PRECEDENTIAL

        UNITED STATES COURT OF APPEALS
             FOR THE THIRD CIRCUIT
                 _______________

                      No. 17-2933
                    _______________

     GALO COBA; COBA LANDSCAPING AND
CONSTRUCTION, INC., individually, and on behalf of other
    members of the general public similarly situated,
                                              Appellants

                              v.

               FORD MOTOR COMPANY
                  _______________

      On Appeal from the United States District Court
               for the District of New Jersey
                 (D.C. No. 2-12-cv-01622)
        District Judge: Honorable Kevin McNulty
                     _______________

        Submitted Under Third Circuit LAR 34.1(a)
                    January 23, 2019

  Before: JORDAN, KRAUSE, and ROTH, Circuit Judges

              (Opinion Filed: July 26, 2019)




John E. Stobart
Ryan Wu, I
Capstone Law
1875 Century Park East
Suite 1000
Los Angeles, CA 90067

      Counsel for Appellant
Robert M. Palumbos
Andrew R. Sperl
Duane Morris
30 South 17th Street
United Plaza
Philadelphia, PA 19103

John M. Thomas
Dykema
2723 South State Street
Suite 400
Ann Arbor, MI 48104

Karol C. Walker
LeClairRyan
1037 Raymond Boulevard
One Riverfront Plaza, 16th Floor
Newark, NJ 07102

      Counsel for Appellee

                      _______________

                OPINION OF THE COURT
                    _______________

KRAUSE, Circuit Judge.

        Problematic as a lemon of a vehicle may be, the
problem becomes more complex when it is peeling. This
appeal involves a putative consumer class action seeking
damages resulting from the delamination, i.e., peeling and
flaking, of the lining of certain Ford truck fuel tanks between
2001 and 2010—a problem that plagued numerous Ford F-
Series and E-Series vehicles in multiple countries and that,
according to Appellant Galo Coba, Ford knew stemmed from
a defect. It requires us to resolve two open questions for our
Court: first, whether a district court’s denial of class
certification divests the court of jurisdiction in a case where
its jurisdiction was predicated solely on the Class Action
Fairness Act, 28 U.S.C. § 1332(d); and second, whether a
warranty that covers only defects in “materials or


                              2
workmanship” extends to design defects under New Jersey
common law. We must also evaluate whether a reasonable
jury could conclude on this record that Ford knew the alleged
fuel-tank defect was the cause of the delamination problem at
the relevant time.

       Because we conclude that the District Court properly
exercised its jurisdiction, that the materials-or-workmanship
warranty did not cover design defects, and that the record
evidence of Ford’s knowledge about the defect does not
create a triable issue, we will affirm the District Court’s entry
of summary judgment in favor of Ford on all of Coba’s
claims.

 I.   Background

       A.   Ford’s Fuel Tank Troubles

       Beginning in 2001 and continuing over the decade that
followed, Ford received waves of complaints from customers
who purchased certain F-Series and E-Series vehicles
reporting similar types of malfunction related to their
vehicles’ fuel tanks. The fuel tanks used in certain vehicle
models were susceptible to a problem known as
“delamination,” whereby particles of the tank lining would
separate from the underlying metal and mix with the vehicle’s
fuel. As the fuel carrying those particles makes its way
through the vehicle’s fuel system, the particles can clog the
fuel filter, which constrains fuel flow to the engine and
reduces power. The particles can also damage fuel-system
components, such as injectors. If left untreated, the problem
eventually may lead to difficulties starting the engine or
keeping the vehicle running.

       In 2001, when Ford first received reports that some of
its vehicles were exhibiting fuel-tank delamination problems,
the complaints came exclusively from customers in Brazil.
Over the next few years, cases of delamination cropped up in
the United States, though they were largely clustered in
certain regions. For example, as of January 2004, forty-three
of the eighty-six warranty claims submitted to Ford that
related to delamination had come from customers in Ohio.
Because of the geographically concentrated occurrence of the
delamination complaints, both Ford and the supplier of its

                               3
fuel-tank coatings, Magni Industries, Inc., suspected that
unique qualities in regional fuel supplies were to blame for
delamination.      In particular, as Ford investigated, its
suspicions gravitated toward fuel containing excessive
concentrations of biodiesel, which Ford recommended against
using because its tanks were not authorized to withstand
biodiesel concentrations over 5%. That theory was consistent
with Ford’s data in some respects because Brazil, where the
problem started, did not have established biodiesel
regulations until 2005.

