Case: 14-11648      Date Filed: 06/09/2015      Page: 1 of 22


                                                                  [DO NOT PUBLISH]



                 IN THE UNITED STATES COURT OF APPEALS

                           FOR THE ELEVENTH CIRCUIT
                             ________________________

                                    No. 14-11648
                              ________________________

                          D.C. Docket No. 0:13-cv-60768-JIC



ADAM KARHU,

                                                                      Plaintiff - Appellant,

                                           versus

VITAL PHARMACEUTICALS, INC.,
d.b.a. VPX Sports,

                                                                     Defendant - Appellee.

                              ________________________

                     Appeal from the United States District Court
                         for the Southern District of Florida
                           ________________________

                                      (June 9, 2015)

Before MARTIN and FAY, Circuit Judges, and GOLDBERG, * Judge.




*
 The Honorable Richard W. Goldberg, United States Court of International Trade Judge, sitting
by designation.
                 Case: 14-11648       Date Filed: 06/09/2015       Page: 2 of 22


GOLDBERG, Judge:

      Vital Pharmaceuticals, Inc. (“VPX”) markets a dietary supplement called

VPX Meltdown Fat Incinerator (“Meltdown”), which it advertises for fat loss.

Adam Karhu purchased the supplement in reliance on Meltdown’s advertising.

Karhu brought class-action suit, alleging that Meltdown’s advertising is false,

insofar as Meltdown does not aid fat loss.

      Karhu moved to certify class of nationwide Meltdown purchasers as well as

a subclass of New York purchasers. Certification is governed by Federal Rule of

Civil Procedure 23. Under Rule 23, certification is proper where the proposed

classes satisfy an implicit ascertainability requirement, the four requirements listed

in Rule 23(a), and the requirements listed in any of Rule 23(b)(1), (2), or (3).

Little v. T-Mobile USA, Inc., 

, 1304 (11th Cir. 2012). The district

court denied Karhu’s motion, holding that the proposed classes satisfied neither

Rule 23’s implicit ascertainability requirement, nor the requirements listed in either

Rule 23(b)(2) or (3).1 Karhu moved to alter or amend the order denying class

certification, which the district court also denied.

      Karhu appeals. He claims that the district court erred to hold that (1) neither

proposed class satisfied the ascertainability requirement, and (2) the New York

subclass failed to satisfy Rule 23(b)(3)’s requirements. We hold that the district


      1
          Karhu had not argued for certification pursuant to Rule 23(b)(1).
                                                 2
              Case: 14-11648     Date Filed: 06/09/2015    Page: 3 of 22


court’s ascertainability decision was proper. We therefore affirm without reaching

the district court’s Rule 23(b)(3) decision.

                                 I. BACKGROUND

   A. Legal Framework for the Ascertainability Requirement

      “The burden of establishing the[ requirements of certification under Rule 23]

is on the plaintiff who seeks to certify the suit as a class action.” Heaven v. Trust

Co. Bank, 

, 737 (11th Cir. 1997). Rule 23 implicitly requires that the

“proposed class is adequately defined and clearly ascertainable.” 

(internal quotation marks omitted).

      In the past, this court has stated that a class is not ascertainable unless the

class definition contains objective criteria that allow for class members to be

identified in an administratively feasible way. Bussey v. Macon Cnty. Greyhound

Park, Inc., 562 F. App’x. 782, 787 (11th Cir. 2014). Identifying class members is

administratively feasible when it is a “manageable process that does not require

much, if any, individual inquiry.” 

marks omitted).

   B. The District Court’s Order Denying Karhu’s Motion for Class Certification

      Invoking these rules, the district court denied Karhu’s motion for class

certification, holding that Karhu had failed to establish that his proposed classes

were ascertainable. Although Karhu’s class definitions contained objective

criteria, Karhu “ha[d] failed to propose a realistic method of identifying the


                                           3
             Case: 14-11648     Date Filed: 06/09/2015   Page: 4 of 22


individuals who purchased Meltdown.” Karhu had proposed that the court use

VPX’s “sales data,” but VPX sold primarily to “distributors and retailers,” such

that VPX’s records could not be used to determine “the identities of most [class]

members.”

      The court also considered, apparently of its own accord, whether allowing

class members to come forward and identify themselves through sales receipts or

affidavits would render the classes ascertainable. The court rejected the receipts-

based method on grounds that Meltdown’s low cost meant most class members

would not retain their proof of purchase.

