FISHER

Affirmed.

Roger A. Colaizzi, Esq. (argued), Christopher S. Crook, Esq., Venable, Washington, DC, Mark A. Grace, Esq., Cohen & Grace, Pittsburgh, PA, Randall K. Miller, Esq., Venable, Vienna, VA, for Appellant. Gretchen L. Jankowski, Esq. (argued), Mackenzie A. Baird, Esq., Jordan M. Webster, Esq., Buchanan Ingersoll & Rooney, Pittsburgh, PA, for Appellee.
Before: FISHER, JORDAN and HARDIMAN, Circuit Judges.

OPINION OF THE COURT

FISHER, Circuit Judge.

In this false advertising case, Euro–Pro Operating, LLC (“Euro–Pro”) appeals the District Court's order granting a motion for a preliminary injunction brought by Groupe SEB USA, Inc. (“SEB”). The District Court found that two advertising claims on Euro–Pro's steam irons likely violated section 43(a) of the Lanham Act, 15 U.S.C. § 1125(a), and enjoined Euro–Pro from using those claims. Euro–Pro raises several issues on appeal, but we principally consider how courts should interpret an advertising claim when the packaging or label unambiguously defines a claim term. The District Court decided that the packaging's definition of a claim term applies to the claim's explicit message. Based on this decision, the District Court disregarded consumer survey evidence offering alternative meanings for the claim term. We agree with the District Court and find its approach firmly based in false advertising law and logic. And because we conclude that the District Court did not otherwise abuse its discretion in granting the preliminary injunction, we will affirm.

I.

A.

SEB distributes and sells various household consumer products under several brand names throughout the country. This case involves SEB's electric steam irons sold under the Rowenta brand name, namely the Rowenta Focus, Model No. DW5080 (“Rowenta DW5080”), and the Rowenta Steamium, Model No. DW9080 (“Rowenta DW9080”). Euro–Pro manufactures, markets, and distributes kitchen and household appliances. It sells these products under the Shark brand name. The dispute here arises from advertising claims on the packaging of two Shark steam irons, the Shark Professional, Model No. GI405–55 (“Shark 405”), and the Shark Ultimate Professional, Model No. GI505–55 (“Shark 505”).

The Shark 405 packaging includes two advertising claims. First, text on the bottom right of the front packaging asserts that the Shark 405 offers “MORE POWERFUL STEAM vs. Rowenta®†† at half the price.” J.A. at A3, A805. The “††” characters refer to a fine-print footnote on the bottom of the packaging, which states that the claim is “††[b]ased on independent comparative steam burst testing to Rowenta DW5080 (grams/shot).” Id. Text on the top right of the front packaging also asserts that the Shark 405 delivers “# 1 MOST POWERFUL STEAM*.” Id. Again, there is a fine-print reference to this claim on the bottom of the packaging that states the Shark 405 “ *[o]ffers more grams per minute (maximum steam setting while bursting before water spots appear) when compared to leading competition in the same price range, at time of printing.” Id. The Shark 505 packaging makes substantially the same claims.

Additionally, both the Shark 405 and the Shark 505 include hang tags on the steam irons for store displays. The hang tags claim that the Shark steam irons deliver “MORE POWERFUL STEAM vs. Rowenta ... at half the price.” J.A. at A4. The hang tags also include a reference stating that the claim is “[b]ased on independent comparative steam burst testing” to the respective Rowenta steam irons in “(grams/shot).” Id.

SEB first learned of the comparative advertising claims on the Shark steam irons in October 2013. Soon thereafter, SEB directed its internal laboratory to conduct testing to determine whether the claims were true. The lab ran tests comparing the Shark 505 and the Rowenta DW9080. The tests measured (1) the variable steam rate in grams per minute according to International Electrical Corporation (“IEC”) 60311 protocol and (2) the mass of a shot of steam in grams per shot according to IEC 60311 protocol. The test results showed that the Rowenta DW9080 performed the same as the Shark 505 in terms of variable steam rate in grams per minute, with both measuring 37 grams per minute. In the test measuring grams per shot of steam, the Rowenta DW9080 outperformed the Shark 505, with measurements of 1.34 grams per shot and 1.00 grams per shot, respectively.