       Although Ford could not confirm that biodiesel was
the culprit—and Ford’s engineers sometimes questioned the
biodiesel hypothesis in light of inconclusive testing—Ford’s
leads were compelling enough that it started working with
Magni in 2005 to develop a more biodiesel-resistant coating.
And by February 2007, Ford released an improved coating,
called “A35,” to replace the prior “A36” coating in F-Series
Super Duty trucks. Around the same time, Ford sent a
message to dealers notifying them about the release of the
new tank coating and explaining that fuel tanks in certain
Ford trucks had delaminated, which Ford attributed to “the
use of fuels containing concentrations of bio-diesel greater
than recommended by Ford (5%).” App. 86.

       Ford’s warranty claims did drop after the release of the
A35 coating, but some reports of delamination persisted.
Having not fully solved the problem, Ford continued its
investigation. And by 2010, Ford’s Materials Engineering
department came to believe that biodiesel was not the root
cause after all; instead, acetic and formic acids—which Ford
discovered in fuel samples from service station pumps near a
dealer that encountered numerous delamination complaints—
were more likely the cause all along.

       B.   Coba’s Lawsuit

       Galo Coba, the plaintiff in this case, is one of the Ford-
vehicle customers whose fuel tanks delaminated.               He
purchased two Ford 2006 F-350 Super Duty 6.0L diesel dump
trucks for his landscaping business, Coba Landscaping and
Construction, Inc. He bought the first in October of 2006 and
the second in March of 2007. By March of 2009, both trucks
began exhibiting signs of tank delamination. According to

                               4
Coba, the engines would misfire, the trucks lacked power
when driven up hills, the fuel filters were contaminated with
fuel-tank debris, and the fuel systems rusted.

       He brought the trucks into a Ford dealership, which
replaced the fuel tanks and fuel filters in both trucks at no
cost to Coba. Despite the repairs, Coba had the same
problems over and over again, needing additional
replacements each time. Altogether, Coba replaced the fuel
tank twice in his older truck and three times in his newer
truck. Because several of the replacements occurred after the
trucks’ warranties had expired, Coba spent several thousand
dollars on the fixes.

       Coba filed this class-action lawsuit against Ford Motor
Company in March of 2012. As amended, the operative
complaint asserts claims for breach of express warranty,
violation of the New Jersey Consumer Fraud Act (NJCFA),
and breach of the duty of good faith and fair dealing.1
Although Ford had replaced several of Coba’s fuel tanks
under warranty, Coba alleges that Ford breached its written
warranty—the New Vehicle Limited Warranty (NVLW)—by
failing to adequately repair and replace his tanks, as the
replacements turned out to have the same defects as his
original tanks. The thrust of the implied-covenant-of-good-
faith-and-fair-dealing claim is that when Ford repaired Coba’s
vehicles, it knew that the repairs would not solve Coba’s
delamination problems. Finally, Coba’s NJCFA claim rests
on allegations that Ford purposefully failed to disclose to
Coba and other customers the defect in its fuel tanks.

       The District Court entered summary judgment in
Ford’s favor on all of Coba’s claims. See Coba v. Ford
Motor Co., No. 12-1622, 

, at *13–14
(D.N.J. Sept. 30, 2016); Coba v. Ford Motor Co., No. 12-
1622, 

, at *11 (D.N.J. Aug. 4, 2017). This
appeal followed.




      1
         It also asserts a common law fraud claim, which
Coba is no longer pursuing.


                              5
II.   Discussion

       A.   Jurisdiction

       We address a threshold issue of jurisdiction before
turning to the merits of the District Court’s decision. While
our jurisdiction to hear Coba’s appeal is clear under 28 U.S.C.
§ 1291, the propriety of the District Court’s jurisdiction is less
straightforward and an issue we must address at the outset.

        The District Court initially exercised jurisdiction over
Coba’s suit—a class action asserting state-law claims—
pursuant to the Class Action Fairness Act (CAFA), which
gives district courts “original jurisdiction of any civil action
in which the matter in controversy exceeds . . . $5,000,000 . . .
and is a class action in which . . . any member of a class of
plaintiffs is a citizen of a State different from any defendant.”
28 U.S.C. § 1332(d) (emphasis added). But when the District
Court entered summary judgment on three of Coba’s four
claims in September of 2016, it simultaneously denied Coba’s
motion for class certification as moot even though it had not
yet disposed of the NJCFA claim. Because § 1332(d)
provides original jurisdiction only over “class action[s],” that
ruling raises the question whether the District Court still had
jurisdiction when it entered its final summary judgment order
in August of 2017. Thus, before we address the merits of this
appeal, we must consider an issue of first impression for our
Court: If a federal court properly exercises jurisdiction
pursuant to § 1332(d) at the time a claim is filed or removed,
does a subsequent denial of class certification divest the court
of subject-matter jurisdiction?

