In the
United States Court of Appeals
For the Seventh Circuit
____________________
No. 18‐1010
ARCHIE BEATON,
Plaintiff‐Appellee,
v.
SPEEDYPC SOFTWARE, a British Columbia Company,
Defendant‐Appellant.
____________________
Appeal from the United States District Court for the
Northern District of Illinois, Eastern Division.
No. 13 C 8389 — Andrea R. Wood, Judge.
____________________
ARGUED MAY 30, 2018 — DECIDED OCTOBER 31, 2018
____________________
Before WOOD, Chief Judge, and SYKES and HAMILTON,
Circuit Judges.
WOOD, Chief Judge. When Archie Beaton’s laptop started
misbehaving, he looked for an at‐home fix. An internet search
turned up a product from SpeedyPC Software (“Speedy”)
that offered both a diagnosis and a cure. Beaton took ad‐
vantage of Speedy’s free trial, which warned that his device
was in bad shape and encouraged him to purchase its soft‐
2 No. 18‐1010
ware solution: SpeedyPC Pro. He did. But he was disap‐
pointed with the outcome: despite Speedy’s promises, the
software failed to improve his laptop’s performance.
Beaton became convinced that he was the victim of a scam.
He filed a consumer class action against Speedy, raising both
contract and tort theories. The district court certified a nation‐
wide class and an Illinois subclass of software purchasers.
Hoping to dodge the consumer class action, Speedy turned to
this court for relief. See FED. R. CIV. P. 23(f). Because we find
no abuse of discretion in the district court’s certification or‐
ders, we affirm.
I
The ad for SpeedyPC Pro that Beaton found in August
2012 promised that Speedy’s software would fix common
problems affecting computer speed and performance and un‐
leash the device’s “true potential.” It also offered a free scan
to detect any problems. Beaton decided to give it a try, and so
he downloaded and ran the free trial. After assessing the lap‐
top’s health across five modules, the program told Beaton that
his computer was in critical condition as a result of hundreds
of serious errors.
The free trial prompted Beaton to buy the licensed version
of the software, which (he was promised) would fix the iden‐
tified problems. Beaton was sold. Using his personal busi‐
ness’s credit card, he purchased SpeedyPC Pro and ran it on
his laptop. It began by scanning his device, just as the free trial
had done. The program then told Beaton to click on “Fix All.”
Beaton dutifully did so. Yet nothing happened. Beaton ran the
software a few more times, to no avail.
No. 18‐1010 3
Feeling ripped off, and suspecting that his experience was
not unique, Beaton sued Speedy in 2013 on behalf of a class of
consumers defined as “All individuals and entities in the
United States who have purchased SpeedyPC Pro.” Despite
Speedy’s lofty pledges, Beaton claimed, the software failed to
perform as advertised. Instead, it indiscriminately and mis‐
leadingly warned all users that their devices were in critical
condition, scared them into buying SpeedyPC Pro, and then
ran a functionally worthless “fix.” The district court had juris‐
diction over this putative class action under the Class Action
Fairness Act, 28 U.S.C. § 1332(d)(2).
Speedy twice tried, and twice failed, to get the lawsuit
thrown out. The district court first rejected its effort to have
the complaint dismissed for failure to state a claim on which
relief could be granted. Speedy then tried a motion to dismiss
on forum non conveniens grounds, based on the fact that the
software’s End User License Agreement (“the Agreement”)
contained a choice‐of‐law provision selecting the law of Brit‐
ish Columbia (Canada) to govern any claims arising from it.
The district court, however, decided to retain the case without
definitively resolving the choice‐of‐law issue at that juncture.
Four years after the suit was filed, Beaton moved to certify
a class and subclass of software purchasers. Beaton’s pro‐
posed class definition was narrower than the one in his com‐
plaint. It included “[a]ll individuals living in the United States
who downloaded a free trial of SpeedyPC Pro and thereafter
purchased the full version between October 28, 2011 and No‐
vember 21, 2014.” He also proposed a subclass of class mem‐
bers “who reside in Illinois” and several other states.
The district court certified Beaton’s class claims for
breaches of the implied warranties of fitness for a particular
4 No. 18‐1010
purpose and merchantability. On behalf of a subclass consist‐
ing only of Illinois residents, the court certified claims for
fraudulent misrepresentation under the Illinois Consumer
Fraud and Deceptive Business Practices Act (ICFA). It rejected
the proposed subclass insofar as it included residents from
other states, because Beaton failed to identify the relevant con‐
sumer‐protection laws of those states.
