Ann Bell v. Albertson Companies, Inc.

• Court: Court of Appeals for the Seventh Circuit
• Decided: December 7, 2020
• Docket: 19-2741
• Status: Published

Opinion

In the

United States Court of Appeals For the Seventh Circuit ____________________ Nos. 19-2581 & 19-2741 ANN BELL, et al., Plaintiffs-Appellants, v.

PUBLIX SUPER MARKETS, INC., et al., Defendants-Appellees. ____________________

ANN BELL, et al., Plaintiffs-Appellants,

v.

ALBERTSON COMPANIES, INC., et al., Defendants-Appellees. ____________________

Appeals from the United States District Court for the Northern District of Illinois, Eastern Division. No. 1:16-cv-05802 — Gary Feinerman, Judge. ____________________

ARGUED SEPTEMBER 17, 2020 — DECIDED DECEMBER 7, 2020 ____________________ 2 Nos. 19-2581 & 19-2741

Before KANNE and HAMILTON, Circuit Judges.* HAMILTON, Circuit Judge. This case is about Parmesan cheese—specifically the kind sold in familiar shaker tubes in grocery stores across the country. The defendants sell these products with labels advertising them as “100% Grated Par- mesan Cheese.” The plaintiffs say these products are not 100 percent cheese, but rather contain between four and nine per- cent added cellulose powder and potassium sorbate, as is ev- ident to a consumer who takes the time to read the fine print of an ingredient list on the back of the package. Plaintiffs claim that these ingredient lists show that the prominent “100%” labeling is deceptive under state consumer-protection laws. The Judicial Panel on Multidistrict Litigation transferred numerous similar actions to the Northern District of Illinois for consolidated pretrial proceedings under 28 U.S.C. § 1407. Plaintiffs then reorganized their claims into five amended consolidated complaints, organized by defendant. In a series of orders, the transferee district court ultimately dismissed the plaintiffs’ deceptive labeling claims (“the 100% claims”) with prejudice for failure to state a claim. Plaintiffs appeal those dismissals. With respect to three of the plaintiffs’ consolidated com- plaints, the 100% claims should have survived the defendants’ motion to dismiss. Plaintiffs have plausibly alleged that the prominent “100%” labeling deceives a substantial portion of reasonable consumers, and their claims are not preempted by

*Then-Circuit Judge Barrett was a member of the panel when this case was submitted but did not participate in the decision and judgment. The appeal is resolved by a quorum of the panel pursuant to 28 U.S.C. § 46(d). Nos. 19-2581 & 19-2741 3

federal law. For reasons specific to the management of the multidistrict litigation, however, we lack appellate jurisdic- tion to review the district court’s dismissal of the 100% claims in two of the plaintiffs’ consolidated complaints (against Pub- lix and Target/ICCO) because the appeals were filed too late. In Part I, we address the merits of the 100% claims. In Part II, we explain why we lack appellate jurisdiction over the district court’s dismissal of the latter two complaints. I. The 100% Claims A. The Defendants’ Products and Plaintiffs’ Claims The defendants are manufacturers of cheese and major food retailers: Kraft Heinz, the ICCO-Cheese Company, Tar- get, Wal-Mart, SuperValu, Albertson’s, and Publix Supermar- kets. All sell grated cheese products that are sold from the gro- cery aisles rather than from refrigerated dairy cases. The cheese products are prominently labeled on the front “100% Grated Parmesan Cheese.” (Plaintiffs also complain about a few products labeled “100% Grated Parmesan & Romano Cheese” or “100% Grated Three Cheese Blend,” but those var- iations do not matter for our analysis.) On the back or side, the products include the required list of ingredients in fine print. Those lists show that they contain cellulose powder and potassium sorbate to prevent the grated cheese from caking and getting moldy, respectively. Plaintiffs bought the defendants’ products with the “100%” labels and allege they were deceived. Plaintiffs con- tend that defendants’ prominent claims that their products are “100% Grated Parmesan Cheese” are deceptive because they are likely to mislead a significant portion of reasonable consumers, who will focus on the prominent “100%” on the 4 Nos. 19-2581 & 19-2741

front labels without checking the fine print on the back show- ing that the products are not 100% cheese. Plaintiffs assert these claims under fourteen state con- sumer protection statutes spanning ten states.1 These statutes are known as “Little-FTC Acts” because they are patterned on the Federal Trade Commission Act (FTCA). See Henry N. But- ler & Joshua D. Wright, Are State Consumer Protection Acts Re- ally Little-FTC Acts?, 63 Fla. L. Rev. 163, 165 (2011). The Little- FTC Acts broadly prohibit unfair business practices, includ- ing deceptive advertising. Unlike the federal act, however, these state statutes provide private rights of action to comple- ment enforcement by the government. These statutes “all require plaintiffs to prove that the rele- vant labels are likely to deceive reasonable consumers,” which “requires a probability that a significant portion of the general consuming public or of targeted consumers, acting reasonably in the circumstances, could be misled.” Beardsall v.

1 Plaintiffs bring claims under the Alabama Deceptive Trade Practices Act,

Ala. Code § 8-19-1 et seq.; California Consumers Legal Remedies Act, Cal. Civ. Code § 1750 et seq.; California Unfair Competition Law, Cal. Bus. & Prof. Code § 17200 et seq.; Connecticut Unfair Trade Practices Act, Conn. Gen. Stat. § 42-110b; Florida Deceptive and Unfair Trade Practices Act, Fla. Stat. § 501.201 et seq.; Illinois Consumer Fraud and Deceptive Business Practices Act, 815 ILCS 505/2; Michigan Consumer Protection Act, Mich. Comp. Laws § 445.903 et seq.; Minnesota Unlawful Trade Practices Act, Minn. Stat. § 325D.09 et seq.; Minnesota Deceptive Trade Practices Act, Minn. Stat. § 325D.44 et seq.; Minnesota False Statement in Advertising Act, Minn. Stat. § 325F.67; Minnesota Prevention of Consumer Fraud Act, Minn. Stat. § 325F.68 et seq.; Missouri Merchandising Practices Act, Mo. Rev. Stat. § 407.010 et seq.; New Jersey Consumer Fraud Act, N.J. Stat. Ann. § 56:8-1 et seq.; and New York Consumer Protection from Deceptive Acts and Practices Act, N.Y. Gen. Bus. Law §§ 349 & 350. Nos. 19-2581 & 19-2741 5

CVS Pharmacy, Inc., 953 F.3d 969, 972−73 (7th Cir. 2020), quot- ing Ebner v. Fresh, Inc., 838 F.3d 958, 965 (9th Cir. 2016); Suchanek v. Sturm Foods, Inc., 764 F.3d 750, 756−57 (7th Cir. 2014). While these are all state statutes, the federal Class Ac- tion Fairness Act of 2005 has pushed many class actions under them into federal courts. See 28 U.S.C. § 1332(d). The core pro- hibitions of these laws are interpreted for the most part inter- changeably, and the parties have not identified any differ- ences relevant to these appeals. We concentrate on the general prohibition against advertising that is likely to deceive a sub- stantial proportion of reasonable consumers. B. The District Court’s Decision The district court dismissed the plaintiffs’ 100% claims on two grounds. First, the court found that the prominent “100%” claims on the front labels are ambiguous and that a consumer who seeks clarity can find it by reading the ingre- dient list on the back label. Second, the court found that com- mon sense would tell a reasonable consumer that, despite the 100% claims, these cheese products must contain added ingre- dients because they are sold unrefrigerated in the main gro- cery aisles, alongside dried pastas and canned sauces. As to the first ground, the district court applied a new am- biguity rule that it derived from some cases, discussed below, that applied the reasonable consumer test. The court wrote: “Where a plaintiff contends that certain aspects of a product’s packaging are misleading in isolation, but an ingredient la- bel … would dispel any confusion, the crucial issue is whether the misleading content is ambiguous; if so, context can cure the ambiguity and defeat the claim.” In re 100% Grated Parmesan Cheese Marketing and Sales Practices Litig., 275 F. Supp. 3d 910, 922 (N.D. Ill. 2017). The court examined the 6 Nos. 19-2581 & 19-2741