      In accordance with every other Circuit Court to
address this question, we conclude that it does not.2 We start


       2
        See F5 Capital v. Pappas, 

, 75–77 (2d
Cir. 2017); Louisiana v. Am. Nat. Prop. Cas. Co., 

, 639–40 (5th Cir. 2014); Metz v. Unizan Bank, 

, 500–01 (6th Cir. 2011); Buetow v. A.L.S. Enters., Inc.,

, 1182 n.2 (8th Cir. 2011); United Steel, Paper
& Forestry, Rubber, Mfg., Energy, Allied Indus. & Serv.
Workers Int’l Union, AFL-CIO, CLC v. Shell Oil Co., 

, 1089 (9th Cir. 2010); Cunningham Charter Corp.

                                6
with the text: District courts have “original jurisdiction” over
“class action[s],” 28 U.S.C. § 1332(d)(2), which the statute
defines as “civil action[s] filed under [R]ule 23 . . . or [a]
similar State statute or rule of judicial procedure authorizing
an action to be brought . . . as a class action,” 

(emphasis added).             This conferral of
jurisdiction plainly encompasses a suit like Coba’s, which
was “filed under [R]ule 23,” notwithstanding its eventual
failure to become certified under Rule 23. See Metz v. Unizan
Bank, 

, 500 (6th Cir. 2011) (“The ‘filed under’
language shows that it is the time of filing that matters for
determining jurisdiction under CAFA.”); Cunningham
Charter Corp. v. Learjet, Inc., 

, 806 (7th Cir.
2010) (noting that § 1332(d)(1)(B) “defines class action as a
suit filed under a statute or rule authorizing class actions,
even though many such suits cannot be maintained as class
actions because the judge refuses to certify a class”). Indeed,
“[h]ad Congress intended that a properly removed class
action be remanded if a class is not eventually certified, it
could have said so.” United Steel, Paper & Forestry, Rubber,
Mfg., Energy, Allied Indus. & Serv. Workers Int’l Union,
AFL-CIO, CLC v. Shell Oil Co., 

, 1091 (9th
Cir. 2010).

        True, § 1332(d)(8) states that CAFA “shall apply to
any class action before or after the entry of a class
certification order by the court with respect to that action,”
but, as the Seventh Circuit has aptly noted, that subsection
refers to “a” certification order, not “the” certification order,
and the former connotes an indefinite expectation that a
certification order may issue. 

(explaining that subsection (d)(8) at most suggests that a class
“may be certified eventually” (emphasis added)). Moreover,
unlike subsection (d)(2), subsection (d)(8) omits reference to
“jurisdiction,” indicating it pertains not to the scope of
jurisdiction conferred by the statute, but to the timing of
certification in relation to removal. See 

Inc., 

, 806–07 (7th Cir. 2010); Vega v.
T-Mobile USA, Inc., 

, 1268 n.12 (11th Cir.
2009).


                               7
        Beyond CAFA’s text, general jurisdictional principles
also support our conclusion that the denial of class
certification did not divest the District Court of jurisdiction
over the NJCFA claim. Typically, “[j]urisdictional facts are
determined at the time of removal [or filing], not by
subsequent events.” Louisiana v. Am. Nat’l Prop. Cas. Co.,

, 635 (5th Cir. 2014); accord 

; 

–01; United 

–92. Of course, that principle is not absolute. See

(discussing exceptions, such as
mootness doctrine); United 

n.3
(same). However, as Congress did not make any exception
here, it seems “likely that Congress intended that the usual
and long-standing principles apply—post-filing developments
do not defeat jurisdiction if jurisdiction was properly invoked
as of the time of filing.” 

      Assured of the District Court’s jurisdiction, we turn to
the merits of the District Court’s summary judgment ruling.

       B.   The District Court’s Grant of Summary
            Judgment

        On appeal, Coba challenges the District Court’s grant
of summary judgment on his claims for breach of express
warranty, breach of the implied covenant of good faith and
fair dealing, and violation of the NJCFA. We review those
rulings de novo. See Blunt v. Lower Merion Sch. Dist., 

, 265 (3d Cir. 2014). Viewing the evidence “in the
light most favorable” to Coba as the non-moving party,
Plumhoff v. Rickard, 

, 768 (2014), we consider
whether Ford has shown “that there is no genuine dispute as
to any material fact and [that it] is entitled to judgment as a
matter of law,” Fed. R. Civ. P. 56(a). We address each claim
in turn.