      The district court also rejected the affidavit-based method. The court had

several concerns. If, on the one hand, “affidavits of Meltdown purchases [were

accepted] without verification,” VPX would be deprived “of its due process rights

to challenge the claims of each putative class member.” “On the other hand,

allowing VPX to contest each affidavit would require a series of mini-trials” to

determine class membership, which would not be administratively feasible.

Moreover, “[u]sing affidavits to determine class membership would also invite

fraudulent submissions and could dilute the recovery of genuine class members.”

   C. Karhu’s Motion to Alter or Amend the Order Denying Class Certification,
      and the District Court’s Order Denying Karhu’s Motion

      Karhu moved to alter or amend the order denying certification pursuant to

Rule 23(c)(1)(C). Karhu argued, inter alia, that new evidence showed that VPX
                                            4
                Case: 14-11648       Date Filed: 06/09/2015        Page: 5 of 22


sold Meltdown primarily to third-party retailers, such that class members could be

identified by subpoenaing records from retailers. According to Karhu, proposing

that class members could be identified using the records of third-party retailers was

sufficient to satisfy the ascertainability requirement.

       The district court denied Karhu’s motion for reconsideration. The court

reasoned that Karhu’s subpoena-based method was not predicated on new evidence

at all: Karhu had come up with the method by examining VPX’s sales data, which

had been available to Karhu well before he moved for class certification.

Therefore, the district court would not accept such a description as grounds for

reconsideration.

                                       II. DISCUSSION

       Karhu appeals, claiming, inter alia, that the district court abused its

discretion in the order denying class certification by holding that Karhu had failed

to establish that his proposed classes were ascertainable.2 We affirm the district

court’s ascertainability decision.

       As noted, the plaintiff seeking certification bears the burden of establishing

the requirements of Rule 23, including ascertainability. 

.

In order to establish ascertainability, the plaintiff must propose an administratively

       2
          A district court’s denial of class certification is reviewed for an abuse of discretion.

. Under the abuse of discretion standard, factual determinations are
reviewed for clear error, and legal determinations are reviewed de novo. Vega v. T-Mobile, 

, 1264–65 (11th Cir. 2009).

                                                5
              Case: 14-11648     Date Filed: 06/09/2015    Page: 6 of 22


feasible method by which class members can be identified. See Stalley v. ADS

Alliance Data Syst., Inc., 

, 679–80 (M.D. Fla. 2013) (denying

certification because “the Court ha[d] not been presented with reasonable methods

for ascertaining the identity of the [class members, that is] individuals who

answered [the defendant’s] collection calls”); Hill v. T-Mobile, USA, Inc., No.

2:09-cv-1827-VEM, 

, at *10–11 (N.D. Ala. May 16, 2011)

(holding ascertainability not established where plaintiffs had proposed “creating a

class list using T[-]Mobile’s databases,” because plaintiffs “ha[d] not addressed

how to effectively back out from such a list” the identities of persons not eligible

for class-action relief); see also Bussey, 562 F. App’x at 788 (revising class

definition to cover only loyalty-card-using gamblers who lost money during a

gaming session—rather than during a specific game—because plaintiffs “ha[d] not

provided any indication that they have, or even that they can obtain, data about

losses at the game level”); cf. Carrera v. Bayer Corp., 

, 306–07 (3d

Cir. 2013) (“A plaintiff may not merely propose a method of ascertaining a class

without any evidentiary support that the method will be successful.”).

      A plaintiff cannot establish ascertainability simply by asserting that class

members can be identified using the defendant’s records; the plaintiff must also

establish that the records are in fact useful for identification purposes, and that

identification will be administratively feasible. See 

–80


                                           6
              Case: 14-11648      Date Filed: 06/09/2015    Page: 7 of 22


(holding ascertainability not established where plaintiffs proposed that “members

of the class [of actual recipients of defendant’s calls] . . . be identified and notified

based on [the defendant’s] own records,” because the defendant’s records indicated

“merely the intended recipients” (internal quotation marks omitted)); Hill, 

, at *10–11; see also Bussey, 562 F. App’x at 788 (revising class

definition to encompass only persons who could, in fact, be identified using the

defendant’s records).