“100% Grated Parmesan Cheese” label and deemed it ambig- uous. In the court’s view, the phrase could denote that the product is 100% grated, 100% Parmesan, or 100% cheese (or perhaps any two of those). Id. at 923. This ambiguity, the court said, could be cleared up by reading the ingredient list, which would show that the defendants’ products are not 100% cheese, let alone 100% Parmesan cheese. The court thus held that defendants’ “100% Grated Parmesan Cheese” labels are not deceptive as a matter of law. Id. As to the second ground, the court said that common sense dictates that the defendants’ products must contain ingredi- ents other than cheese because they are shelf-stable without refrigeration. Courts have used similar common-sense argu- ments to dismiss labeling claims when no reasonable con- sumer could actually believe the plaintiffs’ alleged interpreta- tion. The California breakfast cereal cases are good examples of this, where courts dismissed claims that the colorful rings and balls of cereal shown on “Froot Loops” and “Cap’n Crunch” boxes promised real fruit within. See Werbel v. Pep- sico, Inc., 2010 WL 2673860, at *6 (N.D. Cal. July 2, 2010); McKinnis v. Kellogg USA, 2007 WL 4766060, at *6 (C.D. Cal. Sept. 19, 2007). Here, the district court thought common sense similarly should defeat plaintiffs’ nothing-but-cheese belief: “Cheese is a dairy product, after all, and reasonable consum- ers are well aware that pure dairy products spoil, grow blue, green, or black fuzz, or otherwise become inedible if left un- refrigerated for an extended period of time.” 100% Grated Par- mesan Cheese, 275 F. Supp. 3d at 923. On appeal, defendants offer a third defense, federal preemption, which the district court did not reach. We disa- gree with all three grounds for dismissal. The plaintiffs’ 100% Nos. 19-2581 & 19-2741 7

labeling claims survive the defendants’ motion to dismiss. Part I-C explains our disagreement with the district court’s reasons for dismissal. Part I-D then explains why plaintiffs’ claims are not preempted by federal law. C. State-Law Deception Under the Reasonable Consumer Standard The plaintiffs’ 100% claims were dismissed on the plead- ings under Rule 12(b)(6), so we accept the plaintiffs’ factual allegations and ask only whether they present “a claim to re- lief that is plausible on its face.” Bell Atlantic Corp. v. Twombly, 550 U.S. 544, 570 (2007). The plaintiffs need only “nudge[] their claims across the line from conceivable to plausible.” Id. In this case, plaintiffs’ claims survive if they have plausibly alleged that the defendants’ front labels likely lead a signifi- cant portion of reasonable consumers to falsely believe some- thing that the back labels belie. We agree with the district court that the context of the entire packaging is relevant. We disagree about applying that principle in this case as a matter of law on the pleadings. What matters most is how real con- sumers understand and react to the advertising. We therefore join our colleagues in at least three other circuits in holding that an accurate fine-print list of ingredients does not fore- close as a matter of law a claim that an ambiguous front label deceives reasonable consumers. Many reasonable consumers do not instinctively parse every front label or read every back label before placing groceries in their carts. 1. Ambiguity and Fine Print on the Back The district court’s dismissal erred by departing from the Rule 12(b)(6) standard and attributing to ordinary supermar- ket shoppers a mode of interpretation more familiar to judges 8 Nos. 19-2581 & 19-2741

trying to interpret statutes in the quiet of their chambers. See Dumont v. Reily Foods Co., 934 F.3d 35, 40 (1st Cir. 2019) (re- versing dismissal of similar deceptive labeling claim on pleadings: “Our dissenting colleague envisions a more eru- dite reader of labels, tipped off by the accent grave on the word ‘crème,’ and armed perhaps with several dictionaries, a bit like a federal judge reading a statute.”). Under the district court’s ambiguity rule, as a matter of law, a front label cannot be deceptive if there is any way to read it that accurately aligned with the back label. And this would be so even if the label actually deceived most consumers, and even if it had been carefully designed to deceive them. We respectfully part company with the approach of our colleague on the district court. Consumer-protection laws do not impose on average consumers an obligation to question the labels they see and to parse them as lawyers might for am- biguities, especially in the seconds usually spent picking a low-cost product. See, e.g., Danone, US, LLC v. Chobani, LLC, 362 F. Supp. 3d 109, 123 (S.D.N.Y. 2019) (“[A] parent walking down the dairy aisle in a grocery store, possibly with a child or two in tow, is not likely to study with great diligence the contents of a complicated product package, searching for and making sense of fine-print disclosures … . Nor does the law expect this of the reasonable consumer.”). With the time afforded by litigation, we can see how “100% Grated Parmesan Cheese” might be interpreted as claiming only that whatever it contains is “100% grated,” or perhaps that whatever cheese it contains is “100% Parmesan.” Another reading, though, and certainly a plausible reading, is that “100%” applies to all three words: it’s all cheese; all the cheese is Parmesan, and it’s all grated. Plaintiffs say they are Nos. 19-2581 & 19-2741 9

prepared to offer evidence that a substantial portion, perhaps even a large majority, of consumers interprets the defendants’ prominent front labels that way. The district court’s ambiguity rule conflicts with decisions of the First, Second, and Ninth Circuits in very similar cases also involving food labels. See Dumont, 934 F.3d at 41; Manti- kas v. Kellogg Co., 910 F.3d 633, 638−39 (2d Cir. 2018); Williams v. Gerber Products Co., 552 F.3d 934, 939 (9th Cir. 2008). In each case, the defendant argued that its fine-print, back-label in- gredient list immunized it from claims that more prominent front-label claims were deceptive because a consumer could read the ingredient list. In each case, a district court had granted a motion to dismiss, and in each case, the appellate court reversed. In each case, there was room to argue about the precise meaning of the front-label claims, such as hazel- nuts in coffee beans in Dumont, cheese crackers made with “Whole Grain” in Mantikas, and fruit juice in drinks for young children in Williams. And in reversing each case, our col- leagues in the other circuits held that the reasonable con- sumer standard does not presume, at least as a matter of law, that reasonable consumers will test prominent front-label claims by examining the fine print on the back label. We agree with that reasoning in Dumont, Mantikas, and Williams. The ambiguity rule for front-label claims would, we fear, encourage deceptive advertising and labeling. Lots of advertising is aimed at creating positive impressions in buy- ers’ minds, either explicitly or more subtly by implication and indirection. And lots of advertising and labeling is ambigu- ous. Deceptive advertisements often intentionally use ambi- guity to mislead consumers while maintaining some level of deniability about the intended meaning. We agree with the 10 Nos. 19-2581 & 19-2741

Second Circuit that a rule that immunized any ambiguous la- bel so long as it is susceptible to one non-deceptive interpre- tation “would validate highly deceptive advertising.” Manti- kas, 910 F.3d at 638. Sticking to the reasonable consumer standard avoids this temptation and stays in touch with real consumer behavior. We stand by the general principle that deceptive advertis- ing claims should take into account all the information avail- able to consumers and the context in which that information is provided and used. See, e.g., Davis v. G.N. Mortgage Corp., 396 F.3d 869, 884 (7th Cir. 2005) (affirming summary judg- ment for defendants on claims that mortgage terms were not made sufficiently clear); see also Beardsall, 953 F.3d at 977−78 (affirming summary judgment for defendants where plain- tiffs testified that they expected products to contain preserva- tives despite label claiming “100% pure” aloe); Fink v. Time Warner Cable, 714 F.3d 739, 742 (2d Cir. 2013) (affirming dis- missal of claims that advertising of internet speeds was de- ceptive; “under certain circumstances, the presence of a dis- claimer or similar clarifying language may defeat a claim of deception”), citing Freeman v. Time, Inc., 68 F.3d 285, 289−90 (9th Cir. 1995) (affirming dismissal of challenge to sweep- stakes mailer where mailer explicitly said plaintiff would win only if he had winning number). And where plaintiffs base deceptive advertising claims on unreasonable or fanciful interpretations of labels or other ad- vertising, dismissal on the pleadings may well be justified. See, e.g., Carrea v. Dreyerʹs Grand Ice Cream, Inc., 475 F. Appʹx 113, 115 (9th Cir. 2012) (“It is implausible that a reasonable consumer would interpret ‘Original Sundae Cone,’ ‘Original Nos. 19-2581 & 19-2741 11

Vanilla,’ and ‘Classic,’ to imply that Drumstick is more whole- some or nutritious than competing products.”); Red v. Kraft Foods, Inc., 2012 WL 5504011, at *3 (C.D. Cal. Oct. 25, 2012) (“Made with Real Vegetables” label on box of crackers could not reasonably mean crackers were “composed of primarily fresh vegetables”); Werbel, 2010 WL 2673860, at *3 (colorful “crunch berries” on Cap’n Crunch box did not reasonably promise fresh fruit); McKinnis, 2007 WL 4766060, at *4 (same for Froot Loops). In addition, nothing we say in this opinion is intended to foreclose defendants from offering evidence to show that con- sumers are not actually misled by their “100% Grated Parme- san Cheese” labels. See Dumont, 934 F.3d at 41 (reversing dis- missal on pleadings where label was ambiguous: “None of this is to say that our dissenting colleague’s reading is by any means unreasonable … . That being said, we think it best that six jurors, rather than three judges, decide.”); Mantikas, 910 F.3d at 638−39 (same where “Whole Grain” cracker label could be reasonably read as promising either predominantly whole grain, or just some whole grain); Williams, 552 F.3d at 939 (same where “packaging picture[d] a number of different fruits, potentially suggesting (falsely) that those fruits or their juices” were included) (emphasis added). This practical and fact-intensive approach to consumer be- havior under deceptive-advertising laws is consistent with the law’s approach under other consumer-protection laws. The ambiguity rule applied in this case, by contrast, departs dramatically from that generally applicable approach. Under the Federal Trade Commission Act, for example, the kind of ambiguity rule applied in the district court here has long been rejected: 12 Nos. 19-2581 & 19-2741