              1. Breach of Express Warranty

      The District Court entered summary judgment on
Coba’s breach-of-express-warranty claim because it
determined that the fuel-tank defect at issue was outside the
scope of Ford’s written warranty, the NVLW. The District
Court reasoned (1) that the NVLW—which provides that
Ford will “repair, replace, or adjust all parts on [his] vehicle


                               8
that are defective in factory-supplied materials or
workmanship,” App. 248—covers only “materials or
workmanship” defects, not design defects, and (2) that the
fuel-tank defect alleged by Coba fell in the design-defect
category. We agree on both points.3

                 a. A Warranty for Defects in “Materials or
                    Workmanship” Does Not Encompass
                    “Design” Defects

       New Jersey law, which governs our interpretation of
the NVLW, see Collins v. Mary Kay, Inc., 

,
181–82 (3d Cir. 2017), does not specifically address whether
a warranty for “materials or workmanship” covers “design”
defects. In the absence of any guidance from New Jersey
courts on this particular issue, we “must predict how [New
Jersey’s] highest court would decide [it]” based upon
“relevant state precedents, analogous decisions, considered
dicta, scholarly works, and any other reliable data tending
convincingly to show how the highest court in the state would



      3
           Coba contends that we cannot affirm on these
grounds because Ford did not argue that design defects were
excluded from the NVLW’s coverage, and the District Court
did not provide notice to Coba that it was considering these
grounds sua sponte. See Fed. R. Civ. P. 56(f)(2) (“After
giving notice and a reasonable time to respond, the court may
. . . grant the motion [for summary judgment] on grounds not
raised by a party . . . .”); see also Couden v. Duffy, 

, 500 (3d Cir. 2006). But here, Ford did request summary
judgment on these grounds. See Ford Motor Company’s
Memorandum in Support of Motion for Summary Judgment
at 36 n.4, ECF No. 130-1 at 70 (Sept. 18, 2015). While Ford
did not flesh out the argument in detail in its summary
judgment briefing, it did make reference to the District
Court’s extensive discussion of that very issue, which was
sufficient to raise the argument. See In re Ins. Brokerage
Antitrust Litig., 

, 262 (3d Cir. 2009)
(recognizing that the relevant question as to waiver is whether
a party “presented the argument with sufficient specificity to
alert the district court”).


                              9
decide the issue at hand.” Berrier v. Simplicity Mfg., Inc.,

, 46 (3d Cir. 2009).

       We start with general principles of contract
interpretation under New Jersey law and give the terms of the
NVLW their “plain and ordinary meaning.” M.J. Paquet, Inc.
v. N.J. Dep’t of Transp., 

, 152 (N.J. 2002). If
those terms are unambiguous, resolution by summary
judgment is appropriate. See Michaels v. Brookchester, Inc.,

, 204 (N.J. 1958) (holding that “the construction
of a written agreement is a matter for the court,” not a jury,
unless “its meaning is uncertain or ambiguous”).

         The plain and ordinary meaning of the term “defect[s]
in . . . materials or workmanship,” App. 248, unambiguously
excludes “design” defects. As an initial matter, the plain
definitions of “workmanship” and “materials” are
conceptually distinct from the definition of “design.”
“Workmanship” is the “the execution or manner of making or
doing something,” Webster’s Third New International
Dictionary 2635 (1993),4 and “materials” are the “the basic
matter (as metal, wood, plastic, fiber) from which the whole
or the greater part of something physical (as a machine, tool,
building, fabric) is made,” 

Both definitions relate
to the execution phase of making an object and connote the
physical realization of something. By contrast, the definition
of “design”—“a preliminary sketch or outline (as a drawing
on paper or a modeling in clay) showing the main features of
something to be executed,” 

to the
preparation stage that guides, and precedes, execution.5 Thus,

       4
          The New Jersey Supreme Court regularly relies on
this dictionary and other versions of it when determining the
plain and ordinary meaning of terms. See, e.g., State v. Tate,

, 1204 (N.J. 2015); Highland Lakes Country
Club & Cmty. Ass’n v. Franzino, 

, 657 (N.J.
2006); Exxon Corp. v. Hunt, 

, 275 (N.J. 1984),
rev’d on other grounds, 

(1986).
       5
         While “workmanship,” “materials,” and “design”
each have multiple alternative definitions, we only highlight
the definitions that are most relevant to the context of the
issue before us, i.e., product development. But these words’

                              10
in the context of product development, defects in
“workmanship” and “materials” are flaws pertaining to the
construction or manufacture of a product, while defects in
“design” are shortcomings that arise in the plans for a
product’s creation. More specifically, a “materials” defect is
a failing in the quality of the actual substances used to make a
product, see Hammel v. Van Sickle, 