The court had the benefit of dueling expert testimony be‐
fore it at the time it made these certification decisions. Bea‐
ton’s expert, Craig Snead, described how the free trial oper‐
ated across devices. Speedy’s expert, Monty Myers, disputed
Snead’s account. Although the court had not yet issued its rul‐
ing on the parties’ cross‐motions to exclude the testimony of
each other’s expert, it ultimately denied both motions (with
minor exceptions) roughly two months later. See FED. R. EVID.
702. In that order, the court noted that it had “considered the
challenged expert testimony for purposes of class certification
only to the extent consistent with the rulings stated.”
At that point, Speedy filed and we granted a petition for
interlocutory appeal of the class certification decisions. See
FED. R. CIV. P. 23(f). We note that Speedy’s petition may have
been untimely, but Beaton chose not to press the issue. The
time limit for an appeal under Rule 23(f) is not statutory, and
so a failure to abide by it does not affect our jurisdiction. See
Bowles v. Russell, 551 U.S. 205, 210–11 (2007); McReynolds v.
Merrill Lynch, Pierce, Fenner & Smith, Inc., 672 F.3d 482, 485
(7th Cir. 2012) (holding that Rule 23(f)’s 14‐day limitations pe‐
riod is not jurisdictional), abrogated on other grounds by Phillips
v. Sheriff of Cook Cnty., 828 F.3d 541 (7th Cir. 2016).
No. 18‐1010 5
II
Before we reach the heart of this appeal—the district
court’s Rule 23 decisions—we address Speedy’s more sub‐
stantial preliminary objections.
A
Speedy complains that the class definitions and legal the‐
ories covered by the court’s certification orders impermissibly
differ from those outlined in the original complaint. Speedy
first attacks the narrowing of the class from everyone in the
United States who had purchased SpeedyPC Pro, to individ‐
ual persons (not entities) who downloaded the free trial and
purchased the licensed software over roughly a three‐year pe‐
riod. This is nothing like what we faced in Supreme Auto
Transport, LLC v. Arcelor Mittal USA, Inc., 902 F.3d 735, 741
(7th Cir. 2018), where the later proposed class greatly ex‐
panded the scope of the litigation beyond what the defend‐
ants could have imagined. We see no reason here why Speedy
is prejudiced by the narrower certified definition. Speedy
complains that it would have conducted discovery differently
had it known about the narrowed class. See Chessie Logistics
Co. v. Krinos Holdings, Inc., 867 F.3d 852, 859 (7th Cir. 2017).
But it has not told us, either in its briefs or at oral argument,
what exactly would have changed. Speedy’s position is fur‐
ther weakened by the fact that the district court allowed ad‐
ditional merits discovery following its certification decision.
District courts may amend class definitions either on motion
or on their own initiative. See FED. R. CIV. P. 23(c)(1)(C); Chap‐
man v. First Index, Inc., 796 F.3d 783, 785 (7th Cir. 2015); Abbott
v. Lockheed Martin Corp., 725 F.3d 803, 807 (7th Cir. 2013). We
6 No. 18‐1010
are satisfied that the court reasonably exercised its discretion
in adopting its class definition.
We similarly find no reversible error in the district court’s
decision to certify Beaton’s two implied warranty claims. It is
immaterial that these legal theories were not spelled out in the
initial complaint. See Chessie Logistics Co., 867 F.3d at 860. As
the Supreme Court and this court constantly remind litigants,
plaintiffs do not need to plead legal theories. Johnson v. City of
Shelby, 135 S. Ct. 346, 346–47 (2014) (per curiam); BRC Rubber
& Plastics, Inc. v. Cont’l Carbon Co., 900 F.3d 529, 540–41
(7th Cir. 2018); King v. Kramer, 763 F.3d 635, 642 (7th Cir. 2014).