It is well established, and critical to the notion of preventing false advertising, that where an ad- vertisement conveys more than one meaning, one of which is false, the advertiser is liable for the misleading variation. Murray Space Shoe Corp. v. FTC, 304 F.2d 270, 272 (2d Cir. 1963); Rhodes Pharmacal Co. v. Federal Trade Commission, 208 F.2d 382, 387 (7th Cir. 1953), modified by re- stating the Commission’s order, 348 U.S. 940 (1953). Nor does our application of that princi- ple in this instance rest upon some mere seman- tic quibble or strained interpretation of words, since that meaning of respondents’ claim which deceives is one which is likely to be understood, and reasonably so, by consumers. A more ap- propriate statement of the principle in this case might thus be that an otherwise false advertisement is not rendered acceptable merely because one possi- ble interpretation of it is not untrue. Nat’l Commʹn On Egg Nutrition v. FTC, 570 F.2d 157, 161 n.4 (7th Cir. 1977) (emphasis added, quoting FTC order); accord, e.g., FTC v. Sterling Drug, Inc., 317 F.2d 669, 674−75 (2d Cir. 1963). Similarly, in trademark and trade-dress claims under the Lanham Act, “likelihood of confusion” is the test, and it is an intensely factual question based on real market conditions and real consumers’ behavior. E.g., AM General Corp. v. Daim- lerChrysler Corp., 311 F.3d 796, 812, 825−28 (7th Cir. 2002); Thomas & Betts Corp. v. Panduit Corp., 138 F.3d 277, 296 (7th Cir. 1998); Roulo v. Russ Berrie & Co., 886 F.2d 931, 937 (7th Cir. 1989); James Burrough Ltd. v. Sign of Beefeater, Inc., 540 F.2d 266, Nos. 19-2581 & 19-2741 13

275 (7th Cir. 1976) (trademarks “must be compared in the light of what occurs in the marketplace, not in the court- room”). Under that test, one key factor is the “degree of care likely to be used by consumers.” E.g., AM General, 311 F.3d at 812. In applying that test, we have often stressed that consum- ers are likely to exhibit a low degree of care when purchasing low-priced, everyday items. Id. at 828; Roulo, 886 F.2d at 937 (“the marketing channels for greeting cards do not promote a careful analysis of the product prior to purchase”). The same holds true with the low-cost groceries here. Similarly, in false advertising claims under the Lanham Act, the law has long recognized that even literally true claims may deceive, that implied messages in advertising may de- ceive, and that what matters is how consumers actually un- derstand the advertising. As a district judge, Judge St. Eve summarized these principles using cases from this court: Federal false advertising claims generally fall into two categories: literal falsity and implied falsity. Where a statement or claim made in ad- vertising is literally false, “the plaintiff need not show that the statement either actually deceived customers or was likely to do so.” Hot Wax, Inc. v. Turtle Wax, Inc., 191 F.3d 813, 819 (7th Cir. 1999); B. Sanfield, [Inc. v. Finlay Fine Jewelry Corp., 168 F.3d 967, 971 (7th Cir. 1999)]. Where a state- ment or claim is literally true or ambiguous, however, a plaintiff must prove that the state- ment “implicitly convey[s] a false impression, [is] misleading in context, or likely to deceive consumers.” Hot Wax, 191 F.3d at 819; see also B. Sanfield, 168 F.3d at 971 (plaintiff must show that 14 Nos. 19-2581 & 19-2741

statement is “misleading in context, as demon- strated by actual consumer confusion).” (quot- ing BASF Corp. v. Old World Trading Co., 41 F.3d 1081, 1089 (7th Cir. 1994)). In other words, “[a] statement is misleading when, although liter- ally true, it implies something that is false.” Ab- bott Labs. v. Mead Johnson & Co., 971 F.2d 6, 13 (7th Cir. 1992). Regardless of the theory advanced by the plain- tiff, “whether a claim is either ‘false’ or ‘mislead- ing’ is an issue of fact rather than law.” Mead Johnson & Co. v. Abbott Labs., 209 F.3d 1032, 1034 (7th Cir. 2000); Abbott Labs., 971 F.2d at 13–15; Castrol, Inc. v. Pennzoil Co., 987 F.2d 939, 943–45 (3d Cir. 1993); Johnson & Johnson * Merck Con- sumer Pharma. Co. v. Smithkline Beecham Corp., 960 F.2d 294, 298 (2d Cir. 1992); see also Walker v. Natʹl Recovery, Inc., 200 F.3d 500, 503 (7th Cir. 1999) (discussing confusion in the context of the FDCA, citing Lanham Act precedent and hold- ing that “[w]hether a given message is confus- ing is … a question of fact, not of law or logic.”). Similarly, “[w]hether consumers are likely to be confused … is ultimately a question of fact” that “may be resolved on summary judgment only ‘if the evidence is so one-sided that there can be no doubt about how the question should be an- swered.’” AutoZone[, Inc. v. Strick, 543 F.3d 923, 929 (7th Cir. 2008)] (quoting Packman v. Chicago Tribune Co., 267 F.3d 628, 642 (7th Cir. 2001); Door Sys., Inc. v. Pro–Line Door Sys., Inc., 83 F.3d 169, 171 (7th Cir. 1996)). “It is not for the judge Nos. 19-2581 & 19-2741 15

to determine, based solely upon his or her own intuitive reaction, whether the advertisement is deceptive.” Johnson & Johnson * Merck, 960 F.2d at 297–98. Rather, the question is “what does the person to whom the advertisement is addressed find to be the message? That is, what does the public perceive the message to be?” Id. (internal citations omitted). LG Electronics U.S.A., Inc. v. Whirlpool Corp., 661 F. Supp. 2d 940, 948 (N.D. Ill. 2009). Similarly, courts often encounter questions about consum- ers’ responses to ambiguity under the Fair Debt Collection Practices Act, 15 U.S.C. § 1692 et seq. Our interpretation of that Act has been informed by our treatment of consumer be- havior under the Lanham Act and other consumer-protection statutes. Carefully crafted, deliberately ambiguous language can violate the debt collection act by misleading consumer- debtors who are entitled to clarity. Pantoja v. Portfolio Recovery Assocs., LLC, 852 F.3d 679, 687 (7th Cir. 2017) (affirming sum- mary judgment for plaintiff), citing Janetos v. Fulton Friedman & Gullace, LLP, 825 F.3d 317, 323 (7th Cir. 2016) (reversing summary judgment for defendant), citing in turn Chuway v. National Action Financial Servcies, Inc., 362 F.3d 944, 947–48 (7th Cir. 2004) (reversing summary judgment for defendant). A plaintiff can prevail by showing, as a matter of fact, that ambiguous language actually misleads debtors. E.g., Lox v. CDA, Ltd., 689 F.3d 818, 822, 824–25 (7th Cir. 2012) (reversing summary judgment for defendant), following Gonzales v. Ar- row Financial Services, LLC, 660 F.3d 1055, 1063 (9th Cir. 2011) (affirming summary judgment for debtor; debt collector could 16 Nos. 19-2581 & 19-2741

not use conditional language to avoid liability based on de- ceptive impression given to debtors). More generally under that act, courts try to take care to avoid confusing a consumer’s understanding of a dunning letter with a federal judge’s understanding: “The intended re- cipients of dunning letters are not federal judges, and judges are not experts in the knowledge and understanding of unso- phisticated consumers facing demands by debt collectors. We are no more entitled to rely on our intuitions in this context than we are in deciding issues of consumer confusion in trademark cases, where the use of survey evidence is rou- tine.” Evory v. RJM Acquisitions Funding LLC, 505 F.3d 769, 776 (7th Cir. 2007) (reversing summary judgment for debt collec- tor in relevant part). While the “unsophisticated consumer” standard that we apply under the debt collection law, see id. at 774, is not identical to the “reasonable consumer” standard under the deceptive advertising statutes, in both arenas courts focus on how real consumers understand the carefully crafted messages aimed at them. See also Chuway, 362 F.3d at 948 (re- versing summary judgment for defendant where ambiguous language could mislead “significant fraction” of consumers). Our comparisons to other consumer-protection laws are intended to show that there is nothing out of the ordinary about applying the state consumer-protection laws in this case based on how consumers actually understand defend- ants’ labels. These questions may not be answered as a matter of law simply because lawyers can construe an ambiguous claim in a way that would not be deceptive. Plaintiffs are en- titled to present evidence on how consumers actually under- stand these labels. Nos. 19-2581 & 19-2741 17

Plaintiffs allege in their complaints that they have actually conducted consumer surveys showing that 85 to 95 percent of consumers understood “100% Grated Parmesan Cheese” to mean that the product contains only cheese, without additives like those defendants use in the challenged products. In op- posing dismissal, plaintiffs have also offered affidavits from linguists offering opinions that the most natural and plausible reading of defendants’ labels is entirely Parmesan cheese that is grated, rather than the district court’s suggested “100% grated” meaning.2

