.” Rice v. Cayetano, 528 U. S. 495, 517. Pp. 9–16.

 (c) This Court first considered whether a university may make race-based admissions decisions in Bakke, 438 U. S. 265. In a deeply splintered decision that produced six different opinions, Justice Powell’s opinion for himself alone would eventually come to “serv[e] as the touchstone for constitutional analysis of race-conscious admissions policies.” Grutter, 539 U. S., at 323. After rejecting three of the University’s four justifications as not sufficiently compelling, Justice Powell turned to its last interest asserted to be compelling—obtaining the educational benefits that flow from a racially diverse student body. Justice Powell found that interest to be “a constitutionally permissible goal for an institution of higher education,” which was entitled as a matter of academic freedom “to make its own judgments as to . . . the selection of its student body.” 438 U. S., at 311–312. But a university’s freedom was not unlimited—“[r]acial and ethnic distinctions of any sort are inherently suspect,” Justice Powell explained, and antipathy toward them was deeply “rooted in our Nation’s constitutional and demographic history.” Id., at 291. Accordingly, a university could not employ a two-track quota system with a specific number of seats reserved for individuals from a preferred ethnic group. Id., at 315. Neither still could a university use race to foreclose an individual from all consideration. Id., at 318. Race could only operate as “a ‘plus’ in a particular applicant’s file,” and even then it had to be weighed in a manner “flexible enough to consider all pertinent elements of diversity in light of the particular qualifications of each applicant.” Id., at 317. Pp. 16–19.

 (d) For years following Bakke, lower courts struggled to determine whether Justice Powell’s decision was “binding precedent.” Grutter, 539 U. S., at 325. Then, in Grutter v. Bollinger, the Court for the first time “endorse[d] Justice Powell’s view that student body diversity is a compelling state interest that can justify the use of race in university admissions.” Ibid. The Grutter majority’s analysis tracked Justice Powell’s in many respects, including its insistence on limits on how universities may consider race in their admissions programs. Those limits, Grutter explained, were intended to guard against two dangers that all race-based government action portends. The first is the risk that the use of race will devolve into “illegitimate . . . stereotyp[ing].” Richmond v. J. A. Croson Co., 488 U. S. 469, 493 (plurality opinion). Admissions programs could thus not operate on the “belief that minority students always (or even consistently) express some characteristic minority viewpoint on any issue.” Grutter, 539 U. S., at 333 (internal quotation marks omitted). The second risk is that race would be used not as a plus, but as a negative—to discriminate against those racial groups that were not the beneficiaries of the race-based preference. A university’s use of race, accordingly, could not occur in a manner that “unduly harm[ed] nonminority applicants.” Id., at 341.

 To manage these concerns, Grutter imposed one final limit on race-based admissions programs: At some point, the Court held, they must end. Id., at 342. Recognizing that “[e]nshrining a permanent justification for racial preferences would offend” the Constitution’s unambiguous guarantee of equal protection, the Court expressed its expectation that, in 25 years, “the use of racial preferences will no longer be necessary to further the interest approved today.” Id., at 343. Pp. 19–21.

 (e) Twenty years have passed since Grutter, with no end to race-based college admissions in sight. But the Court has permitted race-based college admissions only within the confines of narrow restrictions: such admissions programs must comply with strict scrutiny, may never use race as a stereotype or negative, and must—at some point—end. Respondents’ admissions systems fail each of these criteria and must therefore be invalidated under the Equal Protection Clause of the Fourteenth Amendment. Pp. 21–34.

  (1) Respondents fail to operate their race-based admissions programs in a manner that is “sufficiently measurable to permit judicial [review]” under the rubric of strict scrutiny. Fisher v. University of Tex. at Austin, 579 U. S. 365, 381. First, the interests that respondents view as compelling cannot be subjected to meaningful judicial review. Those interests include training future leaders, acquiring new knowledge based on diverse outlooks, promoting a robust marketplace of ideas, and preparing engaged and productive citizens. While these are commendable goals, they are not sufficiently coherent for purposes of strict scrutiny. It is unclear how courts are supposed to measure any of these goals, or if they could, to know when they have been reached so that racial preferences can end. The elusiveness of respondents’ asserted goals is further illustrated by comparing them to recognized compelling interests. For example, courts can discern whether the temporary racial segregation of inmates will prevent harm to those in the prison, see Johnson v. California, 543 U. S. 499, 512–513, but the question whether a particular mix of minority students produces “engaged and productive citizens” or effectively “train[s] future leaders” is standardless.