Rule 8 requires only that a complaint must set forth plausible
facts that, if true, would support a claim for relief. See Ashcroft
v. Iqbal, 556 U.S. 662, 678 (2009); Bell Atlantic Corp. v. Twombly,
550 U.S. 544, 570 (2007); FED. R. CIV. P. 8(a)(2). Even where a
plaintiff initially asserts particular theories of recovery, unless
the change unfairly harms the defendant she is allowed to
switch course and pursue other avenues of relief as litigation
progresses. Chessie Logistics Co., 867 F.3d at 859; Whitaker v.
Milwaukee Cnty., 772 F.3d 803, 808 & n.18 (7th Cir. 2014). Here,
the court’s certification of the implied warranty claims was
permissible as long as Beaton’s allegations were plausible and
Speedy had fair notice of what this suit was about. See Run‐
nion ex rel. Runnion v. Girl Scouts of Greater Chi. & Nw. Ind.,
786 F.3d 510, 517 (7th Cir. 2015). We note as well that applica‐
ble law is no longer in dispute, as the parties now agree that
the implied warranty claims derive from the Agreement,
which chooses the law of British Columbia.
Beaton’s complaint describes Speedy as a company that
sells software products. He alleges that it marketed SpeedyPC
No. 18‐1010 7
Pro in the hope of persuading consumers to purchase the soft‐
ware to fix their computers. And he asserts that customers re‐
lied on the company’s expertise and representations that the
software would improve their devices. For present purposes,
this is enough to provide fair notice that he intends to pursue
warranty claims under the law of British Columbia. See
R.S.B.C. 1996, ch. 410, § 18(a)–(b). It is hard to imagine how
Speedy suffered any “unfair surprise,” given that the “legal
basis for liability is based on the same allegations” about the
sale of worthless software. Whitaker, 772 F.3d at 809 & n.19.
Though Speedy insists that it is worse off because it cannot
move to dismiss on the ground that the Agreement expressly
disclaimed these implied warranties, there is no final judg‐
ment in this case. Nothing prevents Speedy from pursuing
this point on remand.
B
Next, we briefly consider Speedy’s assertion that Beaton is
judicially estopped from seeking relief under the law of Brit‐
ish Columbia because initially he argued for Illinois law. Eq‐
uitable estoppel requires that: (1) the party’s later position is
clearly inconsistent with her earlier one; (2) the party success‐
fully persuaded the court to adopt her first position; and
(3) the party would be unfairly advantaged if not estopped.
Janusz v. City of Chi., 832 F.3d 770, 776 (7th Cir. 2016).
Speedy forfeited its estoppel argument by not raising it be‐
fore the district court. 1st Source Bank v. Neto, 861 F.3d 607,
611–12 (7th Cir. 2017). It merely acknowledged that Beaton
changed his position on whether British Columbia or Illinois
law controlled his contract claims.
8 No. 18‐1010
Even on the merits, Speedy’s estoppel theory falls short. It
is true that Beaton flip‐flopped his position on the source of
his implied warranty claims, and so the first criterion for es‐
toppel may be met. In his opposition to the motion to dismiss
for forum non conveniens, Beaton argued that “[n]one of [his]
claims are based upon [the Agreement].” But by the time he
sought class certification, he sang a different tune, conceding
that the implied warranty “claims derive from the End User
License Agreement.” Still, the other two factors necessary for
estoppel are missing. Beaton may have defeated Speedy’s mo‐
tion to dismiss for forum non conveniens, but he did not per‐
suade the district court that Illinois law controlled. The court
thought that British Columbia law may not apply to Beaton’s
contract claims because they “have little or nothing to do with
the terms of the [Agreement].” But ultimately the court found
that this question did not matter for class certification and so
could safely be postponed. And in any event, we cannot see
how Beaton could derive an unfair advantage by agreeing to
apply the substantive law that Speedy wanted all along.
C
Speedy also contends that the district court lacks personal
jurisdiction over the claims of class members from states other
than Illinois. Its argument relies on the Supreme Court’s deci‐
sion in Bristol‐Myers Squibb Co. v. Superior Court of Cal., San
Francisco Cnty., 137 S. Ct. 1773 (2017). In that mass‐tort action,
there was no connection between the forum and the specific
claims at issue. Under those circumstances, the Supreme
Court held that a state court lacks specific jurisdiction over
No. 18‐1010 9
non‐resident plaintiffs’ claims against non‐resident defend‐
ants. Id. at 1781–82. Speedy seems to be asking us to extend
Bristol‐Myers Squibb to nationwide class actions.