2 We do not address here any issues about the ultimate admissibility of any party’s survey evidence in future proceedings. Cf. Evory, 505 F.3d at 776 (relying on trademark precedents, survey evidence regarding debt collection cases must meet professional standards). In opposing a motion to dismiss for failure to state a claim, a plaintiff may describe the evidence she expects to offer to support factual allega- tions, and nothing prevents a plaintiff from including such allegations in the complaint or from tendering affidavits or other documents to show that those expectations about evidence are realistic. E.g., Geinosky v. City of Chicago, 675 F.3d 743, 745–46 n.1 (7th Cir. 2012); Thomas v. Guardsmark, Inc., 381 F.3d 701, 704 (7th Cir. 2004) (denying motion to strike new mate- rials submitted on appeal); Hrubec v. Nat’l R.R. Passenger Corp., 981 F.2d 962, 963–64 (7th Cir. 1992) (“A plaintiff need not put all of the essential facts in the complaint. He may add them by affidavit or brief—even a brief on appeal.”); Early v. Bankers Life & Cas. Co., 959 F.2d 75, 79 (7th Cir. 1992) (reversing dismissal; plaintiff is free to assert new facts in brief opposing motion to dismiss); Derfus v. City of Chicago, 42 F. Supp. 3d 888, 893 n.1 (N.D. Ill. 2014); Roe I v. Bridgestone Corp., 492 F. Supp. 2d 988, 1007 (S.D. Ind. 2007) (documents submitted in opposing motion to dismiss “are not evidence, but they provide a way for a plaintiff to show a court that there is likely to be some evidentiary weight behind the pleadings the court must evaluate”). It would often be prudent to accompany such eviden- tiary materials with a clear statement that they are offered not to convert the motion to one for summary judgment under Rule 12(d) but only to 18 Nos. 19-2581 & 19-2741

We see no basis for disregarding plaintiffs’ allegations on these points in deciding the motions to dismiss. What matters here is how consumers actually behave—how they perceive advertising and how they make decisions. These are matters of fact, subject to proof that can be tested at trial, even if as judges we might be tempted to debate and speculate further about them. We doubt it would surprise retailers and market- ers if evidence showed that many grocery shoppers make quick decisions that do not involve careful consideration of all information available to them. See, e.g., U.S. Food & Drug Ad- min., Guidance for Industry: Letter Regarding Point of Purchase Food Labeling (Oct. 2009) (“FDA’s research has found that with [Front of Package] labeling, people are less likely to check the Nutrition Facts label on the information panel of foods (usu- ally, the back or side of the package).”); Karen Bradshaw Schulz, Information Flooding, 48 Ind. L. Rev. 755, 782 (2015) (when terms “like ‘low-fat’ and ‘multi-grain’ were written in big, bright letters on foods,” consumers would “focus on the bright claim rather than turning the box around to read the dull, black-and-white nutrition label on the back”), citing Timothy Muller, Structural Information Factors Which Stimulate the Use of Nutrition Information: A Field Experiment, 22 J. Mktg. Res. 143 (1985); Alan M. White, Behavior and Contract, 27 Law & Inequality 135, 158 (2009) (“Billions are spent every year on behavioral research by the marketing industry in order to un- derstand consumer biases and heuristics. Armed with this in- formation, marketers engage in various strategies to increase sales.”).

illustrate the allegations that are the proper focus of the Rule 12(b)(6) mo- tion. Geinosky, 675 F.3d at 745–46 n.1. Nos. 19-2581 & 19-2741 19

Defendants also argue that even the “100% cheese” mean- ing is not misleading, despite the presence of the added cellu- lose and potassium sorbate. They contend that the FDA’s def- inition of “grated cheese” allows them to call their products “grated cheese” because that designation allows addition of cellulose and potassium sorbate. See 21 C.F.R. § 133.146(c). The defense theory seems to be that if the FDA defines “grated cheese” in a way that allows added cellulose and po- tassium sorbate, their products with those additives thus qualify as “100% grated cheese.” We have no quarrel with de- fendants’ ability to call their products “grated cheese.” The problem lies in the “100%,” especially since the pleadings pro- vide reason to think that consumers understand “100% grated cheese” to mean that the cheese does not have the additives. And how could a manufacturer of grated cheese without those additives differentiate its product from these defendants’ products if, pace George Orwell, the courts said the products with the additives could lawfully claim to be “100% cheese”? We address the preemption implications of this argument below in Part I-D, but average consumers are not likely to be aware of the nuances of the FDA’s regulations defining “grated cheese.” When consumers read “100% Grated Parme- san Cheese,” plaintiffs plausibly allege, they are not likely to understand that the phrase really means that the product is a blend of cheese and other substances that regulations allow to be called “cheese.” Rather, both plain meaning and the plain- tiffs’ surveys and linguists plausibly indicate that a significant portion of consumers read the labels as promising pure cheese without added ingredients. 20 Nos. 19-2581 & 19-2741

2. Common Sense Does Not Solve the Problem “Common sense” does not render plaintiffs’ reading un- reasonable. The district court found that no reasonable con- sumer who read the label as promising pure cheese could ac- tually believe it because common sense tells us that pure cheese cannot be shelf stable without refrigeration. That rea- soning strayed from the standard for dismissal for failure to state a claim under Rule 12(b)(6). First, as plaintiffs point out, pure grated Parmesan cheese can be shelf-stable for a long time without refrigeration. See U.S. Dep’t of Agriculture, Does all cheese need to be refrigerated? (July 17, 2019), https://ask.usda.gov/s/article/Does-all-cheese- need-to-be-refrigerated (“[H]ard cheeses such as … grated Parmesan do not require refrigeration for safety, but they will last longer if kept refrigerated.”). One of Kraft’s own patents claims: “Fully cured Parmesan cheese is very hard and keeps almost indefinitely.” U.S. Patent No. 6,242,016 B1, 1. For mil- lennia, people have been making cheese to preserve milk for later consumption, all without modern refrigeration. See, e.g., Casey Quakenbush, Archaeologists Have Discovered the World’s Oldest Cheese Inside an Ancient Egyptian Tomb, TIME (Aug. 20, 2018), https://time.com/5371503/ancient-egypt-tomb-old- cheese/. Moreover, today’s grocery shoppers can often spot unre- frigerated cartons of pure grated Parmesan sold beside the cheese wheels that source them. It is true the defendants’ products are not displayed in that way but are instead shelved in the main grocery aisles. But since pure grated Parmesan can be and sometimes is sold unrefrigerated, common sense is not a substitute here for evidence, and certainly not as a matter of law. Reasonable consumers’ expectations about the need to Nos. 19-2581 & 19-2741 21

use either additives or refrigeration cannot be decided as a matter of law. A comparison to our recent opinion in Beardsall may help illustrate the point. Plaintiffs there complained of labels on aloe vera gels proclaiming “100% aloe vera gel” despite the use of added preservatives that were disclosed in the fine- print lists of ingredients. Beardsall, 953 F.3d at 977. We af- firmed summary judgment for the defendants because the plaintiffs failed to offer evidence of actual deception. The Beardsall plaintiffs failed to offer any consumer surveys show- ing the 100% claim was misleading, and the named plaintiffs admitted in their depositions that they actually expected that the products would include preservatives. Id. We did not suggest, however, that the 100% claim could have been deemed not deceptive on the pleadings. To the con- trary, we cast doubt on the defendants’ argument that such 100% claims were not misleading because the back label could clarify any ambiguity: “We are skeptical of defendants’ posi- tion … that an asterisk pointing to an ingredient list in fine print could save virtually any deceptive slogan claiming pu- rity.” 953 F.3d at 977. The result is different in this case be- cause of the difference between a motion to dismiss on the pleadings and a motion for summary judgment. How reason- able consumers actually understand defendants’ “100% Grated Parmesan Cheese” labels is a question of fact that can- not be resolved on the pleadings. D. Federal Preemption The defendants argue that we should affirm on an alterna- tive ground even if their 100% labels could be deemed mis- leading under state law. Defendants contend that such state- 22 Nos. 19-2581 & 19-2741

law claims are preempted because federal law permits their labeling. The district court did not reach this issue, but it was briefed fully in the district court and on appeal. We find it prudent to resolve this issue of law now and perhaps head off another round of appeals in the multidistrict litigation. Defendants’ argument has two distinct branches. First, they point out that the federal Food, Drug, and Cosmetic Act (FDCA) and its accompanying regulations expressly bar states from imposing labeling requirements that are not “identical” to the FDCA’s, and they contend plaintiffs seek to use state law to impose different labels on them. Second, de- fendants say the FDA approved Kraft’s use of the “100% Grated Parmesan Cheese” label in 1999 and 2000, thus render- ing the plaintiffs’ challenge both conflict-preempted and barred by state-law safe harbor provisions. These arguments do not persuade us. The first reads the FDCA’s express preemption provision too broadly. The second fails at the first step because defendants have not shown that the FDA ap- proved Kraft’s “100%” labeling as nondeceptive. 1. Express Preemption The FDCA delegates to the FDA the power to “promulgate regulations fixing and establishing for any food … a reason- able definition and standard of identity.” 21 U.S.C. § 341. These standards of identity determine what a food product must contain to be marketed under a certain name. They also often establish requirements for how the product must be made or sold. The express preemption provision of the FDCA bars states from “directly or indirectly establish[ing] under any author- ity … any requirement for a food which is the subject of a Nos. 19-2581 & 19-2741 23