      Similarly, a plaintiff cannot satisfy the ascertainability requirement by

proposing that class members self-identify (such as through affidavits) without first

establishing that self-identification is administratively feasible and not otherwise

problematic. See Fisher v. Ciba Specialty Chems. Corp., 

, 301–02

(S.D. Ala. 2006) (rejecting “plaintiffs’ optimistic argument that prospective class

members could be counted on to self-select”); LaBauve v. Olin Corp., 

, 684 (S.D. Ala. 2005) (holding possibility of publication notice does not

establish ascertainability in part because “certain people may respond to

publication notice even though they were not [part of the class]”); Perez v.

Metabolife Int’l, Inc., 

, 269 (S.D. Fla. 2003) (holding

ascertainability not established when “the only evidence likely to be offered in

many instances will be the putative class member’s uncorroborated claim that he or

she used the product”). The potential problems with self-identification-based


                                            7
                Case: 14-11648        Date Filed: 06/09/2015        Page: 8 of 22


ascertainment are intertwined. On the one hand, allowing class members to self-

identify without affording defendants the opportunity to challenge class

membership “provide[s] inadequate procedural protection to . . . [d]efendant[s]”

and “implicate[s their] due process rights.” 

; see also

(citing Perez); cf. Marcus v. BMW of N. Am., LLC,

, 594 (3d Cir. 2012) (“Forcing BMW and Bridgestone to accept as

true absent persons’ declarations that they are members of the class, without

further indicia of reliability, would have serious due process implications.”). 3 On

the other hand, protecting defendants’ due-process rights by allowing them to

challenge each claimant’s class membership is administratively infeasible, because

it requires a “series of mini-trials just to evaluate the threshold issue of which

[persons] are class members.” 

; see also Perez, 218

       3
          Karhu argues that defendants have no due-process right against unverified self-
identification when total liability will be established at trial, and will not change depending on
the number of claims actually made. Because defendants’ total liability will not be affected, it
should not matter to them whether or not they are defrauded.
         This argument has no force. As the Carrera court explained, a defendant’s due-process
right against unverified self-identification is not only about total liability: It is also about
ensuring finality of judgment. 

(“If fraudulent or inaccurate claims
materially reduce true class members’ relief, these class members could argue the named
plaintiff did not adequately represent them . . . . When class members are not adequately
represented . . . they are not bound by the judgment.”). Nor is Allapatah Servs., Inc. v. Exxon
Corp., 

, 1258–59 (11th Cir. 2003), the only case Karhu cites in support of his
argument, to the contrary. In Allapatah, we held that defendants do have a due-process right to
contest individual class members’ claims during postverdict claim distribution, at least when
total liability has not already been established and the defendants’ defenses might affect such
liability. 

was not a class-certification (much less ascertainability) case, and did not
foreclose the possibility that defendants’ due-process right against unverified self-identification
might arise from some other concern besides total liability.


                                                 8
                Case: 14-11648        Date Filed: 06/09/2015         Page: 9 

(“[I]ndividualized mini-trials would be required even on the limited

issue of class membership.”). 4 A plaintiff proposing ascertainment via self-

identification, then, must establish how the self-identification method proposed

will avoid the potential problems just described.5

       In light of these standards, the district court’s ascertainability holding was

not an abuse of discretion. Karhu’s proposal to identify class members using

VPX’s “sales data” was incomplete, insofar as Karhu did not explain how the data

would aid class-member identification. Nor was any potential identification

procedure obvious: VPX’s sales data identified mostly third-party retailers, not

class members. Karhu did not explain to the court that it envisioned a three-step

identification process: (1) Use the sales data to identify third-party retailers, (2)

subpoena the retailers for their records, and (3) use those records to identify class

members. Therefore, the district court acted within its discretion when it rejected

Karhu’s proposal to identify class members via VPX’s “sales data.”



       4
          We do not address whether self-identification-based ascertainment might also implicate
the interests of absent class members in cases where total liability will be established at trial. See

(“It is unfair to absent class members if there is a significant likelihood
their recovery will be diluted by fraudulent or inaccurate claims.”).
       5
          A plaintiff might establish that self-identification-based ascertainment is
administratively feasible and otherwise unproblematic by proposing a case-specific and
demonstrably reliable method for screening each self-identification. See 

(rejecting affidavit screening model for lack of case-specificity and reliability, and
remanding to “afford [the plaintiff] the opportunity to submit a[n affidavit] screening model
specific to this case and prove how the model will be reliable and how it would allow [the
defendant] to challenge the affidavits.”).
                                                  9
              Case: 14-11648        Date Filed: 06/09/2015   Page: 10 of 22


      The district court likewise acted within its discretion when it rejected

identification via affidavit. Because Karhu had not himself proposed an affidavit-

based method, he necessarily had not established how the potential problems with

such a method would be avoided. Without a specific proposal as to how

identification via affidavit would successfully operate, the district court had no

basis to accept the method. We therefore uphold the district court’s

ascertainability holding in full.