, 248 (N.J.
1925) (per curiam); a “workmanship” defect is a deficiency in
the execution of a product’s assembly or construction, see
Henningsen v. Bloomfield Motors, Inc., 

, 79 (N.J.
1960); and a “design” defect is a flaw inherent in the
product’s intended operation and construction, see O’Brien v.
Muskin Corp., 

, 304 (N.J. 1983).6

       Historical practice in products liability litigation,
dating back more than a century, reflects a consistent
understanding of the distinctions among these categories.
See, e.g., Lombard Corp. v. Quality Aluminum Prod. Co., 

, 338 (6th Cir. 1958) (“A defect in material is a
defect in quality. . . . A defect in workmanship is a defect in
the way some part of the machine is constructed. . . . Design,
on the contrary, involves the overall plan of construction and
operation.”); Moss v. Smith, 

, 385 (Cal. 1919) (“It
is conceded that the engine and clutch of the automobile in
question were defective, but the appellants claim that the



other definitions would not alter our reasoning; if anything,
they strengthen the meaning we ascribe to them. See, e.g.,
Webster’s Third New International Dictionary 2635 (1993)
(defining “workmanship” as “the quality imparted to a thing
in the process of making”); 

(defining “materials”
as “the finished stuff of which something physical (as an
article of clothing) is made”); 

(defining “design” as
“a mental project or scheme in which means to an end are laid
down”).
       6
        Although not relevant to the issues we address today,
we note that O’Brien was superseded by N.J.S.A. 2A:58C–
3a(2) to the extent it concerns the “consumer expectations”
doctrine. Dewey v. R.J. Reynolds Tobacco Co., 

, 1252 (N.J. 1990).


                              11
defects were those of design instead of material or
workmanship.”).7

       In light of this common law, it is unsurprising that
courts have regularly rejected arguments like Coba’s that a
design defect is within the scope of a materials-and-
workmanship warranty clause. See, e.g., Bruce Martin
Constr., Inc. v. CTB, Inc., 

, 753–54 (8th Cir.

       7
         See also S. Gas & Gasoline Engine Co. v. Adams &
Peters, 

, 677 (Tex. Civ. App. 1917), rev’d on
other grounds, 

(Tex. Comm’n App. 1921) (“A
careful examination of the voluminous testimony of the
witness . . . discloses that there was much of it that went to
other matters than mere defects in the design or plan of the
engine and its various parts, that is, to defects in
workmanship and material . . . .”); Dalton Adding Mach.
Sales Co. v. Denton, 

, 203 (Okla. 1925) (debating
whether a warranty covers only “defective materials and
workmanship” or also covers defects of “design”); Murdock
v. A. A. Sutain, Ltd., 

, 431 (Sup. Ct. 1955)
(“[T]he materials and workmanship were inferior. The
design, however, was the same . . . .”); Simmons v. Gibbs
Mfg. Co., 

, 822 (N.D. Ohio 1959) (“[T]he
criticism of the expert related only to the design of the top,
and not to the materials and workmanship.”); Shelby Mut. Ins.
Co. of Shelby, Ohio v. Ferber Sheet Metal Works, Inc., 

, 749 (Fla. Dist. Ct. App. 1963) (“[I]n his opinion
the roof developed the leak because of faulty design of the
flashing rather than due to the materials and workmanship
furnished by appellee, and . . . the architect on the job was
responsible for the design . . . .”); Totten v. Gruzen, 

, 5 (N.J. 1968) (“[L]iability may rest on architects and
engineers on the basis of improper design as well as on
contractors for defective materials, equipment and
workmanship.”); Falcon Tankers, Inc. v. Litton Sys., Inc., 

, 234 (Del. Super. Ct. 1972) (“The [issue] is whether
the language used in the guarantee clause limits damages so
as to preclude damages for design defects (as opposed to
defects in materials and workmanship) . . . .”).




                              12
2013); Voelker v. Porsche Cars N. Am., Inc., 

,
526–27 (7th Cir. 2003); Davidson v. Apple, Inc., No. 16-CV-
04942-LHK, 

, at *11–12 (N.D. Cal. Mar.
14, 2017); Rollolazo v. BMW of N. Am., LLC, No. CV 16-
00966, 

, at *8–9 (C.D. Cal. May 2, 2017);
Robinson v. Kia Motors Am., Inc., No. 13-006, 

, at *12 (D.N.J. Sept. 11, 2015); Nelson v. Nissan N.
Am., Inc., No. CIV. 11-5712, 

, at *2–3
(D.N.J. Dec. 19, 2014); Orthoflex, Inc. v. ThermoTek, Inc.,
No. 3:11–CV–0870–D, 3:10–CV–2618–D, 

, at *8 (N.D. Tex. Aug. 9, 2013); Rice v. Sunbeam
Prods., Inc., No. CV 12-7923, 

, at *12 (C.D.
Cal. Jan. 7, 2013); Horvath v. LG Elecs. Mobilecomm U.S.A.,
Inc., No. 3:11-CV-01576-H-RBB, 

, at *5
(S.D. Cal. Feb. 13, 2012).