While briefing the issue now before us—class
certification—in the district court, neither party raised
personal jurisdiction. Thus, we have no need to opine on this
question, because it does not bear directly on our
determination. See Abelesz v. OTP Bank, 692 F.3d 638, 652–53
(7th Cir. 2012) (a court’s personal jurisdiction and class
certification decisions were “only tangentially related” and so
the former could not be evaluated on a Rule 23(f) appeal
(quoting Poulos Caesars World, Inc., 379 F.3d 654, 671–72 (9th
Cir. 2004))). On remand, Speedy is free to explain if and how
it preserved this point and how Bristol‐Myers Squibb applies in
these circumstances. For his part, Beaton will be free to
contend that Speedy waived this defense through its conduct.
See, e.g., H‐D Mich., LLC v. Hellenic Duty Free Shops S.A.,
694 F.3d 827, 848 (7th Cir. 2012). On a Rule 23(f) appeal, it is
not for us to take the first bite of this apple.
III
Now we turn to the main event: the district court’s deci‐
sion to certify the nationwide class and the Illinois subclass.
To certify a class under Federal Rule of Civil Procedure 23, a
district court must rigorously analyze whether the plaintiff
satisfies the rule’s requirements. Blow v. Bijora, Inc., 855 F.3d
793, 806 (7th Cir. 2017). Rule 23(a) sets forth four universal re‐
quirements for class actions: “numerosity, typicality, com‐
monality, and adequacy of representation.” Messner v.
Northshore Univ. HealthSys., 669 F.3d 802, 811 (7th Cir. 2012).
Rule 23(b) then identifies particular types of classes, which
have different criteria. Where, as here, certification is sought
10 No. 18‐1010
under Rule 23(b)(3), common questions of law or fact must
predominate over individual inquiries, and class treatment
must be the superior method of resolving the controversy. Id.
In evaluating these factors, the court must go beyond the
pleadings and, to the extent necessary, take evidence on dis‐
puted issues that are material to certification. Bell v. PNC Bank,
Nat’l Ass’n, 800 F.3d 360, 377 (7th Cir. 2015); Szabo v. Bridgeport
Machs., Inc., 249 F.3d 672, 675–76 (7th Cir. 2001). At this early
stage in the litigation, the merits are not on the table. Abbott,
725 F.3d at 810 (describing class definition as a “tool of case
management”); Messner, 669 F.3d at 811 (class certification
should not be turned into a “dress rehearsal for the trial on
the merits”). Beaton bears the burden of showing that each
requirement is met by a preponderance of the evidence.
Steimel v. Wernert, 823 F.3d 902, 917 (7th Cir. 2016).
We review the district court’s class certification orders def‐
erentially, leaving considerable room for the exercise of judg‐
ment unless the factual determinations are clearly erroneous
or there are errors of law. Reliable Money Order, Inc. v.
McKnight Sales Co., Inc., 704 F.3d 489, 498 (7th Cir. 2013).
A
Speedy complains generically that the district court failed
to give its evidence adequate attention. We see no basis for
that accusation. The court referred to Beaton’s pleadings in
providing the case’s background, and then it considered evi‐
dence submitted by Beaton and Speedy. A district court may
abuse its discretion by omitting key factual and legal analysis.
See Priddy v. Health Care Serv. Corp., 870 F.3d 657, 661 (7th Cir.
2017). But it has no obligation to describe every part of the
record.
No. 18‐1010 11
Speedy also specifically challenges the district court’s
findings on commonality, typicality, and adequacy of repre‐
sentation for purposes of Rule 23(a). (It concedes that numer‐
osity is not at issue.) We consider each of these in turn.
B
To satisfy the commonality requirement found in
Rule 23(a)(2), there needs to be one or more common ques‐
tions of law or fact that are capable of class‐wide resolution
and are central to the claims’ validity. Bell, 800 F.3d at 374. The
district court identified several such issues:
Can the customers avail themselves of any
implied warranties, or is the Agreement’s
disclaimer valid?
What functions did the marketing materials
represent that the software would perform?
Did the software perform those functions?
Speedy takes exception to some of these questions, but most
are amenable to class‐wide resolution. See Nikka Traders Inc.
v. Gizella Pastry Ltd. (2012), D.L.R. 4th 120, para. 65 (Can. B.C.