      Karhu’s arguments that we construe the ascertainability requirement too

strictly do not convince. For example, Karhu is incorrect that a strict

ascertainability requirement conflicts with Klay v. Humana, Inc., 

,

1271–72 (11th Cir. 2004). In Klay, we held that the district court did not abuse its

discretion by finding that class-action resolution was superior under Rule 23(b)(3),

even though individualized issues of reliance, causation, and damages would have

to be resolved as to particular class members. 

individualized

issues implicated the manageability of the case—a consideration under Rule

23(b)(3)—we reasoned that manageability concerns “will rarely, if ever, be in

[themselves] sufficient to prevent certification of a class.” 

Karhu

argues that a strict ascertainability requirement violates the Klay principle that a

concern about case manageability should not stand in the way of certification.




                                             10
             Case: 14-11648       Date Filed: 06/09/2015   Page: 11 of 22


      Not so. Klay addressed manageability concerns that a court might face after

class members have already been identified—for example, concerns about whether

particular class members are entitled to relief in light of individualized reliance,

causation, and damages issues. 

Ascertainability, by contrast,

addresses whether class members can be identified at all, at least in any

administratively feasible (or manageable) way. Put differently, the manageability

concern at the heart of the ascertainability requirement is prior to, hence more

fundamental than, the manageability concern addressed in Klay. Klay therefore

presents no bar to our holding.

      Karhu is also incorrect that a strict ascertainability requirement will

eradicate small-dollar class-action claims. Karhu argues that small-dollar plaintiffs

will not be able to propose an administratively feasible method by which class

members can be identified. Karhu’s own briefing illustrates why his fear is

unfounded. According to Karhu, his counsel has before succeeded in proposing

administratively feasible identification methods. For example, in In re Scotts EZ

Seed Litig., No. 12-cv-4727, (S.D.N.Y. Jun. 15, 2012), plaintiffs established that

many purchasers of EZ Seed could be identified by subpoenaing third-party

retailor Wal-Mart for its customer records (and in fact did so subpoena Wal-Mart).

The district court later held that Karhu had satisfied the ascertainability




                                           11
               Case: 14-11648        Date Filed: 06/09/2015       Page: 12 of 22


requirement. In re Scotts EZ Seed Litig., No. 12-cv-4727, 

, at

*5–6 (S.D.N.Y. Jan. 26, 2015).

       Relatedly, Karhu might have satisfied the ascertainability requirement in this

very case. In his motion to alter or amend the order denying class certification,

Karhu explained that VPX sold Meltdown primarily to third-party retailers, and

proposed identifying class members by subpoenaing the retailers for their records.

The district court took no issue with the abstract principle that a plaintiff could

satisfy the ascertainability requirement by proposing a subpoena-based method for

identifying class members. Rather, the district court held only that Karhu should

have proposed the method in his class-certification papers, instead of only upon

moving to alter or amend. Had Karhu done so, he might well have satisfied the

ascertainability requirement before the district court.6

       In sum, a plaintiff establishes Rule 23’s implicit ascertainability requirement

by proposing an administratively feasible method by which class members can be

identified. In this case, Karhu’s bare proposal that the district court ascertain class



       6
         We do not address the question of what, precisely, Karhu would need to produce in
order to establish that the records of third-party retailers could be used to identify class members
in an administratively feasible manner. In this regard, we express no opinion as to whether the
Carrera court set an appropriate bar for subpoena- or third-party-retailer based ascertainment.
See 

–09 (“Carrera argues he will be able to show class membership
using retailer[s’] records of sales made with loyalty cards, e.g., CVS ExtraCare cards, and
records of online sales. . . . But there is no evidence that a single purchaser of WeightSmart could
be identified using records of . . . online sales. There is no evidence that retailers even have
records for the relevant period.”)
                                                12
               Case: 14-11648   Date Filed: 06/09/2015   Page: 13 of 22


members through VPX’s “sales data” was insufficient to satisfy the ascertainability

requirement.