 Second, respondents’ admissions programs fail to articulate a meaningful connection between the means they employ and the goals they pursue. To achieve the educational benefits of diversity, respondents measure the racial composition of their classes using racial categories that are plainly overbroad (expressing, for example, no concern whether South Asian or East Asian students are adequately represented as “Asian”); arbitrary or undefined (the use of the category “Hispanic”); or underinclusive (no category at all for Middle Eastern students). The unclear connection between the goals that respondents seek and the means they employ preclude courts from meaningfully scrutinizing respondents’ admissions programs.

 The universities’ main response to these criticisms is “trust us.” They assert that universities are owed deference when using race to benefit some applicants but not others. While this Court has recognized a “tradition of giving a degree of deference to a university’s academic decisions,” it has made clear that deference must exist “within constitutionally prescribed limits.” Grutter, 539 U. S., at 328. Respondents have failed to present an exceedingly persuasive justification for separating students on the basis of race that is measurable and concrete enough to permit judicial review, as the Equal Protection Clause requires. Pp. 22–26.

  (2) Respondents’ race-based admissions systems also fail to comply with the Equal Protection Clause’s twin commands that race may never be used as a “negative” and that it may not operate as a stereotype. The First Circuit found that Harvard’s consideration of race has resulted in fewer admissions of Asian-American students. Respondents’ assertion that race is never a negative factor in their admissions programs cannot withstand scrutiny. College admissions are zero-sum, and a benefit provided to some applicants but not to others necessarily advantages the former at the expense of the latter.

 Respondents admissions programs are infirm for a second reason as well: They require stereotyping—the very thing Grutter foreswore. When a university admits students “on the basis of race, it engages in the offensive and demeaning assumption that [students] of a particular race, because of their race, think alike.” Miller v. Johnson, 515 U. S. 900, 911–912. Such stereotyping is contrary to the “core purpose” of the Equal Protection Clause. Palmore, 466 U. S., at 432. Pp. 26–29.

  (3) Respondents’ admissions programs also lack a “logical end point” as Grutter required. 539 U. S., at 342. Respondents suggest that the end of race-based admissions programs will occur once meaningful representation and diversity are achieved on college campuses. Such measures of success amount to little more than comparing the racial breakdown of the incoming class and comparing it to some other metric, such as the racial makeup of the previous incoming class or the population in general, to see whether some proportional goal has been reached. The problem with this approach is well established: “[O]utright racial balancing” is “patently unconstitutional.” Fisher, 570 U. S., at 311. Respondents’ second proffered end point—when students receive the educational benefits of diversity—fares no better. As explained, it is unclear how a court is supposed to determine if or when such goals would be adequately met. Third, respondents suggest the 25-year expectation in Grutter means that race-based preferences must be allowed to continue until at least 2028. The Court’s statement in Grutter, however, reflected only that Court’s expectation that race-based preferences would, by 2028, be unnecessary in the context of racial diversity on college campuses. Finally, respondents argue that the frequent reviews they conduct to determine whether racial preferences are still necessary obviates the need for an end point. But Grutter never suggested that periodic review can make unconstitutional conduct constitutional. Pp. 29–34.

 (f) Because Harvard’s and UNC’s admissions programs lack sufficiently focused and measurable objectives warranting the use of race, unavoidably employ race in a negative manner, involve racial stereotyping, and lack meaningful end points, those admissions programs cannot be reconciled with the guarantees of the Equal Protection Clause. At the same time, nothing prohibits universities from considering an applicant’s discussion of how race affected the applicant’s life, so long as that discussion is concretely tied to a quality of character or unique ability that the particular applicant can contribute to the university. Many universities have for too long wrongly concluded that the touchstone of an individual’s identity is not challenges bested, skills built, or lessons learned, but the color of their skin. This Nation’s constitutional history does not tolerate that choice. Pp. 39–40.