Because SEB's internal test results were inconsistent with the Shark advertising claims, SEB commissioned SLG Prüfund Zertifizierungs GmbH (“SLG”), an independent laboratory based in Germany, to conduct independent tests based on the Shark claims. SLG tested three steam irons of each model in accordance with IEC 60311 protocol, and it delivered its findings to SEB in a comprehensive thirty-eight page report (“SLG Test Report”). The SLG Test Report showed that the Rowenta DW5080 and the Rowenta DW9080 outperformed the Shark 405 and the Shark 505, respectively, in terms of grams per minute. For the test measuring steam power in grams per shot, the SLG Test Report showed that two of the three Shark 405 steam irons performed worse than all three Rowenta DW5080 steam irons, but one Shark 405 steam iron outperformed all three Rowenta DW5080 steam irons. The Rowenta DW5080's average performance was higher than the Shark 405's average performance. The SLG Test Report also showed that two of the three Rowenta DW9080 steam irons performed better in grams per shot than all three Shark 505 steam irons, and one Rowenta DW9080 performed worse than all three Shark 505 steam irons. The Rowenta DW9080's average performance was higher than the Shark 505's average performance.

B.

On January 29, 2014, SEB filed a complaint in the United States District Court for the Western District of Pennsylvania, asserting claims for false advertising under the Lanham Act, 15 U.S.C. § 1125(a), and for unfair competition under Pennsylvania common law. The following day, SEB moved for a preliminary injunction to enjoin Euro–Pro from making the claims on the Shark 405 and the Shark 505.

The District Court held an evidentiary hearing on March 19, 2014, to address SEB's motion for a preliminary injunction. At the hearing, SEB introduced the aforementioned internal test results and the independent SLG Test Report to show that the claims on the Shark steam irons are false. Euro–Pro introduced testimony and a study from its scientific expert, Dr. Abid Kemal (collectively referred to as the “Kemal Report”). According to the Kemal Report, steam power is the kinetic energy of a steam burst divided by the duration of the burst. Using this measurement for steam power, the Kemal Report showed that the Shark 405 and the Shark 505 deliver more powerful steam than the Rowenta DW5080 and the Rowenta DW9080, respectively. The Kemal Report also showed that “the mass of a shot of steam expelled from [the Shark steam irons] is comparable to the mass of a shot of steam (grams/shot) expelled from [the respective Rowenta steam irons].” J.A. at A909. Additionally, Euro–Pro introduced a consumer survey report prepared by Dr. Gary Ford (“the Ford Survey”) showing that consumers do not have a uniform understanding of the meaning of the phrase “more powerful steam.”

The District Court also heard testimony from SEB's marketing director, Scott Pollard, about the harm to the Rowenta brand caused by the Shark claims. Pollard testified that SEB had invested substantial resources to promote Rowenta as the best brand of steam irons in the eyes of retailers and consumers. According to Pollard, the direct reference to Rowenta on the lower-priced Shark steam irons likely would erode the Rowenta brand's reputation in the eyes of retailers, current consumers, and future consumers.

The District Court granted SEB's preliminary injunction motion. The District Court first concluded that SEB established a likelihood of success on the merits because it demonstrated that the Shark claims are literally false. The District Court next found that SEB had successfully demonstrated a likelihood of irreparable harm in the absence of preliminary relief, relying in large part on Pollard's testimony about the impact on the reputation of the Rowenta brand and on SEB's goodwill. Finally, the District Court concluded that the balance of harms and the public interest favored granting the preliminary injunction. Notably, the preliminary injunction required Euro–Pro to place stickers over the claims on the Shark packaging and remove the hang tags from the steam irons.

Euro–Pro filed this timely appeal on May 15, 2014. Euro–Pro filed motions to stay the preliminary injunction in the District Court and in this Court, but both motions were denied.

II.

The District Court had jurisdiction under 28 U.S.C. §§ 1331 and 1367. We have jurisdiction over the appeal from the order granting the preliminary injunction pursuant to 28 U.S.C. § 1292(a)(1). “We review an order granting a preliminary injunction for abuse of discretion, the factual findings for clear error, and the determinations of questions of law de novo.” Zambelli Fireworks Mfg. Co. v. Wood, 592 F.3d 412, 418 (3d Cir.2010) (internal quotation marks and citation omitted). We review the details of equitable relief for abuse of discretion. Anderson v. Davila, 125 F.3d 148, 159 (3d Cir.1997).

III.