                                   III. Conclusion

      Because we uphold the district court’s ascertainability decision, we affirm

without reaching the district court’s Rule 23(b)(3) decision.

      AFFIRMED.




                                         13
             Case: 14-11648     Date Filed: 06/09/2015    Page: 14 of 22


MARTIN, Circuit Judge, concurring:

      The vehicle of the class action was intended to “vindicat[e] . . . the rights of

groups of people who individually would be without effective strength to bring

their opponents into court at all.” Amchem Prods., Inc. v. Windsor, 

,

617, 

, 2246 (1997) (citations omitted). Today’s majority denies

relief based on the court-created doctrine of ascertainability, which, as a principle,

could erode this purpose. Some courts have held that a prospective class of

consumers of a small-dollar product is not ascertainable if the only way they can

be identified is through self-identification. Though the facts of this case present

this court with no opportunity to join those, I write separately to address the

problems with such a holding. Specifically, self-identification can and should be a

sufficient means of ascertaining a class, particularly for a class of consumers of a

cheap and unique product like the one at issue here. I therefore reject the District

Court’s reasoning and its potential implications. However, because I agree with

the majority that Mr. Karhu failed to sufficiently make a self-identification

argument at the class-certification stage, I concur in the judgment.

                                          I.

      Vital Pharmaceuticals, Inc. (VPX) is a Florida corporation that

manufactures, advertises, and sells a dietary supplement, modestly named “VPX

Meltdown Fat Incinerator” (Meltdown). VPX advertises, also modestly, that


                                          14
             Case: 14-11648      Date Filed: 06/09/2015    Page: 15 of 22


Meltdown “burns fat for 6+ hours” by causing a “29% thermogenic increase” and a

“56% increase in fat utilization.” Adam Karhu, a New York resident who

purchased Meltdown, sued VPX, claiming that Meltdown is not effective for this

advertised purpose.

      Mr. Karhu filed his suit as a class action. He moved pursuant to Federal

Rule of Civil Procedure 23(b)(3) for certification of a class defined as “all persons

in the United States who purchased Meltdown from April 4, 2008 to date” as well

as a subclass of “all Class members who purchased the product in New York.” He

asserted claims on behalf of the nationwide class for (1) violation of the

Magnuson-Moss Warranty Act, 15 U.S.C. §§ 2301–12; (2) breach of express

warranty; (3) unjust enrichment; and (4) violation of the Florida Deceptive and

Unfair Trade Practices Act, Fla. Stat. §§ 501.201–13. He also asserted a claim on

behalf of only the New York subclass for violations of New York General

Business Law § 349.

      The District Court denied Mr. Karhu’s request for class certification

primarily because it concluded that the members of Mr. Karhu’s proposed classes

were not ascertainable. First, although Mr. Karhu suggested that VPX’s “sales

data would allow the Court to identify members,” the District Court found that

“VPX . . . makes the bulk of its sales to distributors and retailers and sells directly

to consumers relatively infrequently.” Thus, VPX would not have a record of most


                                           15
              Case: 14-11648     Date Filed: 06/09/2015    Page: 16 of 22


class members. Further, because a bottle of Meltdown is “a relatively small

purchase, . . . purchasers are less likely to retain receipts or other records” to show

they have purchased Meltdown. Finally, the Court refused to “trust individuals to

identify themselves as class members through the submission of affidavits.” Doing

so would “deprive VPX of its due process rights to challenge the claims of each

putative class member.” “On the other hand, allowing VPX to contest each

affidavit would require a series of mini-trials and defeat the purpose of class-action

treatment.”

                                          II.

      A plaintiff seeking class certification must demonstrate that his claim meets

the express requirements of Rule 23(a): (1) numerosity; (2) commonality; (3)

typicality; and (4) adequacy. See Fed. R. Civ. P. 23(a)(1)–(4). However, courts

have also read Rule 23 to contain an implicit, unwritten requirement: that a

proposed class be “adequately defined and clearly ascertainable.” Little v. T-

Mobile USA, Inc., 

, 1304 (11th Cir. 2012) (quotation omitted).