No. 20–1199, 980 F. 3d 157; No. 21–707, 567 F. Supp. 3d 580, reversed.

 Roberts, C. J., delivered the opinion of the Court, in which Thomas, Alito, Gorsuch, Kavanaugh, and Barrett, JJ., joined. Thomas, J., filed a concurring opinion. Gorsuch, J., filed a concurring opinion, in which Thomas, J., joined. Kavanaugh, J., filed a concurring opinion. Sotomayor, J., filed a dissenting opinion, in which Kagan, J., joined, and in which Jackson, J., joined as it applies to No. 21–707. Jackson, J., filed a dissenting opinion in No. 21–707, in which Sotomayor and Kagan, JJ., joined. Jackson, J., took no part in the consideration or decision of the case in No. 20–1199.

Notes
¹ Together with No. 21–707, Students for Fair Admissions, Inc. v. University of North Carolina et al., on certiorari before judgment to the United States Court of Appeals for the Fourth Circuit.

Petitioner Gerald Groff is an Evangelical Christian who believes for religious reasons that Sunday should be devoted to worship and rest. In 2012, Groff took a mail delivery job with the United States Postal Service. Groff’s position generally did not involve Sunday work, but that changed after USPS agreed to begin facilitating Sunday deliveries for Amazon. To avoid the requirement to work Sundays on a rotating basis, Groff transferred to a rural USPS station that did not make Sunday deliveries. After Amazon deliveries began at that station as well, Groff remained unwilling to work Sundays, and USPS redistributed Groff’s Sunday deliveries to other USPS staff. Groff received “progressive discipline” for failing to work on Sundays, and he eventually resigned.

   Groff sued under Title VII of the Civil Rights Act of 1964, asserting that USPS could have accommodated his Sunday Sabbath practice “without undue hardship on the conduct of [USPS’s] business.” 42 U. S. C. §2000e(j). The District Court granted summary judgment to USPS. The Third Circuit affirmed based on this Court’s decision in Trans World Airlines, Inc. v. Hardison, 432 U. S. 63, which it construed to mean “that requiring an employer ‘to bear more than a de minimis cost’ to provide a religious accommodation is an undue hardship.” 35 F. 4th 162, 174, n. 18 (quoting 432 U. S., at 84). The Third Circuit found the de minimis cost standard met here, concluding that exempting Groff from Sunday work had “imposed on his coworkers, disrupted the workplace and workflow, and diminished employee morale.” 35 F. 4th, at 175.

Held: Title VII requires an employer that denies a religious accommodation to show that the burden of granting an accommodation would re sult in substantial increased costs in relation to the conduct of its particular business.

Pp. 4–21.   (a) This case presents the Court’s first opportunity in nearly 50 years to explain the contours of Hardison. The background of that decision helps to explain the Court’s disposition of this case. Pp. 4–15.

   (1) Title VII of the Civil Rights Act of 1964 made it unlawful for covered employers “to fail or refuse to hire or to discharge any individual, or otherwise to discriminate against any individual with respect to his compensation, terms, conditions, or privileges [of] employment, because of such individual’s . . . religion.” §2000e–2(a)(1). As originally enacted, Title VII did not spell out what it meant by discrimination “because of . . . religion.” Subsequent regulations issued by the EEOC obligated employers “to make reasonable accommodations to the religious needs of employees” whenever doing so would not create “undue hardship on the conduct of the employer’s business.” 29 CFR §1605.1 (1968). In 1970, however, the Sixth Circuit held that Title VII did not require an employer “to accede to or accommodate” a Sabbath religious practice because to do so “would raise grave” Establishment Clause questions. Dewey v. Reynolds Metals Co., 429 F. 2d 324, 334. This Court affirmed Dewey by an evenly divided vote. See 402 U. S. 689. Congress responded by amending Title VII in 1972 to track the EEOC’s regulatory language and to clarify that employers must “reasonably accommodate. . . an employee’s or prospective employee’s religious observance or practice” unless the employer is “unable” to do so “without undue hardship on the conduct of the employer’s business.” §2000e(j). Pp. 4–6.