“A preliminary injunction is an extraordinary remedy never awarded as of right.” Winter v. Natural Res. Def. Council, Inc., 555 U.S. 7, 24, 129 S.Ct. 365, 172 L.Ed.2d 249 (2008). Awarding preliminary relief, therefore, is only appropriate “upon a clear showing that the plaintiff is entitled to such relief.” Id. at 22, 129 S.Ct. 365. “A plaintiff seeking a preliminary injunction must establish that he is likely to succeed on the merits, that he is likely to suffer irreparable harm in the absence of preliminary relief, that the balance of equities tips in his favor, and that an injunction is in the public interest.” Id. at 20, 129 S.Ct. 365.

On appeal, Euro–Pro challenges the District Court's conclusions on the first and second factors in the preliminary injunction test: first, that SEB established a likelihood of success on the merits; and second, that SEB showed a likelihood of irreparable harm without preliminary relief. Euro–Pro also contends that the District Court's injunction violates the First Amendment and is overbroad.

A.

SEB brought its false advertising claims pursuant to the Lanham Act and Pennsylvania common law. Section 43(a) of the Lanham Act provides in pertinent part:

(1) Any person who, on or in connection with any goods or services, or any container for goods, uses in commerce any word, term, name, symbol, or device, or any combination thereof, or any false designation of origin, false or misleading description of fact, or false or misleading representation of fact, which

...

(B) in commercial advertising or promotion, misrepresents the nature, characteristics, qualities, or geographic origin of his or her or another person's goods, services, or commercial activities,

shall be liable in a civil action by any person who believes that he or she is or is likely to be damaged by such act.
15 U.S.C. § 1125(a)(1)(B).

To establish a claim for false advertising, a Lanham Act plaintiff must prove five elements:

1) that the defendant has made false or misleading statements as to his own product [or another's]; 2) that there is actual deception or at least a tendency to deceive a substantial portion of the intended audience; 3) that the deception is material in that it is likely to influence purchasing decisions; 4) that the advertised goods traveled in interstate commerce; and 5) that there is a likelihood of injury to the plaintiff in terms of declining sales, loss of good will, etc.
Pernod Ricard USA, LLC v. Bacardi U.S.A., Inc., 653 F.3d 241, 248 (3d Cir.2011) (alteration in original) (quoting Warner–Lambert Co. v. Breathasure, Inc., 204 F.3d 87, 91–92 (3d Cir.2000)).

A plaintiff can prevail in a false advertising action if it proves that the advertisement “is either (1) literally false or (2) literally true or ambiguous, but has the tendency to deceive consumers.” Novartis Consumer Health, Inc. v. Johnson & Johnson–Merck Consumer Pharm. Co., 290 F.3d 578, 586 (3d Cir.2002). Proof of literal falsity relieves the plaintiff of its burden to prove actual consumer deception. Id. Here, the only dispute is whether the Shark claims are literally false.

“A determination of literal falsity rests on an analysis of the message in context.” Johnson & Johnson–Merck Consumer Pharm. Co. v. Rhone–Poulenc Rorer Pharm., Inc., 19 F.3d 125, 129 (3d Cir.1994). In deciding whether an advertising claim is literally false, a court must decide first whether the claim conveys an unambiguous message and second whether that unambiguous message is false. Novartis, 290 F.3d at 586. “A ‘literally false’ message may be either explicit or ‘conveyed by necessary implication when, considering the advertisement in its entirety, the audience would recognize the claim as readily as if it had been explicitly stated.’ ” Id. at 586–87 (quoting Clorox Co. P.R. v. Proctor & Gamble Commercial Co., 228 F.3d 24, 35 (1st Cir.2000)). Unless the claim is unambiguous, however, it cannot be literally false. Id. at 587. “ ‘The greater the degree to which a message relies upon the viewer or consumer to integrate its components and draw the apparent conclusion ... the less likely it is that a finding of literal falsity will be supported.’ ” Id. (quoting United Indus. Corp. v. Clorox Co., 140 F.3d 1175, 1181 (8th Cir.1998)). We review a district court's findings that an advertising claim is unambiguous and literally false for clear error. See id. at 589.

The District Court analyzed the two advertising claims at issue separately. It first determined that Euro–Pro's claim that the Shark steam irons offer “MORE POWERFUL STEAM vs. Rowenta” is unambiguous. The District Court found that the footnote reference to this claim governs the claim's meaning, as the packaging explicitly claims that the Shark steam irons offer more powerful steam measured in grams per shot than the respective Rowenta steam irons. The District Court also determined that the “# 1 MOST POWERFUL STEAM” claim is unambiguous but for different reasons. Recognizing that the reference to this claim explicitly restricts the claim to comparisons to steam irons in the same price range and that Rowenta steam irons are in a higher price range, the District Court still found an unambiguous message of superiority over Rowenta steam irons conveyed by necessary implication due to the claim's close proximity to the “MORE POWERFUL STEAM vs. Rowenta” claim.