        The two contrary district court decisions on which
Coba relies do not persuade us otherwise. Koulajian v. Trek
Bicycle Corp. provides almost no analysis to support its bare
conclusion that a “warranty’s reference to ‘workmanship’
could refer to . . . designs as well as to implementation of
those designs,” and thus, it offers nothing helpful for us to
consider. No. 90-Civ-3156, 

, at *2 (S.D.N.Y.
Feb. 11, 1992). And the logic of In re Saturn L-Series Timing
Chain Products Liability Litigation is that “design is
integrated into each step of the manufacturing process and
affects both materials and workmanship.” MDL No. 1920,

, at *15 (D. Neb. Nov. 7, 2008). That is
true, but it misses the point: While a design might dictate
what material or workmanship is required, it does not speak
to their quality.

       In short, we conclude that, under New Jersey law, a
warranty that limits its coverage to defects in “materials” and
“workmanship” does not, without more, apply to defects in
“design.” While parties are free to redefine words in their
contracts in ways that deviate from plain and ordinary
meaning, they did not do so here.             “Materials” and




                              13
“workmanship” in the NVLW carry their plain meaning, and
the warranty therefore does not extend to design defects.8

                 b. The Fuel Tank Defect Was a Design
                    Defect

       Having concluded that the NVLW does not cover
design defects, we must determine whether the fuel-tank-
delamination problem, as alleged, reflected a defect in design.
We agree with the District Court that it does, so the court
properly entered summary judgment on Coba’s breach-of-
warranty claim.

       Accounting for the differences between design,
materials, and workmanship defects, 

Section
II.B.1.a, the alleged flaw in Ford’s fuel tanks has all the
trappings of a design defect. The fundamental nature of the
defect relates to the “overall plan of construction and
operation” of the fuel tanks. Lombard 

. The problem, as consistently described by Coba, was
      8
         Coba asserts that Ford’s repeated replacements of
Coba’s fuel tanks constitute a course of performance that
should be given “controlling weight” in interpreting these
terms. Appellant’s Reply Br. 5. It is true that, for contracts
governed by New Jersey’s Uniform Commercial Code, which
Ford concedes is applicable, course of performance may be
used to “explain[]” or “supplement[]” a written agreement’s
terms. See N.J. Stat. Ann. § 12A:2-202; see also 

208. But, under the UCC, a meaning suggested by a course
of performance is trumped by the express terms of an
agreement. See 

208(2) (“The express terms of
the agreement and any such course of performance . . . shall
be construed whenever reasonable as consistent with each
other; but when such construction is unreasonable, express
terms shall control course of performance . . . .”). In any
event, Ford’s course of performance is not inconsistent with
our interpretation of the NVLW: Its willingness to repair and
replace some of Coba’s malfunctioning fuel tanks without
quibbling over whether and how the parts were defective
appears motivated by a desire to retain customer goodwill
rather than by an obligation to replace parts afflicted with
design defects.


                              14
not a low-quality supply of the A35 and A36 coatings or a
problem in the process for applying them to Ford’s fuel tanks;
rather, it was Ford’s plan to use those coatings at all in
constructing its fuel tanks.9 And that flawed-design theory is
consistent with the evidence on which Coba relies, including
Ford’s own conclusion in 2010 that tanks were delaminating
because “the A36 and A35 fuel tank coatings cannot tolerate
a constant supply of acetic and formic acids in fuel.” App.
1203. It is also consistent with Coba’s allegations that “[a]ll”
of the vehicles manufactured this way suffer from a
“common” issue, App. 83, and that the “root cause” of
delamination was that “all of the tanks at issue, as designed,
were susceptible to delamination” when exposed to certain
acids, Memorandum in Support of Motion for Class
Certification at 2, ECF No. 132 (Sept. 18, 2015) (emphasis
added). See Schwartz v. Volvo N. Am. Corp., 

,
941 n.5 (Ala. 1989) (noting that a design defect exists when
“every product of a line is defective” (emphasis omitted)); cf.
also Restatement (Third) of Torts: Prod. Liab. § 2 (1998)
(distinguishing design defects from “manufacturing defects,”
which occur where a “product departs from its intended
design”).