   (2) Hardison concerned an employment dispute that arose prior to the 1972 amendments to Title VII. In 1967, Trans World Airlines hired Larry Hardison to work in a department that operated “24 hours per day, 365 days per year” and played an “essential role” for TWA by providing parts needed to repair and maintain aircraft. Hardison, 432 U. S., at 66. Hardison later underwent a religious conversion and began missing work to observe the Sabbath. Initial conflicts with Hardison’s work schedule were resolved, but conflicts resurfaced when he transferred to another position in which he lacked the seniority to avoid work during his Sabbath. Attempts at accommodation failed, and TWA discharged Hardison for insubordination.

  Hardison sued TWA and his union, and the Eighth Circuit sided with Hardison. The Eighth Circuit found that reasonable accommodations were available to TWA, and rejected the defendants’ Establishment Clause arguments. Hardison v. Trans World Airlines, Inc., 527 F. 2d 33, 42–44. This Court granted certiorari. TWA’s petition for certiorari asked this Court to decide whether the 1972 amendment of Title VII violated the Establishment Clause as applied by the Eighth Circuit, particularly insofar as that decision had approved an accommodation that allegedly overrode seniority rights granted by the relevant collective bargaining agreement. At the time, some thought that the Court’s now-abrogated decision in Lemon v. Kurtzman, 403 U. S. 602—which adopted a test under which any law whose “principal or primary effect” “was to advance religion” was unconstitutional, id., at 612–613—posed a serious problem for the 1972 amendment of Title VII. Ultimately, however, constitutional concerns played no on-stage role in the Court’s decision in Hardison. Instead, the Court’s opinion stated that “the principal issue on which TWA and the union came to this Court” was whether Title VII “require[s] an employer and a union who have agreed on a seniority system to deprive senior employees of their seniority rights in order to accommodate a junior employee’s religious practices.” Hardison, 432 U. S., at 83, and n. 14. The Court held that Title VII imposed no such requirement. Id., at 83, and n. 14. This conclusion, the Court found, was “supported by the fact that seniority systems are afforded special treatment under Title VII itself.” Id., at 81. Applying this interpretation of Title VII and disagreeing with the Eighth Circuit’s evaluation of the factual record, the Court identified no way in which TWA, without violating seniority rights, could have feasibly accommodated Hardison’s request for an exemption from work on his Sabbath.

  The parties had not focused on determining when increased costs amount to “undue hardship” under Title VII separately from the seniority issue. But the Court’s opinion in Hardison contained this oft-quoted sentence: “To require TWA to bear more than a de minimis cost in order to give Hardison Saturdays off is an undue hardship.” Although many lower courts later viewed this line as the authoritative interpretation of the statutory term “undue hardship,” the context renders that reading doubtful. In responding to Justice Marshall’s dissent, the Court described the governing standard quite differently, stating three times that an accommodation is not required when it entails “substantial” “costs” or “expenditures.” Id., at 83, n. 14. Pp. 6–12.

   (3) Even though Hardison’s reference to “de minimis” was undercut by conflicting language and was fleeting in comparison to its discussion of the “principal issue” of seniority rights, lower courts have latched on to “de minimis” as the governing standard. To be sure, many courts have understood that the protection for religious adherents is greater than “more than . . . de minimis” might suggest when read in isolation. But diverse religious groups tell the Court that the “de minimis” standard has been used to deny even minor accommodations. The EEOC has also accepted Hardison as prescribing a “more than a de minimis cost” test, 29 CFR §1605.2(e)(1), though it has tried to soften its impact, cautioning against extending the phrase to cover such things as the “administrative costs” involved in reworking schedules, the “infrequent” or temporary “payment of premium wages for a substitute,” and “voluntary substitutes and swaps” when they are not contrary to a “bona fide seniority system.” §§1605.2(e)(1), (2). Yet some courts have rejected even the EEOC’s gloss on “de minimis,” rejecting accommodations the EEOC’s guidelines consider to be ordinarily required. The Court agrees with the Solicitor General that Hardison does not compel courts to read the “more than de minimis” standard “literally” or in a manner that undermines Hardison’s references to “substantial” cost. Tr. of Oral Arg. 107. Pp. 12–15.