With respect to the question of falsity, the District Court found that both claims are false because all the scientific evidence that measured steam power in grams per shot and grams per minute—the measurements for steam power provided on the Shark packaging—disproved Euro–Pro's claims of superiority over Rowenta. The District Court rejected Euro–Pro's scientific evidence, the Kemal Report, as irrelevant because it did not measure steam power in grams per shot or grams per minute. The District Court also observed that Euro–Pro failed to come forward with any other evidence that actually supported its claims.

1.

We agree with the District Court that the “MORE POWERFUL STEAM vs. Rowenta” claim is unambiguous. When a product's packaging includes an advertising claim and unambiguously defines a claim term, the packaging's definition of the claim term applies to the claim's explicit message. As explained below, we think this rule is consistent with false advertising law and common sense.

In certain cases, determining the message conveyed by a claim is a simple exercise because the claim is explicit and unambiguous. See Novartis, 290 F.3d at 586. And so it is here. To make something explicit is to state it clearly and precisely. Therefore, when Euro–Pro took the affirmative step to include a reference on the Shark packaging that clearly defined the key term in its claim—that steam power is measured in grams per shot—it made an explicit claim. The claim is also unambiguous because grams per shot is a unit of measurement provided by the IEC, the leading independent publisher of standards for electrotechnology, including steam irons. Thus, there is no “ ‘apparent conclusion’ ” to be drawn about this claim's meaning, id. at 587 (quoting United Indus., 140 F.3d at 1181), nor is its meaning “balanced between several plausible meanings,” Clorox Co. P.R., 228 F.3d at 35. There is only one available conclusion and only one plausible meaning—the claim means exactly what the reference on the packaging says it does.

Moreover, as we previously discussed, courts deciding whether a claim is literally false must view the claim in the context of the entire advertisement. See Rhone–Poulenc, 19 F.3d at 129. Here, the reference that defines the meaning of steam power is on the Shark packaging, and the claim expressly links to the reference using a symbol—“††” on the Shark 405 and “†” on the Shark 505. Thus, ignoring the reference in our analysis would be not only to read the claim out of context, but also to ignore part of the claim itself denoted by the symbol.

Our holding is also consistent with other areas of the law where courts interpreting a term's meaning apply a specific definition if one is provided by the author. See, e.g., Meese v. Keene, 481 U.S. 465, 484, 107 S.Ct. 1862, 95 L.Ed.2d 415 (1987) (“It is axiomatic that the statutory definition of the term excludes unstated meanings of that term.”); Phillips v. AWH Corp., 415 F.3d 1303, 1316 (Fed.Cir.2005) (en banc) (“[O]ur cases recognize that the specification may reveal a special definition given to a claim term by the patentee that differs from the meaning it would otherwise possess. In such cases, the inventor's lexicography governs.”); J.C. Penney Life Ins. Co. v. Pilosi, 393 F.3d 356, 363 (3d Cir.2004) (applying Pennsylvania law to interpret an insurance contract, and explaining that words expressly defined in a policy will be given that definition by courts interpreting the policy); 12 Richard A. Lord, Williston on Contracts, § 34:11, at 123 (4th ed. 2012) (“Another method for excluding usage is to have the contract define terms in a manner that is different from the industry or trade definitions for those terms. Then the contract definitions govern and usage is inapplicable....”). We see no reason to depart from this principle here.

We therefore agree entirely with the District Court that the reference's definition of steam power governs the term's meaning in the “MORE POWERFUL STEAM vs. Rowenta” claim. Accordingly, the claim's explicit and unambiguous message is that the Shark steam irons offer more powerful steam measured in grams per shot than the respective Rowenta steam irons.

The fact that the references are in fine-print footnotes and presumably less likely to be read by consumers does not alter our analysis, as Euro–Pro urges it should. We understand that other courts have held that footnote disclaimers purporting to make a false or misleading claim literally true cannot cure the claim's false or misleading message. See, e.g., Am. Home Prods. Corp. v. Johnson & Johnson, 654 F.Supp. 568, 590 (S.D.N.Y.1987). We have not addressed this issue, see Pernod, 653 F.3d at 252 n. 13 (declining to address the situation when an allegedly misleading claim is corrected by a true statement contained in fine print), and we do not decide it today. Our rather unremarkable holding here is analytically distinct. It is that what a product's packaging says a claim term means is in fact part of the claim's explicit message. If that explicit message is both unambiguous and false, the claim is literally false.