      Historically, courts analyzing ascertainability have required something quite

narrow. “Ascertainability has traditionally been defined as the existence of a class

whose members can be identified by reference to objective criteria in the class

definition.” Daniel Luks, Ascertainability in the Third Circuit: Name That Class

Member, 82 Fordham L. Rev. 2359, 2369 (2014). The leading class action treatise


                                           16
             Case: 14-11648      Date Filed: 06/09/2015    Page: 17 of 22


similarly notes that “courts essentially focus on the question of whether the class

can be ascertained by objective criteria.” Newberg on Class Actions § 3.3 (5th

ed.); see also McBean v. City of New York, 

, 133 (S.D.N.Y. 2009)

(“A class is ascertainable when defined by objective criteria that are

administratively feasible, without a subjective determination.”). What is more, that

treatise cautions that a “court need not know the identity of each class member

before certification; ascertainability requires only that the court be able to identify

class members at some stage of the proceeding.” Newberg on Class Actions § 3.3

(5th ed.).

       Our Court’s approval of the class in Fitzpatrick v. General Mills, Inc., 

(11th Cir. 2011), exemplifies this approach. There, we approved of a

district court’s certification of a class of all purchasers of YoPlus, a probiotic

yogurt, in the State of Florida. 

(noting that the district court order

“conducted a detailed analysis,” was a “scholarly work reflecting careful attention

to the requirements of” Rule 23, and “is sound and in accord with federal and state

law”). The district court in Fitzpatrick approved of the class despite the “likely

difficulties” of “identifying those consumers that bought Yo-Plus, and of that

group, who paid a premium for it,” and “calculating the appropriate compensation

for each plaintiff.” Fitzpatrick v. General Mills, Inc., 

, 702 (S.D.

Fla. Jan. 11, 2010). Yogurt consumers, after all, are unlikely to retain receipts


                                           17
               Case: 14-11648       Date Filed: 06/09/2015       Page: 18 of 22


proving their purchases. But the District Court insisted that “this difficulty [was]

not in itself a sufficient basis to prevent certification of the class.” 

omitted). We easily allowed certification of that class without questioning

its ascertainability.

                                               III.

       The record here indicates that Mr. Karhu could have made a good case for

the ascertainably of his proposed class based on consumer affidavits. 1 Many

courts have grappled with whether affidavits can be a reliable way to identify class

membership. I see these courts looking to at least two factors. First, courts look to

the value of the product being challenged in the class action to determine the

likelihood that fraudulent claims will be filed. “A simple statement or affidavit

may be sufficient where claims are small.” Newberg on Class Actions § 18:54 (4th

ed.). Thus, while courts have certified classes of purchasers of low-value items,

such as supplements or bottled beverages, based solely on consumer self-

identification, see, e.g., McCrary v. Elations Co., LLC, No. EDCV 13-00242 JGB,

, at *7–9 (C.D. Cal. Jan. 13, 2014) (allowing “class members to

self-identify” that they purchased “an over-the-counter supplement sold in retailers

throughout [a] state”), they have been more reluctant to certify classes when the


       1
          I agree with the majority that Mr. Karhu forfeited his argument that he could identify
purchasers of Meltdown through third-party sales records, since he raised this argument for the
first time in his motion for reconsideration.
                                                18
             Case: 14-11648     Date Filed: 06/09/2015    Page: 19 of 22


per-class-member claim is larger, see, e.g., In re Hulu Privacy Litig., No. C 11-

03764 LB, 

, at *16 (N.D. Cal. June 17, 2014) (finding no

ascertainability because “at $2,500 per class member, [the claims] are not small”).

      Second, courts have evaluated the “the likelihood of a potential class

member being able to accurately identify themselves as a purchaser of the

allegedly deceptive product.” Randolph v. J.M. Smucker Co., 

, 689

(S.D. Fla. 2014). Where a challenged product is similar to other unchallenged

products on the market, consumers may find it hard to know whether they

purchased the challenged product. For instance, in Randolph, the Southern District

of Florida declined to certify a class of purchasers of certain Crisco oils. The court

noted that there were at least nine different Crisco oils on the market, only four of

which were the subject of that case. 

Beyond that, even among those

four oils, the allegedly deceptive label being challenged “was not placed on all four

oils uniformly throughout the class period.” 

concluded that “the

likelihood that an individual would recall not only which specific kind of oil, but

also, when that oil was purchased, complicates identification of the putative class.”