      As Coba alleged a design defect, and the NVLW
covered only materials and workmanship defects, the District
Court properly granted summary judgment on Coba’s breach-
of-warranty claim.

              2. Breach of Covenant of Good Faith and Fair
                 Dealing

       Because Coba did not have any right to repair or
replacement of his fuel tanks under the NVLW, he also could
not prevail on his claim for breach of the implied covenant of
good faith and fair dealing. New Jersey recognizes an
       9
          Although the allegations in Coba’s complaint
suggested that the fuel tanks might suffer from a
“manufacturing defect,” App. 83, the summary judgment
record is devoid of any evidence supporting the existence of
such a defect. And in neither his summary judgment briefing
nor his briefing on appeal has Coba argued that the fuel tanks’
manufacturing process was defective.


                              15
implied covenant of good faith and fair dealing in every
contract, Wilson v. Amerada Hess Corp., 

,
1126 (N.J. 2001), but to state a claim that it was breached, a
plaintiff must have “the right . . . to receive the fruits of the
contract” and must show that the defendant had “improper
motive” when interfering with that right, Wade v. Kessler
Inst., 

, 1259–60 (N.J. 2002) (citation omitted).

       Here, Coba alleges that Ford breached the covenant of
good faith and fair dealing implied in the NVLW by repairing
and replacing his tanks, while “knowing that those repairs and
replacements would not fix or remedy the [f]uel [t]ank
[d]efect.” App. 129. But even assuming Ford possessed an
improper motive—a questionable notion given the evolving
nature of Ford’s knowledge of a design defect—the NVLW
did not cover design defects, so tank repair and replacement
were not “fruits of the [NVLW]” that Coba had a “right . . . to
receive.” 

.

              3. New Jersey Consumer Fraud Act

       To prove a violation of the New Jersey Consumer
Fraud Act, N.J. Stat. Ann. §§ 56:8–1 to –210, a plaintiff must
establish “that the defendant engaged in an unlawful practice
that caused an ascertainable loss to the plaintiff,” Frederico v.
Home Depot, 

, 202 (3d Cir. 2007) (citing Cox v.
Sears Roebuck & Co., 

, 462–65 (N.J. 1994)).
There are three general types of “unlawful practices”:
“affirmative acts, knowing omissions, and regulation
violations.” 

647 A.2d at 462). A plaintiff
asserting a claim based on an omission must demonstrate that
the defendant “(1) knowingly concealed (2) a material fact (3)
with the intention that plaintiff rely upon the concealment.”
Judge v. Blackfin Yacht Corp., 

, 541 (N.J.
Super. Ct. App. Div. 2003); see also N.J. Stat. Ann. § 56:8-2.

       Here, Coba’s NJCFA claim rests on two theories, both
predicated on omissions by Ford: (1) that Ford knew and did
not disclose that the fuel tank suffered from a design defect
that caused delamination, and (2) that even if Ford did not
know the cause of the delamination, it failed to disclose the
risk. The District Court held, as to the first, that Coba failed
to put forth sufficient evidence of Ford’s knowledge of the
design defect, and, as to the second, that “the information

                               16
about the risk of delamination that Ford had available to it at
the time [Coba purchased his trucks] was not material.”
Coba, 

, at *4–9. For the reasons explained
below, we agree with both conclusions.

                 a. Ford’s Knowledge of the Design Defect

      To prevail on the theory that Ford failed to disclose a
known design defect, Coba would need to show that Ford had
that knowledge at the time of his purchases—i.e., before
March 9, 2007, when Coba purchased his second truck.

       Viewing the evidence “in the light most favorable” to
Coba, 

, no reasonable jury could
find Ford had that knowledge. Internal email correspondence
shows that, as early as 2005, Ford knew that the problem had
existed for several years and was investigating its cause. But
the evidence does not show that Ford knew that the cause was
the design of its tanks. To the contrary, it shows that
throughout the relevant period, Ford suspected the problem
was the improper use by certain customers of fuel with high
biodiesel concentrations, which seemed plausible in light of
the geographic clustering of delamination occurrence and the
phenomenon’s origin in Brazil. In September 2006, which
was one month before Coba purchased his first Ford vehicle,
a meeting of Ford managers and engineers concluded that
“[t]he cause for damaged fuel tanks is biodiesel (both refined
and the home brewed type) with bio concentrations greater
than 20% (Ford only authorizes concentrations up to 5%).”
App. 1145. And biodiesel fuel remained the prime suspect in
February 2007, just before Coba bought his second truck, as
apparent in both the message Ford then sent to dealers
explaining that fuel containing biodiesel at high
concentrations might cause delamination, and its development
of the more biodiesel-resistant A35 coating that it released
that month. As the District Court concluded, there was no
genuine dispute that at the time Coba bought his trucks, Ford
“believ[ed] that the problem was due to instances of
contaminated fuel, affecting a limited number of tanks, rather
than a defectively designed tank.” Coba, 

,
at *8.