Nor does the presence of consumer survey evidence showing alternative meanings for a defined term affect our holding. Euro–Pro would have us ignore the packaging's definition of steam power and instead credit consumer survey evidence demonstrating that the meaning of steam power is ambiguous. According to Euro–Pro, the District Court's decision to ignore the Ford Survey is inconsistent with our decision in Pernod Ricard USA, LLC v. Bacardi U.S.A., Inc., 653 F.3d 241 (3d Cir.2011). The crux of Euro–Pro's argument is that consumer surveys must be considered by courts in determining whether a claim's message is ambiguous. As explained below, Euro–Pro's argument does not hold up.

In Pernod, we addressed whether courts must always consider survey evidence showing that consumers are misled by an advertising claim. There, the appellant asserted that the name of a brand of rum, “Havana Club,” misled consumers about the brand's geographic origin. Id. at 247. Beneath the “Havana Club” name, the label prominently stated that it was “Puerto Rican Rum,” an accurate statement of where the rum was distilled. Id. at 245–46. The District Court found that the label made no false or misleading statement, so it disregarded consumer survey evidence showing that eighteen percent of consumers were confused about the brand's geographic origin. See id. at 247–48.

We held that the district court properly disregarded the consumer survey evidence. Our conclusion rested on the principle “that there is and must be a point at which language is used plainly enough that the question ceases to be ‘what does this mean’ and becomes instead ‘now that it is clear what this means, what is the legal consequence.’ ” Id. at 251. Applying this principle, we observed that the label contained a “factually accurate, unambiguous statement of geographic origin,” prominently stating that it was “Puerto Rican Rum.” Id. at 252. As a consequence, we concluded that no reasonable consumer could be misled by the “Havana Club” name when it was considered in the context of this prominent truthful statement on the label. Id. at 252–53. Consumer survey evidence was therefore immaterial because the Lanham Act does not prohibit a claim that “reasonable people would have to acknowledge is not false or misleading.” Id. at 253. But we cautioned that judges should not “lightly disregard” consumer surveys because they may reveal “potential ambiguities in an advertisement” that show reasonable consumers may in fact be misled by the advertisement. Id. at 254–55. Finally, we noted that “a district court's decision to disregard survey evidence is reviewable de novo, since it is founded on a legal conclusion based on underlying facts, that is that no reasonable consumer would be misled by an advertisement.” Id. at 255 n. 18.

As our discussion of Pernod demonstrates, it is readily distinguishable from the issue before us here. Unlike Pernod, the case before us involves claims of literal falsity, so evidence of actual consumer deception is not required. See Novartis, 290 F.3d at 586. By disregarding the consumer survey evidence in this case, the District Court did not make the same legal conclusion we recognized in Pernod: that no consumers could be misled by the advertisement. The District Court instead made a factual finding about what the claim means and that its message is clear and unambiguous.

Pernod does not license courts to use consumer survey evidence to define the meaning of words in an advertising claim. In fact, our analysis in Pernod recognized that words may be used plainly enough and carry baseline meanings such that consumer survey evidence is irrelevant. See 653 F.3d at 251 (discussing Mead Johnson & Co. v. Abbott Labs., 201 F.3d 883, 886 (7th Cir.2000), opinion amended on denial of reh'g, 209 F.3d 1032 (7th Cir.2000) (explaining that “never before has survey research been used to determine the meaning of words, or to set the standard to which objectively verifiable claims must be held”)). In this case, Euro–Pro plainly explained on the packaging what it meant by its claim, so we are puzzled by Euro–Pro's characterization of the District Court's approach as a court inserting its “own perception” ahead of consumer perception. Appellant's Opening Br. at 33. Far from using its own perception of the claim's meaning, the District Court used the definition provided by Euro–Pro in the reference, and, concluding that Euro–Pro's message was explicit and unambiguous, it reasonably declined to substitute the uninformed first impressions of consumers about the claim's meaning. See Mead Johnson, 201 F.3d at 886. Euro–Pro chose a definition for steam power and now must live with it. It cannot use a consumer survey to create an ambiguity out of whole cloth. Accordingly, we conclude that the District Court did not err in failing to consider the Ford Survey in its analysis.