Sup. Ct.) (describing the elements of claim for the implied
warranty for fitness for a particular purpose); Dream Carpets
Ltd. v. Sandhedrai, [2009] B.C.W.L.D 5070, para. 68 (Can. B.C.
Prov. Ct.) (elements for implied warranty of merchantability);
Dubey v. Pub. Storage, Inc., 395 Ill. App. 3d 342, 353 (2009)
(same for ICFA). And we can see additional common ques‐
tions, including whether Speedy typically deals in goods re‐
lated to this software and whether a reasonable consumer
12 No. 18‐1010
would be deceived by the advertisements’ representations.
Commonality is easily satisfied.
C
Second, we consider typicality. See Rule 23(a)(3); Oshana
v. Coca‐Cola Co., 472 F.3d 506, 514 (7th Cir. 2006). This requires
us to evaluate whether Beaton’s claims arise from the same
events or course of conduct that gives rise to the putative class
members’ claims. The individual claims may feature some
factual variations as long as they “have the same essential
characteristics.” Id. (citation omitted).
The district court thought this requirement satisfied be‐
cause Beaton “appears to have seen the same representations
as the other users of Speedy’s free software, and the software
appears to operate in the same way on each computer.” Un‐
like Speedy, we do not take exception with the court’s use of
the word “appears” to describe the match between Beaton’s
claim and that of the other class members. This semantic
choice suggests only that the court’s determinations are pre‐
liminary, as they should be. See Messner, 669 F.3d at 811.
On the merits, neither of the court’s findings reflects an
abuse of discretion. We begin with the finding that Beaton
saw the same representations as other users. Speedy empha‐
sizes that some customers bought the software through third‐
party platforms, which could advertise as they saw fit. Yet the
advertisements in the record, drawn from various sites, fea‐
ture almost identical language. The class members were thus
exposed to the same message (and promises) from Speedy.
Next, we turn to the court’s determination that the free
trial operated the same way across devices. Based on a review
No. 18‐1010 13
of the free trial’s source code, Beaton’s expert, Snead, con‐
cluded that the software was programmed to operate uni‐
formly on all PCs, independent of any differences among in‐
dividual devices. He asserted that the software universally re‐
ported “problems” and “errors,” mislabeled innocuous and
routine features, and issued a low performance rating before
any scan had begun. In his view, the scan failed to account for
factors that do influence a device’s performance, and it incor‐
porated factors that have no impact. Speedy’s senior director
of technical operations confirmed that the scan identified as
problems characteristics that might not affect performance.
Speedy asks us to reject this evidence because Snead ex‐
amined only the source code for the free trial’s scanning por‐
tion, as opposed to the scanning or repair portions of the li‐
censed software. It is not clear how similar the two scanning
programs are, but that does not matter for our purposes. The
district court was entitled to credit the evidence indicating
that the free trial scan software did not differentiate between
devices before declaring them to be in “critical condition.”
This is sufficient to show that Beaton’s claims are typical. He
focuses on Speedy’s uniform (alleged) misrepresentation of
computer health to induce users to buy its product. Though
Speedy issued 19 different versions of the software during the
class period, Snead opined that “the primary features and
functionality remained consistent” across versions. Speedy’s
expert, Myers, disagreed with Snead’s conclusions, and the
company pointed to positive survey responses and third‐
party reviews to argue that Beaton’s experience was atypical.
But that just indicates that there are merits issues to be re‐
solved. For class certification purposes, the district court
needed only to find by a preponderance of the evidence that
the software scanned Beaton’s device in the same way as it
14 No. 18‐1010
scanned other class members’ computers. We see no reason to
reject its conclusion.
But, Speedy argues, the district court did not say out loud
that it weighed both expert reports and found Snead’s conclu‐
sions more persuasive. In fact, the court did not mention My‐
ers’s report at all. Speedy sees this as a glaring omission be‐
cause the court had yet to rule on the Rule 702 cross‐motions.
It points out that a district court should not certify a class, and
thereby raise the stakes of the litigation, based on faulty opin‐
ion evidence. Instead, it “must conclusively rule on any chal‐
lenge to the expert’s qualifications or submissions prior to rul‐
ing on a class certification motion,” if the “expert’s report or
testimony is critical to class certification.” Am. Honda Motor
Co., Inc. v. Allen, 600 F.3d 813, 814–15 (7th Cir. 2010). Speedy
concludes that the court erred by not doing so.