standard of identity … that is not identical to such standard of identity or that is not identical to the requirement of section 343(g)” of the Act. 21 U.S.C. § 343-1(a)(1). The FDA’s standard of identity for “grated cheeses” allows the defendants to add anticaking agents (cellulose powder) and antimycotics (po- tassium sorbate) and to call the product “grated cheese.” 21 C.F.R. § 133.146(c). In fact, the standard of identity requires defendants to call their products “Grated Parmesan Cheese.” “If only one variety of cheese is used, the name of the food is ‘grated ____ cheese’, the name of the cheese filling the blank.” 21 C.F.R. § 133.146(d)(3)(i). Defendants point out that the standard of identity does not explicitly prohibit them from labeling their products as “100%” cheese. From this silence, they argue that any state- law ruling that prohibited them from using the phrase “100% cheese” would establish a requirement not identical to those set out in the FDA’s standard of identity. The conclusion does not follow from the statute because defendants read the express preemption provision too broadly. The problem for the defense is that nothing in the standard of identity says whether products named “Grated Parmesan Cheese” may be labeled with an additional modi- fier such as “100%.” The key question becomes what that si- lence means for the FDCA’s preemption provision. In Turek v. General Mills, Inc., 662 F.3d 423 (7th Cir. 2011), we held that the FDCA preempted the plaintiffs’ attempt to use state law to require that disclosure language be added to a food label when federal regulations did not explicitly require it. “Even if the disclaimers that the plaintiff wants added would be consistent with the requirements imposed by the Food, Drug, and Cosmetic Act, consistency is not the test; 24 Nos. 19-2581 & 19-2741

identity is.” Id. at 427. Thus, when a standard of identity lists labeling disclosures that are affirmatively required, state law may not tack on further required disclosures that the federal standard does not mention. That would “establish … a[] re- quirement … not identical to” the federal standard. 21 U.S.C. § 343-1(a)(1); see also Nemphos v. Nestle Waters North America, Inc., 775 F.3d 616, 625 (4th Cir. 2015) (affirming dismissal; state-law claim seeking remedy of additional warning on la- bel for excessive fluoride in bottled water was preempted). In this case, however, it is defendants’ labels, not plaintiffs’ claims, that depart from the standard of identity. Plaintiffs are not seeking to add labeling requirements. They seek only to stop defendants from voluntarily adding deceptive language to the federally permitted labels. The federal standard of iden- tity does not address that language, the “100%.” If state law prevents such deception, it will not establish any new require- ment different from the standard of identity. In fact, the FDCA already provides generally that “a food shall be deemed to be misbranded” if its labeling is “false or mislead- ing in any particular.” 21 U.S.C. § 343(a)(1). Absent contrary language in a standard of identity that protects a particular statement, § 343-1 does not expressly preempt state-law pro- hibitions on deceptive statements that sellers add voluntarily to their labels or advertising. See, e.g., Durnford v. Muscle- Pharm Corp., 907 F.3d 595, 603−04 (9th Cir. 2018) (reversing dismissal of state-law claim that seller’s label made false claim about source of protein in supplement where federal stand- ards were silent about source of protein); Hawkins v. Kroger Co., 906 F.3d 763, 770−71 (9th Cir. 2018) (reversing dismissal of state-law claim that food label’s claim of “zero trans fat” was deceptive where federal regulations did not expressly au- thorize that claim outside fine-print list of nutritional facts); Nos. 19-2581 & 19-2741 25

Astiana v. Hain Celestial Group, Inc., 783 F.3d 753, 758 (9th Cir. 2015) (reversing dismissal of state-law claim that cosmetics la- bel’s use of “natural” was deceptive where federal regulations did not address use of “natural”). By the logic of these decisions, state-law claims challeng- ing defendants’ voluntary addition of “100%” to their labels are not preempted. A simple example illustrates this limit for express preemption. If an FDCA standard of identity said that a vegetable label must indicate the vegetable’s “color, date of harvest, and common name,” the preemption provision would prohibit a state from adding a further requirement that all vegetable labels also list the country of origin. The absence of such a requirement in the federal law operates to exclude it. But if a seller chose voluntarily to add a country of origin— and lied about it—then § 343-1(a)(1) would not preempt state law from requiring the seller to remove the voluntarily-added lie. After all, there are all sorts of potentially misleading addi- tions that standards of identity do not explicitly ban. See Durnford, 907 F.3d at 603−04 (seller’s false claim about source of protein was not expressly authorized by federal regula- tions, so state-law claim for deception was not preempted); Astiana, 783 F.3d at 758 (same where cosmetics seller’s decep- tive use of “natural” was not expressly authorized by federal regulations). Closer to this case, as in our hypothetical standard for a vegetable, the actual standard for grated cheese says nothing about a cheese’s country of origin. Suppose a defendant here labeled its product “Grated Parmesan Cheese, 100% from It- aly.” If the cheese did not actually come from Italy, state-law claims for deceptive advertising would not be preempted 26 Nos. 19-2581 & 19-2741

simply because the federal standard of identity does not ex- plicitly ban such a statement. Such a result would stretch the FDCA’s “not identical to” language for express preemption beyond its breaking point. The FDCA’s preemption provision means that, while states may not require sellers to add further labeling that is not required by federal law, they may prevent sellers from volun- tarily adding deceptive content that is not required by federal law. Plaintiffs’ “100%” claims are thus not preempted by § 343-1(a)(1).3

3 Plaintiffs’ briefs suggest that the front labels could also be fixed if they read “100% Grated Parmesan Cheese with additives and preservatives.” That remedy would be preempted. It would require the defendants to list optional ingredients on the front label—an added requirement that con- flicts with federal law. Recall that the FDCA also preempts states from es- tablishing a labeling requirement “not identical to the requirement of sec- tion 343(g).” 21 U.S.C. § 343-1(a)(1). Section 343(g) provides that “a food shall be deemed to be misbranded … if it purports to be … a food for which a definition and standard of identity has been prescribed … unless (1) it conforms to such definition and standard, and (2) its label bears the name of the food … and, insofar as may be required by such regulations, the common names of optional ingredients (other than spices, flavoring, and col- oring) present in such food.” 21 U.S.C. § 343(g) (emphasis added). If FDA regulations do not require a certain food label to include optional ingredi- ents, then state law may not impose such a requirement. In general, FDA regulations do not require optional ingredients on front labels. See 21 C.F.R. § 130.11 (requiring it only when both a particular standard of identity requires it and it would “significantly differentiate between two or more foods that comply with the same standard”). The standard of identity governing grated cheeses does not require optional ingredients to be declared on the front label. Unlike some standards that do require this (e.g., Macaroni products, 21 C.F.R. § 139.110(f)), the stand- ard for grated cheese uses common-form language mandating only that each ingredient be declared “as required by the applicable sections of Nos. 19-2581 & 19-2741 27

2. Conflict Preemption and Safe Harbors This brings us to the defendants’ last line of defense in these appeals. They contend that the FDA actually approved Kraft’s use of the “100% Grated Parmesan Cheese” label in 1999 and 2000. The argument distorts the facts. The FDA’s standard of identity for Parmesan cheese requires that the cheese have been aged for at least ten months. In 1999, the FDA issued Kraft a temporary permit “to market test a prod- uct designated as ‘100% Grated Parmesan Cheese’ that devi- ates from the U.S. standards of identity for Parmesan cheese and grated cheeses” in that it used “a different enzyme tech- nology that fully cures the cheese in 6 months rather than 10 months.” 64 F.R. 16743-01 (1999), citing 21 C.F.R. § 133.165, which requires a ten-month curing period for Parmesan cheese. In 2000, the FDA extended the permit. 65 F.R. 83040- 01 (2000). Kraft needed these permits to experiment with a six- month curing period. And that was what the FDA approved. There is no indication in these permits that the FDA assessed whether Kraft’s proposed “100% Grated Parmesan Cheese” label was nondeceptive under federal law, and certainly not that it approved the 100% claim. The FDA only noted in pass- ing that Kraft’s “products will bear the name ‘100% Grated

parts 101 and 130” of the FDA regulations. 21 C.F.R. § 133.146(e). Part 130 includes the general regulation disclaiming the need to list optional ingre- dients on a front label. Part 101 requires that “chemical preservative[s]” such as cellulose powder or potassium sorbate be disclosed only some- where “on the food or on its container.” 21 C.F.R. § 101.22(c). Because the FDCA and its accompanying regulations explicitly disclaim the defend- ants’ need to disclose their use of cellulose or potassium sorbate on the front label, § 343-1 preempts state laws that would require that disclosure. 28 Nos. 19-2581 & 19-2741