Turning to the “# 1 MOST POWERFUL STEAM” claim, we again agree with the District Court that this claim unambiguously conveys that Shark steam irons deliver more powerful steam than Rowenta steam irons. Unlike the “MORE POWERFUL STEAM vs. Rowenta” claim, however, the relevant message here is not explicit. The corresponding reference to the “# 1 MOST POWERFUL STEAM” claim states that the Shark steam irons “[o]ffer[ ] more grams per minute ... when compared to leading competition in the same price range,” and the parties agree that Rowenta steam irons are in a different price range. But, as we discussed earlier, a literally false claim may also be conveyed by necessary implication when considering the advertisement in its entirety. See Novartis, 290 F.3d at 586–87. The question here is whether, “based on a facial analysis of the product name or advertising, ... the consumer will unavoidably receive a false message.” Id. at 587. Here, the answer is yes. The “# 1 MOST POWERFUL STEAM” claim appears directly above the “MORE POWERFUL STEAM vs. Rowenta” claim, and the proximity of the two claims necessarily and unavoidably conveys a message that Shark steam irons offer the most powerful steam, even when compared to Rowenta steam irons. We therefore cannot say the District Court's finding is clearly erroneous.

2.

Having decided that the claims convey unambiguous messages, the next question is whether those messages are false. We find no clear error in the District Court's determination that the messages are false. The District Court reasonably relied on SEB's internal test results and the SLG Test Report. Both tests measured steam power in grams per shot and grams per minute—the measurements for steam power provided on the Shark packaging—in accordance with independent, objective standards promulgated by the IEC. Both tests also showed that the Rowenta steam irons either outperformed or performed as well as the Shark steam irons. Moreover, the Kemal Report acknowledged that there is no difference in grams per shot of steam between the Shark steam irons and the respective Rowenta steam irons. Put simply, all the relevant evidence before the District Court refuted Euro–Pro's claims of superiority.

Euro–Pro makes one final argument in an effort to overcome the District Court's finding of literal falsity. According to Euro–Pro, the District Court improperly shifted the burden of proof away from SEB to Euro–Pro. In addition to the rule that the party seeking preliminary relief bears the burden of satisfying the four-factor test, ECRI v. McGraw–Hill, Inc., 809 F.2d 223, 226 (3d Cir.1987), the general rule in false advertising cases is that a plaintiff bears the burden of proving falsity, Novartis, 290 F.3d at 589. But in Novartis, we recognized an exception to the general rule and held that “a court may find that a completely unsubstantiated advertising claim by the defendant is per se false without additional evidence from the plaintiff to that effect.” Id. at 590. Euro–Pro argues that the Novartis exception only applies when a defendant refuses to present any evidence to support the truth of its claim. According to Euro–Pro, unlike the defendant in Novartis, Euro–Pro provided “robust” and uncontroverted evidence—the Kemal Report. Appellant's Opening Br. at 24.

We do not read Novartis so narrowly. Euro–Pro fails to appreciate that the Kemal Report is mostly irrelevant to the messages actually conveyed by the Shark claims. The Kemal Report's primary conclusion is that the Shark steam irons have more powerful steam than the respective Rowenta steam irons when steam power is measured by calculating the kinetic energy of a steam burst over the burst's duration. But Euro–Pro does not, and cannot, argue that the Kemal Report supports the claims that the Shark steam irons offer more powerful steam measured in grams per shot or grams per minute than the respective Rowenta steam irons. In fact, as we previously mentioned, the Kemal Report concedes that the Shark steam irons deliver the same grams per shot of steam as the respective Rowenta steam irons. Therefore, Euro–Pro's claims are entirely unsubstantiated and exactly like the claims in Novartis.

But even though Novartis permits a finding of falsity based on Euro–Pro's failure to come forward with any evidence to support its claims, we note that the District Court relied on SEB's affirmative showing of falsity at least as much as it relied on Euro–Pro's failure to substantiate its claims. We therefore cannot say that the District Court shifted the burden of proof at all. Thus, the District Court's finding that the claims are false is not clearly erroneous.

Accordingly, the District Court correctly decided that SEB established a likelihood of success on the merits.

B.

Euro–Pro next argues that the District Court erred by finding that SEB established a likelihood of irreparable harm in the absence of preliminary relief.