If this was error (a point we need not resolve), it was harm‐
less. See Messner, 669 F.3d at 814. In its Rule 702 ruling, the
district court made clear that it had considered only the expert
testimony it later deemed admissible. Speedy gives us no rea‐
son to doubt the district court’s assurance. And it is also worth
recalling that the district court permitted additional merits
discovery after its certification decision. Had Speedy wished
to pursue the expert qualifications issue further, it could have
done so. We thus find no abuse of discretion in the court’s rul‐
ing on typicality.
D
The last requirement is adequate representation. See FED.
R. CIV. P. 23(a)(4). A named plaintiff must be a member of the
putative class and have the same interest and injury as other
members. Conrad v. Boiron, Inc., 869 F.3d 536, 539 (7th Cir.
No. 18‐1010 15
2017). A representative might be inadequate if he is subject to
a substantial defense unique to him. CE Design Ltd. v. King Ar‐
chitectural Metals, Inc., 637 F.3d 721, 726, 728 (7th Cir. 2011).
The district court generously characterized Speedy’s ade‐
quacy challenge as “scattershot.” We need not catalog every
objection Speedy raises, but we have considered all of them,
and we will mention a few. First, Speedy claims Beaton is not
actually a class member because he did not purchase the soft‐
ware as an individual. It cites the credit card statement billing
the charge to Beaton’s business, Chlorine Free Products Asso‐
ciation, for which he was the sole shareholder. But Beaton
averred in his declaration that he purchased the software for
a laptop that he personally owned and used for primarily per‐
sonal reasons. The software subscription was in Beaton’s
name. The district court did not clearly err in finding that Bea‐
ton purchased the software in his personal capacity.
Next, Speedy accuses Beaton of spoliating evidence—an
act that (it says) makes him an inadequate representative. But
spoliation is a harsh word for what happened (or so the dis‐
trict court could conclude). Beaton deleted a potentially use‐
ful email and took his laptop to an IT professional for repairs,
where his data were lost when the hard drive was reformat‐
ted. The district court rejected Speedy’s interpretation of this
incident when it denied Speedy’s motion for sanctions. It
found as a fact that Beaton did not intend to destroy evidence.
Speedy offers no reason for us to revisit that conclusion.
Speedy also launches a multipart attack on Beaton’s cred‐
ibility. It makes much ado of Beaton’s decades‐old man‐
slaughter conviction. But assaults on the credibility of a
named plaintiff must be supported by admissible evidence. Id.
at 728. Wholly unrelated criminal history does not fit that bill.
16 No. 18‐1010
See FED. R. EVID. 609(b) (a conviction’s probative value must
substantially outweigh its prejudicial effect in order to intro‐
duce it to impeach a witness over 10 years after his release);
e.g., United States v. Rogers, 542 F.3d 197, 201 (7th Cir. 2008).
Beaton’s various “lies” during discovery underlie
Speedy’s next attempt to discredit him. Some of these alleged
discrepancies are minor, such as his omitting a marijuana con‐
viction when asked about his criminal background. Beaton’s
supposed inconsistency in describing his laptop usage—that
he uses his laptop primarily for personal reasons but also for
business ones—is nothing of the sort; in fact, his statements
are consistent. Speedy does, however, point out one relevant
discrepancy. In both the complaint and his first set of inter‐
rogatories, Beaton professed to have purchased the software
for $39.94, while his credit card statement says that he paid
only $9.97. The district court did not abuse its discretion, how‐
ever, in concluding that Beaton’s credibility was not severely
undermined by this detail. See CE Design, 637 F.3d at 728. We
see no reason to disturb the court’s determination that Beaton
was an acceptable class representative.
Speedy also throws barbs at plaintiff’s counsel, Edelson
PC, citing allegations of wrongdoing made against the firm in
another case. Yet Speedy points to no evidence that Edelson is
unqualified, has created a conflict between the firm and the
putative class, or has violated a specific ethical rule. Speedy
may dislike Edelson PC, and we can assume it is not a fan of
class actions, but “general distaste for the class‐action device”
will not preclude certification. Mejdrech v. Met‐Coil Sys. Corp.,
319 F.3d 910, 912 (7th Cir. 2003). Nothing in this record per‐
suades us to consider Speedy’s request for sanctions under
Federal Rule of Appellate Procedure 38. The request is, in any
No. 18‐1010 17
event, procedurally irregular: Rule 38 requires sanctions re‐
quests to be filed in separate motions, see Vexol, S.A. de C.V. v.