Parmesan Cheese’.” 65 F.R. 83040-01. The FDA’s approval of Kraft’s shorter six-month curing period therefore poses no conflict with the plaintiffs’ state-law labeling claims: “Conflict preemption applies when there is an actual conflict between state and federal law such that it is impossible for a person to obey both.” Nelson v. Great Lakes Educational Loan Servs., Inc., 928 F.3d 639, 646 (7th Cir. 2019). Likewise, the plaintiffs’ claims are not barred by state-law safe harbor provisions embedded in the Little-FTC Acts. These safe harbor provisions prevent federal-state conflicts by protecting business practices that the federal government has specifically approved. The Illinois Consumer Fraud Act, for example, bars claims based on actions that are “specifically authorized by laws administered by any regulatory body or officer acting under statutory authority of this State or the United States.” 815 ILCS 505/10b(1). Safe harbor provisions like these do not foreclose the plaintiffs’ claims because, again, the FDA did not authorize the defendants’ “100% Grated Parmesan Cheese” label as nondeceptive. See Vanzant v. Hill’s Pet Nutrition, Inc., 934 F.3d 730, 738 (7th Cir. 2019) (holding safe harbor provision did not apply where an FDA policy guide did not “specifically authorize” the defendant’s prescription labeling on its pet food). For these reasons, plaintiffs’ state-law claims are not preempted to the extent that they would require the defend- ants to remove “100%” from their labels—leaving only “Grated Parmesan Cheese.” The FDCA allows defendants to name their products “Grated Parmesan Cheese” even though they contain ingredients other than cheese. And state law may not require the defendants to disclose optional ingredients on the front label. But if the plaintiffs ultimately succeed on the Nos. 19-2581 & 19-2741 29

merits, state law could prevent the defendants from voluntar- ily including the “100%” labeling. II. Appellate Jurisdiction We have jurisdiction over the appeals addressing the 100% claims within three of the plaintiffs’ consolidated complaints, but not over the appeals arising from the remaining two. The district court dismissed the plaintiffs’ 100% claims in three consolidated complaints, those against Kraft Heinz Co., Al- bertsons Co., Albertsons LLC, SuperValu Inc., Wal-Mart Stores, Inc. and ICCO-Cheese Co. Other claims remain pend- ing in the district court in those cases, but the district court invoked Federal Rule of Civil Procedure 54(b) to enter partial final judgments on those claims on the ground that there was no just reason for delay. This was a practical and sensible use of Rule 54(b) to enable review of the 100% claims, which seem to be the core of the overall litigation, without forcing the par- ties and the district court to work through all the other claims first. Ironically, however, we encounter problems with our ap- pellate jurisdiction in the two cases in which all claims were dismissed as to all parties: the cases against Publix and against Target/ICCO. In those cases, plaintiffs filed their no- tices of appeal too late to invoke our jurisdiction. We reach this split result for reasons specific to how this multidistrict litigation was managed. To explain, we first recount the key events in the district court, then note some of the challenges of managing final judgments and appeals in multidistrict liti- gation, and finally explain why we lack jurisdiction over the appeals against Publix and Target/ICCO. 30 Nos. 19-2581 & 19-2741

A. Procedural History In an order issued August 24, 2017, the district court first dismissed all the plaintiffs’ 100% labeling claims, but without prejudice to filing amended complaints. See 100% Grated Par- mesan Cheese Marketing and Sales Practices Litig., 275 F. Supp. 3d 910, 927 (N.D. Ill. 2017). Plaintiffs then filed five amended consolidated class action complaints, which were organized around the defendants. See Dkt. 225 against Kraft Heinz Co.; Dkt. 226 against Publix Super Markets, Inc.; Dkt. 227 against Albertsons Co., Albertsons LLC, and SuperValu Inc.; Dkt. 228 against Target Corp. and ICCO-Cheese Co.; Dkt. 229 against Wal-Mart Stores, Inc. and ICCO-Cheese Co. Defendants re- sponded to those consolidated complaints with a fresh round of motions to dismiss on the pleadings, setting off a new round of briefing on the merits. The district court ruled on those motions on November 1, 2018, dismissing all of the 100% labeling claims within these consolidated complaints, as well as all other claims in the Tar- get/ICCO complaint. See 100% Grated Parmesan Cheese, 348 F. Supp. 3d 797, 818 (N.D. Ill. 2018) (“The 100% claims are dis- missed in their entirety, as are all Anticaking claims against Target/ICCO.”). The November 1, 2018 order denied dismis- sal of some “anticaking” claims against several defendants. Because the Publix complaint brought only 100% labeling claims, the November 1, 2018 order dismissed all claims against all parties in both the Publix and Target/ICCO com- plaints. The November 1, 2018 order did not say explicitly that these dismissals were with prejudice, but it ended: “No Anti- caking claims may proceed against Publix and Target, which are dismissed from this litigation.” Id. Despite this language, Nos. 19-2581 & 19-2741 31

the district court did not at that time enter a separate written judgment in favor of Publix or Target. The plaintiffs tried to treat the November 1, 2018 order as if it had been a dismissal without prejudice by filing a motion under Rule 15(a)(2) for leave to file amended complaints against Publix and Target/ICCO. See Dkt. 306. Those defend- ants immediately opposed the motion to amend, arguing that the November 1, 2018 order had dismissed all claims against them and amounted to final judgments in their favor. Dkt. 324 at 5–6 (Dec. 21, 2018) (Publix); Dkt. 325 at 6–7 (Dec. 21, 2018) (Target/ICCO). Defendants argued that the liberal amend- ment standards of Rule 15 therefore did not apply and that the final judgments in favor of Publix and Target/ICCO could be set aside only under Rules 59 or 60, which plaintiffs had not invoked and could not satisfy. Accordingly, whether the November 1, 2018 order had amounted to a final judgment in favor of Publix and Target/ICCO was openly disputed from the outset, and in time for the parties or the court to resolve any ambiguity. No one did that, however. On July 16, 2019, the district court denied plaintiffs’ mo- tion to file amended complaints against Publix and Tar- get/ICCO. 100% Grated Parmesan Cheese Marketing and Sales Practices Litig., 393 F. Supp. 3d 745, 756 (N.D. Ill. 2019). In the July 16, 2019 order, the district court did not address whether it had already in effect entered final judgments in favor of Publix and Target/ICCO back on November 1, 2018, but with- out entering separate Rule 58 judgments. The court wrote that it “need not resolve the dispute because Plaintiffs’ motion fails even under the more lenient Rule 15(a)(2) standard.” Id. at 753–54. The court concluded: “The claims against Publix and 32 Nos. 19-2581 & 19-2741

Target/ICCO”—who were “dismissed as defendants” back in November 2018—“remain dismissed.” Id. at 751, 765. After the July 16, 2019 order, the court still did not enter any separate Rule 58(a) final judgments. It was not until Au- gust 26, 2019, that the court (upon joint requests by the par- ties) finally entered judgments based on its previous dismis- sals. The court entered four judgments. It entered Rule 58 judgments in three of the individual transferred cases within the Publix and Target/ICCO tracks. And it entered final judg- ment under Rule 54(b) “as to the ‘100% claims’ in the consoli- dated amended class action complaints [Dkts.] 225 227 228 229.” Missing from this list is Dkt. 226, the complaint against Publix, which contained only dismissed “100% claims” and thus had no remaining claims that brought it within Rule 54(b)’s purview. It is unclear why the list includes Dkt. 228, the Target/ICCO complaint. Both its 100% and anticaking claims had been dismissed in November 2018. Perhaps its in- clusion reflected the ongoing confusion over the effect of those dismissals. Or it could be that this Rule 54(b) judgment simply listed each complaint that asserted claims beyond the “100% claims” that the court was separating out for appeal. In any event, the intended effect of this judgment was to start the appeal clock on the 100% claims that had not yet been the sub- ject of appealable final judgments. As we explain in Part II-D below, the problem for plain- tiffs’ appeals against Publix and Target/ICCO is that, despite the absence of separate Rule 58 judgments accompanying the November 1, 2018 order that said those defendants were “dis- missed from this litigation,” more than 150 days had passed. In 2002, Federal Rule of Civil Procedure 58 and Federal Rule of Appellate Procedure 4(a) were amended to try to solve the Nos. 19-2581 & 19-2741 33

problems that can arise when a district court never gets around to entering a separate Rule 58(a) judgment. The solu- tion was to deem a final judgment to have been entered 150 days after the entry in the civil docket of the order that appar- ently ended the case and is being appealed. See Fed. R. Civ. P. 58(c); Fed. R. App. P. 4(a)(7); Committee Notes on Rules (2002 Amendment). That description fits the November 1, 2018 dis- missal of all claims in the Target/ICCO and Publix cases. The 150 days from November 1, 2018 ran on March 31, 2019, so we treat these judgments as if they were entered on Monday, April 1, 2019, the first business day after the 150th day. Plain- tiffs then had 30 more days, until May 1, to file notices of ap- peal from the November 1, 2018 dismissals of Publix and Tar- get/ICCO. Plaintiffs missed that deadline by more than three months. B. Final Judgments in Multidistrict Litigation The problem of appellate jurisdiction has an additional wrinkle here because these appeals arise from multidistrict litigation in which a district court and parties may use spe- cialized procedures to manage the pretrial proceedings in the related cases. Congress enacted 28 U.S.C. § 1407 in 1968 to manage more effectively complex sets of related lawsuits pending in multi- ple districts. The basic tool under § 1407(a) gave the new Ju- dicial Panel on Multidistrict Litigation the power to transfer related cases to one district court for “coordinated or consoli- dated pretrial proceedings.” Absent settlement or pretrial res- olution, § 1407(a) provides that such transferred actions “shall be remanded by the panel at or before the conclusion of such pretrial proceedings to the district from which it was trans- ferred … .” During the pretrial proceedings, however, the 34 Nos. 19-2581 & 19-2741

transferee court has extensive case management powers, and those powers are especially broad when the parties and court agree. Coordinated or consolidated pretrial proceedings can be used to streamline litigation, to manage discovery to avoid duplication and waste, and to narrow issues by agreement or by motions to dismiss or for summary judgment, and espe- cially by settlement. The default rule is that separate actions transferred for those pretrial proceedings retain their separate identities, especially for purposes of entering final judgments and pursuing appeals. Yet transferee courts and parties may choose to manage those cases in ways that can change that default rule and give up the separate identities of the original suits transferred to the MDL litigation. See, e.g., In re Refriger- ant Compressors Antitrust Litig., 731 F.3d 586, 588 (6th Cir. 2013) (one category of antitrust plaintiffs—indirect purchas- ers—filed a single “consolidated amended complaint” that combined all of their allegations; dismissal of some indirect purchasers’ claims were not final judgments while other claims asserted in the consolidated amended complaint re- mained pending). A related issue reached the Supreme Court in Gelboim v. Bank of America Corp., 574 U.S. 405 (2015). In Gelboim, the dis- trict judge dismissed just one of the many individual actions transferred and consolidated under the MDL’s umbrella. The plaintiffs in that case appealed, but the Second Circuit dis- missed their appeal for lack of a final judgment because other cases transferred to the MDL remained pending in the district court. The Supreme Court reversed the dismissal of the ap- peal. It held that the dismissal of the single complaint had been an appealable final judgment. The Court explained: Nos. 19-2581 & 19-2741 35