Id.; see also Jones v. ConAgra Foods, Inc., No. C 12-01633 CRB, 

, at *10 (N.D. Cal. June 13, 2014) (noting a “subjective memory problem”

because “there were literally dozens of varieties with different can sizes,

ingredients, and labeling over time and some Hunt’s cans included the challenged


                                          19
               Case: 14-11648       Date Filed: 06/09/2015       Page: 20 of 22


language, while others included no such language at all” (quotation marks

omitted)).

       I would combine the reasoning of these courts. Looking to the two factors—

(1) the value of each class member’s claim, and (2) the likelihood that potential

class members could accurately identify themselves—I conclude that Mr. Karhu

likely could have shown ascertainability. First, a bottle of Meltdown costs around

$23, a small amount not likely to invite fraudulent claims. Second, Mr. Karhu

notes that “Meltdown is a unique product in name, function, and appearance.”

This is in contrast to Randolph, where there were a number of versions of Crisco

on the market, and only some were being challenged. This record includes no

evidence that consumers would be unable to recall whether they are class

members. A class composed of purchasers of Meltdown for a certain period of

time could therefore be ascertainable. 2


       2
          The cases the majority cites rule on facts different from those here. In Fisher v. Ciba
Specialty Chemicals Corp., 

(S.D. Ala. 2006), the court held that a class of
landowners was not adequately defined because the class definition included a requirement that
each parcel of land not be “income-producing.” 

The court rejected the argument that
“prospective class members could be counted on to self-select” because the question of whether
a parcel of land is “income-producing” cannot be reduced to a simple yes-or-no question. 

“Depositions would need to be taken, documents would have to be produced, and argument
would need to be heard, . . . to assess whether [any lot] was or was not ‘income-producing.’” 

LaBauve v. Olin Corp., 

(S.D. Ala. 2005), the court declined to
certify a certain class of fishermen because it thought that “people may respond to publication
notice even though they were not fishing in the particular area of concern during the particular
temporal interval of concern” since the geographic and temporal scope of the class of fishermen
was “so amorphous.” 

Finally, in Perez v. Metabolife International, Inc., 

(S.D. Fla. 2003), the court was wary of certifying a class of purchasers of an over-
the-counter weight loss supplement because the supplement was “only one of several producing
                                               20
               Case: 14-11648        Date Filed: 06/09/2015       Page: 21 of 22


                                               IV.

       Unfortunately for the putative class, Mr. Karhu failed to argue this point in

his class-certification motion. As the majority notes, Mr. Karhu, did “not

established how the usual problems with [affidavits] would be avoided.” Panel

Op. 10. Mr. Karhu, it says, failed to offer a “specific proposal as to how

identification via affidavit would successfully operate.” 

the District

Court apparently considered the self-identification argument “of its own accord”

below. 

As I’ve said, I believe that self-identification would probably be a

sufficient means of ascertaining a class of purchasers of a product like Meltdown.

In any event, I read today’s majority opinion narrowly. Mr. Karhu simply did not

adequately argue his class was ascertainable before the District Court. Class

representatives in future cases may more clearly explain to district courts how

affidavits will reliably show class membership based on the two factors I noted

above, and I expect that district courts will closely consider those arguments.

       To hold otherwise—rejecting affidavits as a legitimate means of class

identification in every case—would make it considerably more difficult for

consumers to bring class-action claims on small-dollar products where consumers

and companies are unlikely to keep or retain records of purchases. These include



containing ephedra, at least two others of which have very similar names.” 

Here, the
class is clearly defined, and there was little risk of confusion since there was no similar product
on the market.
                                                21
             Case: 14-11648      Date Filed: 06/09/2015    Page: 22 of 22


most low-cost products typically sold in corner stores or vending machines—

products like chewing gum, bottled soft drinks, or cigarettes—all of which are

routinely bought with cash. But claims like these are precisely the ones that the

mechanism of the class-action device was designed to foster. See 

(“[T]he class action device, at its very core, is designed for cases like this

where a large number of consumers have been defrauded but no one consumer has

suffered an injury sufficiently large as to justify bringing an individual lawsuit.”).

I would like to see our courts continue to clarify the ascertainability doctrine so as

not to eradicate the small-dollar consumer class action.

                                          VI.

      The record here has led me to conclude that although Mr. Karhu failed to

properly argue that affidavits were sufficient to show ascertainability in his class-

certification motions to the District Court, he had a strong case to make. When

timely presented, I would hold that affidavits are a sufficient means of

identification for purchasers of a cheap, unique product like Meltdown. I concur in

the judgment.




                                           22