We recently clarified the standard for irreparable harm in Lanham Act cases in Ferring Pharmaceuticals, Inc. v. Watson Pharmaceuticals, Inc., 765 F.3d 205 (3d Cir.2014). There, we held that “a party seeking a preliminary injunction in a Lanham Act case is not entitled to a presumption of irreparable harm but rather is required to demonstrate that she is likely to suffer irreparable harm if an injunction is not granted.” Id. at 217. Of particular relevance to this case, our analysis in Ferring relied in large part on the Supreme Court's decision in eBay Inc. v. MercExchange, L.L.C., 547 U.S. 388, 126 S.Ct. 1837, 164 L.Ed.2d 641 (2006). In eBay, the Supreme Court held that “broad classifications” and “categorical rule[s]” about when injunctions should issue are inconsistent with exercising “equitable discretion” pursuant to traditional equitable principles. 547 U.S. at 393–94, 126 S.Ct. 1837. Like the Patent Act at issue in eBay, “[t]he Lanham Act's injunctive relief provision is premised upon traditional principles of equity.” Ferring, 765 F.3d at 214 (citing 15 U.S.C. § 1116(a)). It follows that a presumption, or categorical rule, of irreparable harm in Lanham Act cases is inconsistent with exercising discretion according to traditional equitable principles. Id. at 215–16. Our decision in Ferring also emphasized that courts may award preliminary injunctive relief only upon a “ ‘ clear showing ’ ” of a likelihood of irreparable harm. Id. at 217 (quoting Winter, 555 U.S. at 22, 129 S.Ct. 365).

Although the District Court below did not have the benefit of our holding in Ferring, it presciently declined to apply a presumption of irreparable harm, at least overtly. The District Court decided that:

[A]lthough the Third Circuit held in [ Kos Pharmaceuticals, Inc. v. Andrx Corp., 369 F.3d 700, 726 (3d Cir.2004) ] that a court may find a presumption of irreparable injury if a likelihood of success on the merits is proved, the Supreme Court's later decisions in eBay and Winter ... indicate that such a presumption no longer exists in the Lanham Act context. Consequently, [SEB] bears the burden of showing that it is likely to suffer irreparable harm in the absence of preliminary injunctive relief.
J.A. at A17.

In spite of the District Court's express disavowal of a presumption, Euro–Pro contends that the District Court erred by applying a relaxed standard and a de facto presumption in determining that SEB demonstrated a likelihood of irreparable harm. On the one hand, we agree with Euro–Pro that portions of the District Court's opinion may be construed as applying a relaxed standard. For example, the District Court stated that a likelihood of irreparable harm is proven if a plaintiff establishes “a reasonable basis for the belief that it is likely to be damaged as a result of the false advertising.” J.A. at A18 (internal alterations and quotation marks omitted). In Novartis, we rejected this very standard. 290 F.3d at 595. The District Court also cited repeatedly to a case that relied, at least in part, on a presumption of irreparable harm. J.A. at A18–20 (citing W.L. Gore & Assocs., Inc. v. Totes Inc., 788 F.Supp. 800, 811 (D.Del.1992)). But other parts of the District Court's opinion, including its detailed discussion of the specific claims, the relationship between the competing products, and SEB's explanation of the likely injury to the Rowenta brand's reputation, as well as its conclusion that SEB “convincingly demonstrated” a likelihood of irreparable harm, are consistent with Ferring. J.A. at A18–20. It is therefore unclear whether the District Court's reference to the wrong standard actually affected the substance of its analysis.

We need not dwell on the question, however, because even if the District Court erred by reciting and applying the wrong standard, we may uphold the District Court's finding of a likelihood of irreparable harm if it is supported by sufficient evidence in the record. See Novartis, 290 F.3d at 595–96. Here, the record contains sufficient evidence of likely harm to the Rowenta brand's reputation and SEB's goodwill. See S & R Corp. v. Jiffy Lube Int'l, Inc., 968 F.2d 371, 378 (3d Cir.1992) (“Grounds for irreparable injury include loss of control of reputation, loss of trade, and loss of goodwill.”). The District Court credited the testimony of Pollard, SEB's marketing director, that Rowenta steam irons enjoy strong reputations among retailers and consumers as the premier steam-iron brand on the market, and that this first-rate reputation is the result of substantial SEB investments in advertising, promotion, and product development. In addition, the District Court found that Rowenta and Shark steam irons compete against each other, that they are often sold side-by-side on retail shelves, and that relative steam power is an important factor for consumers. And most importantly, the District Court credited Pollard's testimony that the claims on the Shark steam irons, which, to be clear, are “literally false, unsubstantiated comparative claims that identify its competitor by name,” would likely harm the Rowenta brand's reputation among retailers and consumers, especially because Shark steam irons are lower-priced. J.A. at A19–20. Finally, the District Court found that the harm to SEB would be impossible to calculate monetarily.