“Cases consolidated for MDL pretrial proceedings ordinarily retain their separate identities, so an order disposing of one of the discrete cases in its entirety should qualify under [28 U.S.C.] § 1291 as an appealable final decision.” 574 U.S. at 413. In other words, “When the transferee court overseeing pre- trial proceedings in multidistrict litigation grants a defend- ant’s dispositive motion ‘on all issues in some transferred cases, [those cases] become immediately appealable … while cases where other issues remain would not be appealable at that time.’” Id. at 415, quoting D. Herr, Multidistrict Litigation Manual § 9:21, p. 312 (2014). So the default rule is that each transferred case retains its separate identity for purposes of entering judgments and appealing. In footnote 3, however, Gelboim flagged a case-manage- ment choice that can change that default rule so that many transferred cases are fused together for purposes of final judg- ment and appeal: Parties may elect to file a “master complaint” and a corresponding “consolidated answer,” which supersede prior individual pleadings. In such a case, the transferee court may treat the master pleadings as merging the discrete ac- tions for the duration of the MDL pretrial pro- ceedings. In re Refrigerant Compressors Antitrust Litigation, 731 F.3d 586, 590–592 (C.A.6 2013). No merger occurs, however, when “the master complaint is not meant to be a pleading with le- gal effect but only an administrative summary of the claims brought by all the plaintiffs.” Id., at 590. 36 Nos. 19-2581 & 19-2741

574 U.S. at 413 n.3. This separate rule may apply to a tool that transferee judges and parties often use to manage multidis- trict litigation. As noted during oral argument in Gelboim, dis- trict courts often direct plaintiffs to file “a consolidated com- plaint … [a]nd then … simply resolve one of the claims in the master complaint without entering judgment in the individ- ual action … . The district court uses this administrative com- plaint to manage the litigation.” Tr. of Oral Arg. at 56, Gelboim, 574 U.S. 405 (2014) (No. 13-1174). We read note 3 in Gelboim as endorsing the Sixth Circuit’s approach in Refrigerant Compressors to ascertaining the legal effect of a consolidated complaint in multidistrict litigation, though the Sixth Circuit approach would persuade us even without the Supreme Court’s endorsement. That approach fo- cuses pragmatically on the behavior of the district court and the parties to determine whether they treated the consoli- dated complaint as the “operative pleading” or merely “an administrative summary.” 731 F.3d at 591. This approach holds the court and the parties to their actions to prevent them from springing traps by treating a consolidated complaint as the real complaint in the district court but then denying its importance and effect once a party tries to appeal. The test also avoids odd results that would flow from a contrary rule. As the Sixth Circuit posited: What at any rate would be the alternative? The [plaintiffs] suggest looking at their original in- dividual complaints rather than at the new con- solidated complaint. But the new complaint su- perseded the old ones, and it makes little sense to ascertain appellate jurisdiction by looking at the ghosts of departed pleadings. Besides, when Nos. 19-2581 & 19-2741 37

deciding whether the district court properly granted a motion to dismiss, we would have to look at the new complaint anyway. How odd it would be to write an opinion that talks about one complaint in the jurisdiction section and an- other in the merits section. Id. at 586. Under Refrigerant Compressors’ pragmatic inquiry, relevant signals include (1) how the plaintiffs labeled the new com- plaint, (2) whether the plaintiffs served the defendants with the new complaint instead of the original pleadings, (3) whether key deadlines were set in relation to the new com- plaint, (4) whether the court entertained motions to dismiss the consolidated complaint, and (5) whether the parties and the court looked solely to the allegations in the consolidated complaint when arguing and deciding such motions. See 731 F.3d at 590–91. This last factor is “perhaps most important[].” Id. at 591. We agree, especially with the emphasis on the par- ties’ and the district court’s focus on the allegations in the con- solidated complaint in arguing and deciding Rule 12 motions. But as the Sixth Circuit also explained, the dangers of am- biguity can be avoided if the court and the parties decide ex- plicitly, from the beginning, the legal status of the consoli- dated complaint(s). At the outset of a multidistrict litigation, all parties have an interest in knowing when and how appel- late rights may be triggered or lost. We urge district judges and MDL plaintiffs to indicate clearly whether a consolidated MDL complaint is to be treated as the operative pleading for purposes of judgment and appeal or instead as merely an ad- ministrative convenience. 38 Nos. 19-2581 & 19-2741

C. Legal Status of the Five Consolidated Complaints Such clarity was not reached in this case, so we must de- cide whether the consolidated complaints superseded the in- dividual pleadings to become the operative units for entry of judgments and appeals. Just as in Refrigerant Compressors, the relevant factors show that the five consolidated complaints here became the legally operative pleadings in the district court. The plaintiffs did not fashion this set of complaints as mere administrative summaries. Rather, they filed them as “amended consolidated” complaints. Dkt. 225–229. The court then held a status hearing to ask whether defendants “in- tend[ed] to answer or move to dismiss the amended com- plaints.” Dkt. 230. The defendants said they would move to dismiss, and the court set motions deadlines. Dkt. 232. The court invited and expected Rule 12 motions aimed at the new complaints. By this point, the judge and all parties were treat- ing them as the legally operative pleadings, and the defend- ants moved to dismiss the consolidated complaints, not the original pleadings. When litigating and adjudicating those motions, both sides and the court examined the sufficiency of only the consolidated complaints. Even after the court dis- missed all claims against all parties in the Publix and Tar- get/ICCO complaints, the plaintiffs continued to treat those complaints as the real complaints by seeking leave to amend them under Rule 15. D. Untimeliness of Publix and Target/ICCO Appeals Pulling these threads together, the consolidated Publix and Target/ICCO complaints were legally operative pleadings in two distinct cases under the umbrella of the larger MDL. Nos. 19-2581 & 19-2741 39

The district court’s November 1, 2018 order “had the hall- marks of a final decision.” See Gelboim, 574 U.S. at 414. That order effectively resolved all claims asserted in those two con- solidated complaints. Unlike the court’s August 2017 order, the November 2018 order did not expressly grant leave to amend. Most important, it said Publix and Target/ICCO were “dismissed from this litigation.” 100% Grated Parmesan Cheese, 348 F. Supp. 3d at 818. The order therefore qualified under Rule 54(a) as a judgment in these two cases, and separate Rule 58(a) judgments should have been entered at that time. Unfortunately, the district court did not enter separate Rule 58 judgments in those cases then. The failure to enter separate judgments means that the entry of judgment is de- termined, for purposes of the time to appeal, under the 150- day provision of Federal Rule of Civil Procedure 58(c) and Federal Rule of Appellate Procedure 4(a)(7). This 150-day fail- safe limit “ensure[s] that parties will not be given forever to appeal … when a court fails to set forth a judgment or order on a separate document.” Fed. R. App. P. 4(a)(7), Committee Notes on Rules (2002 Amendment). So we treat these judgments as if they were entered on Monday, April 1, 2019, the first business day after the 150th day. See Fed. R. Civ. P. 58(c)(2)(B); Fed. R. App. P. 4(a)(7)(A)(ii).4 That means the plaintiffs’ deadline to appeal

4 The plaintiffs’ November 28, 2018, motion to amend the Publix and Tar- get/ICCO complaints did not stop this 150-day clock. If the plaintiffs had filed motions under Rule 59 or 60, as the defendants argued might have been appropriate, then their appeal clock would have been tolled until the court decided the motions in July 2019. See Fed. R. App. P. 4(a)(4)(A)(iv– vi). But the plaintiffs framed their motion as one to amend under Rule 15(b), and the district court did not hold it was anything different, so it could not stop the clock. See Shields v. Illinois Dep’t of Corrections, 746 F.3d 40 Nos. 19-2581 & 19-2741

the dismissal of the operative Publix and Target/ICCO com- plaints passed thirty days later on May 1, 2019. See Fed. R. App. P. 4(a)(1)(A) (requiring notice of appeal within 30 days after entry of the judgment). Plaintiffs argue that we should not enforce the 150-day provision because defendants waived their right to insist on it. See Hamer v. Neighborhood Housing Servs. of Chicago, 897 F.3d 835, 839–40 (7th Cir. 2018) (enforcing waiver against defend- ants who admitted appeal was timely, then argued other- wise). There was no waiver here. As soon as plaintiffs filed their Rule 15(b) motion to amend in November 2018, Publix and Target/ICCO signaled their opposition and their view that the court’s November 1, 2018 order amounted to a final judgment in their cases. At that point, defendants teed up the uncertainty posed by the “dismissed from this litigation” term of that order and the court’s failure to enter separate judgments. Any party or the court could have acted promptly—long before the 150 days had run—to clarify the situation, but no one did. Sophisticated counsel on both sides apparently preferred to live with, and perhaps even to gamble on, the uncertainty. We have explained above how we resolve it. Because plaintiffs did not file their notices of appeal until August 2019, their appeals against Publix and Target/ICCO

782, 799 (7th Cir. 2014) (“Filing a post-judgment Rule 15(b) motion does not similarly toll the time to file an appeal.”). Nos. 19-2581 & 19-2741 41

are untimely, and we lack appellate jurisdiction over the dis- trict court’s dismissal of the claims in the consolidated com- plaints against Publix and Target/ICCO.5 Conclusion In Appeal No. 19-2741, the partial final judgments dismiss- ing the 100% claims alleged in the consolidated Kraft, Albert- sons, and Walmart/ICCO complaints are REVERSED and those cases are REMANDED for further proceedings con- sistent with this opinion. Appeal No. 19-2581, involving the consolidated complaints against Publix and Target/ICCO, is DISMISSED for lack of appellate jurisdiction.