      Although Coba posits that the District Court only
reached this conclusion by “construing all facts and drawing

                              17
all inferences . . . in favor of [Ford],” Appellant’s Br. 56, the
evidence to which he points fails to raise a triable issue. Coba
relies primarily on a 2005 email from a Ford engineer
remarking that recent tests of tanks exhibiting delamination
uncovered “no bio-diesel traces” and noting that they were
looking at “different additives that could cause [the]
delamination.” App. 963. But, viewing this evidence in the
light most favorable to Coba, it shows that some Ford
engineers had doubts whether biodiesel was the problem and
they were continuing to investigate. It does not support the
inference, as Coba contends, that Ford knew the problem was
a design defect and that biodiesel was a “pretext,” Appellant’s
Reply Br. 20. Cf. United States v. One 1973 Rolls Royce,
V.I.N. SRH-16266 By & Through Goodman, 

, 809
n.13 (3d Cir. 1994) (distinguishing between knowledge and
“suspicion followed by a failure to make further inquiry”).
Nor is that inference supported either by the correspondence
to which Coba points concerning the mere prevalence of the
delamination problem or by other correspondence that post-
dates his truck purchases and thus has no bearing on Ford’s
earlier knowledge.

       Because there is no genuine dispute of material fact as
to Ford’s knowledge of a design defect, its failure to disclose
that alleged defect does not give rise to liability under the
NJCFA.

                  b. Materiality of Delamination Risk

       Coba fares no better with his alternative theory that
Ford violated the NJCFA by failing to disclose material
information about the risk of delamination. To establish that
information withheld was “material,” Coba would need to
show that “a reasonable [person] would attach importance to
its existence in determining his [or her] choice of action.”
Suarez v. E. Int’l Coll., 

, 89 (N.J. Super. Ct. App.
Div. 2012). But the most favorable evidence in the record for
Coba concerning the rate of delamination comes from Ford’s
expert who analyzed Ford’s warranty database and found that
Ford was replacing Magni-lined steel fuel tanks for model
year 2003-2007 F-series trucks like Coba’s at a rate of less
than 1% across the United States. That replacement rate,
moreover, included all tank replacements, not merely those


                               18
related to delamination, suggesting an even lower
replacement rate for delaminated tanks. And while Coba
criticizes the warranty data as under-inclusive because it
covered only tanks that Ford actually replaced while
excluding those denied warranty coverage, he identifies no
concrete evidence of a higher rate of delamination.10

       In any event, the relevant question is not the actual rate
of delamination viewed in hindsight, but what Ford knew and
therefore could have disclosed to customers about that rate.
And the warranty data—reflecting delamination-based
replacements at a rate of even less than 1%—was the
information Ford had at the time. As to that small percentage,
based on the undisputed evidence that Ford then believed
biodiesel to be the culprit and the recommendation in its
owner’s manual against using those fuels, Ford had every
reason to believe that risk was mitigated—as would any
reasonable customer in possession of that same information.
We therefore agree with the District Court that “[n]o
reasonable factfinder could conclude that this information
would be material to a reasonable consumer prospectively
deciding, in March 2007, whether to purchase a Ford 6.0L
diesel truck.”      Coba, 

, at *10.
Accordingly, Coba’s second NJCFA theory, predicated on
non-disclosure of the risk of delamination, also does not
survive summary judgment.


       10
          For example, Coba relies on a document stating that
“[i]n April 2008, FCSD management indicated a higher than
normal sales volume (500/month) for Diesel Fuel Tanks due
to delamination concerns,” App. 1348, for the proposition that
“Ford admit[ted] that at one point it was replacing over 500
tanks per month due to delamination,” Appellant’s Br. 61.
Inartfully phrased as the document may be, however, on
closer inspection, it is apparent that it cannot plausibly bear
the weight that Coba places on it. To the contrary, it reflects
that—whatever the extent to which delamination concerns
may have contributed in part to the increased monthly sales
volume—in the entire eight years between 2001 and 2008
Ford had identified in total only 448 verified delamination
concerns.


                               19
III.   Conclusion

        For the foregoing reasons, we will affirm the judgment
 of the District Court.




                              20