By finding that SEB established a likelihood of irreparable harm, we are not connecting these facts using a veiled presumption of irreparable harm. Ferring bars such a presumption; we emphasize, however, that Ferring does not bar drawing fair inferences from facts in the record. Indeed, a key lesson from Ferring is that courts considering whether to grant injunctive relief must exercise their equitable discretion in a case-by-case, fact-specific manner. A critical aspect of fact-finding in this and other contexts is drawing reasonable inferences from facts in the record. See generally Anderson v. City of Bessemer City, 470 U.S. 564, 574, 105 S.Ct. 1504, 84 L.Ed.2d 518 (1985) (explaining that the clearly erroneous standard of review applies to findings that rest on “inferences from other facts”). The inference drawn by Pollard, the District Court, and now this Court—that SEB is likely to suffer irreparable harm to its brand reputation and goodwill—is supported not by a general rule or presumption but by the literally false comparative advertising claims at issue, the competitive relationship between the parties and products, and the judgment of Pollard that the harm to SEB's brand reputation and goodwill is impossible to quantify. Nor does Ferring change the rule that harm to reputation and goodwill constitutes irreparable harm, so long as the plaintiff makes a clear showing. Based on the facts of this case, we conclude that SEB clearly showed a likelihood of irreparable harm to its brand reputation and goodwill.

Accordingly, any error committed by the District Court was harmless because there is sufficient evidence in the record to support a finding that SEB is likely to suffer irreparable harm without preliminary relief.

C.

Euro–Pro's final challenge is to the constitutionality and scope of the District Court's injunction. “District Courts are afforded considerable discretion in framing injunctions.” Meyer v. CUNA Mut. Ins. Soc., 648 F.3d 154, 169 (3d Cir.2011). At the same time, an injunction “should be ‘no more burdensome to the defendant than necessary to provide complete relief to plaintiffs.’ ” Novartis, 290 F.3d at 598 (quoting Califano v. Yamasaki, 442 U.S. 682, 702, 99 S.Ct. 2545, 61 L.Ed.2d 176 (1979)). “Moreover, because commercial speech is entitled to appropriate protection under the First Amendment, an injunction restraining allegedly false or misleading speech must be narrowly tailored to cover only the speech most likely to deceive consumers and harm [the plaintiff].” Id. (alteration in original) (internal quotation marks omitted).

Here, the District Court's order granting the preliminary injunction requires Euro–Pro to place stickers over the “MORE POWERFUL STEAM vs. Rowenta” and the “# 1 MOST POWERFUL STEAM” claims on both the Shark 405 and the Shark 505. Also, the order directs Euro–Pro to remove the hang tags from the steam irons.

Commercial speech conveying a literally false message is not protected by the First Amendment. See id. (“We conclude that the injunction does not violate the First Amendment ... because each of these messages is false.”). As we have explained, we agree with the District Court's conclusion that SEB will likely prevail on its false advertising claims. Therefore, we see no First Amendment violation.

Euro–Pro contends that the District Court's injunction is overbroad because it requires Euro–Pro to cover the advertising claims themselves rather than only the references to the claims. Euro–Pro correctly points out that the references are critical to the literal falsity analysis. Without the definitions from the references, the claims about relative steam power may be considered ambiguous, and as such, could not be literally false. See id. at 587. Thus, Euro–Pro argues that the injunction should have targeted only the references.

We disagree with Euro–Pro's narrow characterization of its advertising claims. Although the references provide the definition for steam power that the District Court appropriately adopted in this case, the references and the advertising claims together compose the literally false messages. Therefore, the injunction is not overbroad because it is limited to reaching claims that are literally false. See Castrol Inc. v. Pennzoil Co., 987 F.2d 939, 949 (3d Cir.1993). Moreover, the logic underlying Euro–Pro's argument would create an unworkable framework. Under Euro–Pro's suggested approach, district courts could not just enjoin the dissemination of literally false advertising claims, but they also would need to parse each part of those literally false claims to see if the removal of a word or a portion here and there would render the remainder true. We cannot say that the District Court abused its discretion when it required Euro–Pro to place stickers over the entirety of the false advertising claims rather than only part of them.

IV. CONCLUSION

For the foregoing reasons, we find no abuse of discretion by the District Court granting SEB's motion for a preliminary injunction. Accordingly, we will affirm.