5 Because we have no appellate jurisdiction over the claims against Publix,

we do not decide whether that complaint’s allegations of minimal diver- sity were enough to establish subject matter jurisdiction under the Class Action Fairness Act. 42 Nos. 19-2581 & 19-2741

KANNE, Circuit Judge, concurring. I join in Judge Hamil- ton’s opinion reversing the district court’s dismissal of Plain- tiffs’ deceptive-advertising claims. And I emphasize at the outset my agreement with Judge Hamilton’s well-articulated reasons for why we lack jurisdiction over the appeals against Publix and Target/ICCO and why Defendants’ various preemption arguments are unavailing. I have nothing to add with respect to those conclusions. I write separately only to further address today’s primary holding: that the district court erred in concluding that De- fendants’ “100% Grated Parmesan Cheese” labels cannot de- ceive a reasonable consumer as a matter of law. Although I agree with that holding, too, I think it worthwhile to further discuss its basis and contours for the benefit of district courts. The district court’s reasoned analysis thoroughly synthe- sized a mass of cases to come to an understandable, even if incorrect, rule: that ambiguity on a product’s front label can be cured by clarity on its back. But ultimately, I believe this rule is impractical and conflicts with the more abundant and persuasive authorities that properly distinguish questions of law and questions of fact in the deceptive-advertising context. It’s impractical because, while lawyers and judges can find ambiguity in just about anything, that’s not what we expect of the reasonable consumer. Under the district court’s ap- proach, defendants could often proffer some alternative, non- deceptive reading of the front label—or fashion the label pre- cisely so that it can bear one plausibly non-deceptive read- ing—regardless of whether the reasonable consumer (or some significant portion of reasonable consumers) would read it that way. And the court might admit some degree of ambigu- ity, in which case the ingredient list dictates the outcome—all Nos. 19-2581 & 19-2741 43

while reasonable consumers are actually deceived by the label without ever even noticing any ambiguity. That, at bottom, is the flaw in the district court’s rule: a court could decide as a matter of law that a statement is not de- ceptive even where it could deceive reasonable consumers as a matter of fact. It assumes reasonable consumers not only no- tice ambiguities but then investigate to resolve them, either by scouring the fine print or, even less likely, reading up on the shelf life of Parmesan cheese. It assumes too much. Moreover, the district court’s ambiguity rule conflicts with the weight of authority. It’s well settled that a label is not de- ceptive as a matter of law when the plaintiff’s interpretation is so facially illogical, implausible, or fanciful that no reason- able consumer would think it—and that dismissal is war- ranted in those circumstances. See, e.g., Bober v. Glaxo Wellcome PLC, 246 F.3d 934, 940 (7th Cir. 2001) (“As a matter of law, none of the three statements … is deceptive” because the label “eliminates any possibility of deception” and “can only be read” in a nondeceptive way.); see also Forouzesh v. Starbucks Corp., 714 F. App’x 776, 777 (9th Cir. 2018) (“The statutory claims fail as a matter of law because no reasonable consumer would think … that a 12-ounce ‘iced’ drink, such as iced coffee or iced tea, contains 12 ounces of coffee or tea and no ice.”); Fink v. Time Warner Cable, 714 F.3d 739, 741 (2d Cir. 2013). I consider this principle to stem directly from the motion to dis- miss standard, which requires that a claim to relief be “plau- sible on its face” to survive dismissal. Bell Atlantic Corp. v. Twombly, 550 U.S. 544, 570 (2007). Just as important, however, is the corollary to this princi- ple: that if a plaintiff’s interpretation of a challenged state- ment is not facially illogical, implausible, or fanciful, then a 44 Nos. 19-2581 & 19-2741

court may not conclude that it is nondeceptive as a matter of law. The “determination of the likelihood of deception ‘is an impressionistic one more closely akin to a finding of fact than a conclusion of law.’” Beardsall v. CVS Pharmacy, Inc., 953 F.3d 969, 973 (7th Cir. 2020) (Hamilton, J.) (quoting Suchanek v. Sturm Foods, Inc., 764 F.3d 750, 762 (7th Cir. 2014)). So unless the plaintiff’s reading is implausible on its face, it is for a panel of “jurors, rather than three judges, [to] decide on a full record whether the challenged label ‘has the capacity to mislead’ rea- sonably acting … consumers.” Dumont v. Reily Foods Co., 934 F.3d 35, 41 (1st Cir. 2019) (quoting Aspinall v. Philip Morris Cos., Inc., 813 N.E.2d 476, 488 (Mass. 2004)). The district court here did not conclude that Plaintiffs’ in- terpretation of the “100% Grated Parmesan Cheese” state- ment is illogical, implausible, or fanciful. Rather, in conclud- ing that the statement is ambiguous, it necessarily found the opposite: that reasonable consumers may interpret the state- ment in multiple, plausible ways. Ambiguity, Black’s Law Dic- tionary (11th ed. 2019) (“An uncertainty of meaning … that gives rise to any of two or more quite different but almost equally plausible interpretations.”). Thus, whether the 100% statement is likely to deceive a reasonable consumer is a question of fact that cannot be de- termined on the pleadings. And Judge Hamilton thoroughly explains how the weight of authority suggests that the pres- ence of an ingredient label does not transform this question of fact into a question of law. Beardsall, 953 F.3d at 978 (“[D]is- closure in an ingredient list cannot cure a clearly misleading statement.” (citing Williams v. Gerber Prods. Co., 552 F.3d 934, 939 (9th Cir. 2008))). Nos. 19-2581 & 19-2741 45

This is not news to many courts, including those within this circuit. “Although there are situations where the determi- nation of the possibility for deception can [be] made as a mat- ter of law,” one court recently explained, “the relevant ques- tions in this case—whether a reasonable consumer would be deceived by the products’ name and label and a reasonable consumer’s understanding of the [label]—are not so clear- cut.” Korte v. Pinnacle Foods Grp., LLC., No. 17-CV-199-SMY- SCW, 2018 WL 1508855, at *4 (S.D. Ill. Mar. 27, 2018). That court therefore declined to conclude as a matter of law that the label was not deceptive, even though “the ingredient list on the back of the bottle discloses that it contains … other in- gredients.” Id. at *3. “[T]he allegedly deceptive act must be looked upon in light of the totality of the information made available to the plaintiff,” id. at *4 (quoting Davis v. G.N. Mortg. Corp., 396 F.3d 869, 884 (7th Cir. 2005)), and “whether listing other ingredients on the back would be sufficient to dispel any misunderstanding … the front of the package might cause is a question of fact,” id. Our opinion approves this approach. Defendants make much of the word “clearly” in Beardsall and Williams, which they argue suggests that an ingredient la- bel can cure a misleading statement so long as the statement is not clearly misleading. That one word is simply too weak a basis to depart from the signals projected by cases from the First, Second, and Ninth Circuits—cases in which, as Judge Hamilton correctly notes, “there was room to argue about the precise meaning of the front-label claims.” Determining that a statement is not “clearly misleading” on the pleadings robs the jury of the opportunity to determine, as a matter of fact, whether the statement is “clearly misleading,” just “mislead- ing,” or “not misleading at all.” It also it flips the proper 46 Nos. 19-2581 & 19-2741

motion to dismiss inquiry on its head; courts should ask whether the plaintiff’s interpretation of a statement is facially implausible, not whether the statement is “clearly” enough misleading. And it’s simply a stretch to say that a consumer’s reading of a statement is implausible as a matter of law just because fine print elsewhere on the label could clarify an am- biguity that a reasonable consumer might not have even no- ticed in the first place. Ultimately, a jury might conclude that the 100% statement is unlikely to deceive a reasonable consumer. Take the below example, which is representative of Defendants’ products (though, to be sure, they vary in content and layout). Coupled with other evidence, this label might lead a jury to conclude that a reasonable consumer would interpret the statement to assert only that whatever is in the canister is 100% grated, not 100% Parmesan and not 100% cheese:

But the point is, how a reasonable consumer would inter- pret this label is a question of fact, not a question of law, and the district court erred in deciding it on the pleadings. Ac- cordingly, I join Judge Hamilton’s fine opinion.