Berry Plastics Corp., 882 F.3d 633, 638 (7th Cir. 2018), and it
does not contemplate sanctions against appellees.
IV
After clearing the hurdles posed by Rule 23(a), a person
wishing to bring a class action must also demonstrate that the
action fits under one of the three subsections of Rule 23(b). As
we said, the only one that applies to Beaton is Rule 23(b)(3),
the common‐question variant. It requires the putative class
representative to show that questions of law or fact common
to the class members predominate, and that the class device
is the superior method for adjudicating those claims.
A
The guiding principle behind predominance is whether
the proposed class’s claims arise from a common nucleus of
operative facts and issues. Messner, 669 F.3d at 815. This re‐
quires more than a tally of common questions; the district
court must consider their relative importance. Parko v. Shell
Oil Co., 739 F.3d 1083, 1085 (7th Cir. 2014). On the other hand,
not every issue must be amenable to common resolution; in‐
dividual inquiries may be required after the class phase. Kleen
Prods. LLC v. Int’l Paper Co., 831 F.3d 919, 922 (7th Cir. 2016).
Speedy identified 10 individual issues that allegedly de‐
feated predominance. The district court was not persuaded. It
found that some were best addressed on a class‐wide basis,
and they outweighed the remaining individualized inquiries.
The district court did not abuse its discretion in so con‐
cluding. For example, it will be easy to ascertain from whom
the class members purchased the software. The court found
18 No. 18‐1010
that they all bought it through the portal at the end of the free
trial that redirected customers to two payment platforms.
Similarly, the court found that users saw the same represen‐
tations about the software’s capabilities, and so a common an‐
swer to the question whether a reasonable customer would be
deceived is possible. And based on the court’s preliminary de‐
termination that the software’s diagnostic mechanisms oper‐
ated uniformly across devices, the trier of fact could reach a
single answer on the software’s functionality and value.
Speedy insists that the court needs to inquire individually
about each customer’s level of satisfaction with the product.
But dissatisfaction is not an element of any of the certified
claims. If the product truly serves none of its functions, its us‐
ers’ subjective satisfaction is likely evidence of misrepresenta‐
tion, not that the users were not harmed. See In re Aqua Dots
Prods. Liab. Litig., 654 F.3d 748, 750–51 (7th Cir. 2011) (purchas‐
ers suffered financial loss by paying more for products than
they would have had they known the products’ true quality).
Admittedly, some individualized questions remain. For
instance, what was the class member’s purpose (business or
personal?) in buying the software? Did the class member seek
a refund? What are each customer’s damages? Speedy re‐
minds us that we have frowned upon class treatment as a
poor fit for warranty and fraud claims because they can in‐
volve so many individualized issues. See Szabo, 249 F.3d at
674. But these theories do not automatically fail the predomi‐
nance test. See Amchem Prods., Inc. v. Windsor, 521 U.S. 591, 625
(1997) (certain consumer‐fraud cases readily establish pre‐
dominance); Suchanek v. Sturm Foods, Inc., 764 F.3d 750, 759–
60 (7th Cir. 2014) (the fact that “[e]very consumer fraud case
involves individual elements” does not preclude class ac‐
tions). Speedy misreads Supreme Court precedent in arguing
No. 18‐1010 19
that liability with regard to all class members must be resolved
in a single stroke. See Wal‐Mart Stores, Inc. v. Dukes, 564 U.S.
338, 350 (2011) (requiring resolution in “one stroke” of a
“common contention” central to the common claim); see also
Suchanek, 764 F.3d at 759–60; Pella Corp. v. Saltzman, 606 F.3d
391, 394 (7th Cir. 2010).
The district court recognized that individualized inquiries
could be handled through “streamlined mechanisms” such as
affidavits and proper auditing procedures. We agree.