  (b) The Court holds that showing “more than a de minimis cost,” as that phrase is used in common parlance, does not suffice to establish “undue hardship” under Title VII. Hardison cannot be reduced to that one phrase. In describing an employer’s “undue hardship” defense, Hardison referred repeatedly to “substantial” burdens, and that formulation better explains the decision. The Court understands Hardison to mean that “undue hardship” is shown when a burden is substantial in the overall context of an employer’s business. This fact-specific inquiry comports with both Hardison and the meaning of “undue hardship” in ordinary speech. Pp. 15–21.

   (1) To determine what an employer must prove to defend a denial of a religious accommodation under Title VII, the Court begins with Title VII's text. The statutory term, “hardship,” refers to, at a minimum, “something hard to bear” and suggests something more severe than a mere burden. If Title VII said only that an employer need not be made to suffer a “hardship,” an employer could not escape liability simply by showing that an accommodation would impose some sort of additional costs. Adding the modifier “undue” means that the requisite burden or adversity must rise to an “excessive” or “unjustifiable” level. Understood in this way, “undue hardship” means something very different from a burden that is merely more than de minimis, i.e., “very small or trifling.” The ordinary meaning of “undue hardship” thus points toward a standard closer to Hardison’s references to “substantial additional costs” or “substantial expenditures.” 432 U. S., at 83, n. 14. Further, the Court’s reading of the statutory term comports with pre-1972 EEOC decisions, so nothing in that history plausibly suggests that “undue hardship” in Title VII should be read to mean anything less than its meaning in ordinary use. Cf. George v. McDonough, 596 U. S. ___, ___. And no support exists in other factors discussed by the parties for reducing Hardison to its “more than a de minimis cost” line. Pp. 16–18.

   (2) The parties agree that the “de minimis” test is not right, but they differ in the alternative language they propose. The Court thinks it is enough to say that what an employer must show is that the burden of granting an accommodation would result in substantial increased costs in relation to the conduct of its particular business. Hardison, 432 U. S. at 83, n. 14. Courts must apply the test to take into account all relevant factors in the case at hand, including the particular accommodations at issue and their practical impact in light of the nature, size, and operating cost of an employer. Pp. 18.

   (3) The Court declines to adopt the elaborations of the applicable standard that the parties suggest, either to incorporate Americans with Disabilities Act case law or opine that the EEOC’s construction of Hardison has been basically correct. A good deal of the EEOC’s guidance in this area is sensible and will, in all likelihood, be unaffected by the Court’s clarifying decision. But it would not be prudent to ratify in toto a body of EEOC interpretation that has not had the benefit of the clarification the Court adopts today. What is most important is that “undue hardship” in Title VII means what it says, and courts should resolve whether a hardship would be substantial in the context of an employer’s business in the commonsense manner that it would use in applying any such test. Pp. 18–19.

   (4) The Court also clarifies several recurring issues. First, as the parties agree, Title VII requires an assessment of a possible accommodation’s effect on “the conduct of the employer’s business.” §2000e(j). Impacts on coworkers are relevant only to the extent those impacts go on to affect the conduct of the business. A court must analyze whether that further logical step is shown. Further, a hardship that is attributable to employee animosity to a particular religion, to religion in general, or to the very notion of accommodating religious practice, cannot be considered “undue.” Bias or hostility to a religious practice or accommodation cannot supply a defense.

  Second, Title VII requires that an employer “reasonably accommodate” an employee’s practice of religion, not merely that it assess the reasonableness of a particular possible accommodation or accommodations. Faced with an accommodation request like Groff ’s, an employer must do more that conclude that forcing other employees to work overtime would constitute an undue hardship. Consideration of other options would also be necessary. Pp. 19–20.

  (c) Having clarified the Title VII undue-hardship standard, the Court leaves the context-specific application of that clarified standard in this case to the lower courts in the first instance. Pp. 21.

35 F. 4th 162, vacated and remanded.

 Alito, J., delivered the opinion for a unanimous Court. Sotomayor, J., filed a concurring opinion, in which Jackson, J., joined.