Defendants’ due process rights are not harmed by such case‐
management tools. Mullins v. Direct Dig., LLC, 795 F.3d 654,
667–72 (7th Cir. 2015). Speedy’s attempts to distinguish
Mullins as merely about proving class membership, and not
liability, are unavailing. The company makes the obvious
point that it can neither cross‐examine an affidavit nor depose
every class member. But Speedy will still have the
opportunity to challenge the class members’ credibility. See
Mullins, 795 F.3d at 671. It can obtain the testimony of a
representative sample of the class members and, if necessary,
present evidence contradicting statements found in particular
affidavits.
Speedy also contends that there is a fatal lack of uni‐
formity in the purpose for which each person acquired its
software. We do not see that as a barrier to class treatment,
however. It is true that the law of British Columbia insists that
a particular purpose be brought clearly to the seller’s atten‐
tion. Compare Kobelt Mfg. Co. v. Pac. Rim Engineered Prods.
(1987) Ltd. (2011), 84 B.L.R 4th 189, para. 104 (Can. B.C. Sup.
Ct.) (leaky brakes did not violate an implied warranty because
no implied communication that purchasers intended to use
20 No. 18‐1010
the brakes on drawworks), with Wharton v. Tom Harris Chev‐
rolet Oldsmobile Cadillac Ltd. (2002), 97 B.C.L.R. 3d 307, para.
59–60 (Can. B.C. App. Ct.) (buzzing sound system violated
implied warranty where salesman knew purchasers wanted a
luxury vehicle). But we do not see that flaw here. The people
who used the free trial and then bought SpeedyPC Pro were
all concerned about the health and performance of their com‐
puters. Why they owned a computer is beside the point. To
the extent it is relevant, each user’s specific reason for buying
the software can be established through affidavits, subject to
the defendant’s right to challenge them with evidence.
B
Finally, the district court had several reasons for conclud‐
ing that a class action was the superior way to resolve this dis‐
pute. All are well‐supported. First, common questions of fact
and law predominate. Speedy insists that we should categor‐
ically reject class treatment for implied warranty and con‐
sumer fraud claims because of the choice‐of‐law clause. See,
e.g., Szabo, 249 F.3d at 674; In re Bridgestone/Firestone, Inc.,
288 F.3d 1012, 1020 (7th Cir. 2002). But that makes no sense
here, since all parties agree that British Columbia law controls
for the nationwide class and Illinois law for the subclass. And
there is no risk of inconsistent rules with respect to recogni‐
tion of the contractual choice‐of‐law clause, because that fol‐
lows the forum, Illinois. See Martin v. Reid, 818 F.3d 302, 308
(7th Cir. 2016).
Second, the amount of damages to which each plaintiff
would be entitled is so small that no one would bring this suit
without the option of a class. Suchanek, 764 F.3d at 759–60.
“Rule 23(b)(3) was designed for situations such as this, in
No. 18‐1010 21
which the potential recovery is too slight to support individ‐
ual suits, but injury is substantial in the aggregate.” Murray v.
GMAC Mortg. Corp., 434 F.3d 948, 953 (7th Cir. 2006). The fact
that others have not sued over this software is more likely be‐
cause “only a lunatic or a fanatic sues for $30,” Carnegie v.
Household Int’l, Inc., 376 F.3d 656, 661 (7th Cir. 2004), than it is
because the software is flawless. Consumer class actions are a
crucial deterrent against the proliferation of bogus products
whose sticker price is dwarfed even by a court filing fee (now
$400 for a civil case in federal district court). Though punitive
damages may also deter, few litigants would risk filing suit
on the off‐chance that punitive damages would be recovered
after years of litigation. See Hughes v. Kore of Ind. Enter., Inc.,
731 F.3d 672, 677–78 (7th Cir. 2013). The district court did not
abuse its discretion in finding the class‐action device superior.
V
Defendants spend much time and money fighting Rule 23
certifications to the hilt. Yet “certification is largely independ‐
ent of the merits … and a certified class can go down in flames
on the merits.” Schleicher v. Wendt, 618 F.3d 679, 685 (7th Cir.
2010). We say this not to imply that the merits in this case fa‐
vor either party, but simply to remind defendants that the
class‐action glass is sometimes half‐full: dismissed claims of a
certified class end litigation once and for all. That, after all, is
why settlement classes are so popular.
Finding no abuse of discretion in the district court’s deci‐
sions to certify the nationwide class and the Illinois subclass,
we AFFIRM the court’s certification orders.