The Appeal Is Won Before the Brief Is Written

Here is a sentence that sounds like a paradox and is actually a rule of thumb: most appeals are decided by a question that never appears in the jury charge, never came up at trial, and that most clients have never heard of. It is the standard of review—the legal dial that tells the court of appeals how hard to squint at the decision below.

Turn the dial all the way to the left, to de novo review, and the appellate court looks at the question as though no one had answered it before, owing the trial judge precisely nothing. Turn it all the way to the right, to abuse of discretion, and the trial judge can be wrong in the appellate court's private estimation and still win, so long as the ruling sits somewhere inside the broad zone the law calls reasonable. In between sits clear error, the standard that protects a trial judge's findings of fact and that swallows the overwhelming majority of factual challenges whole.

The same cold record—the same exhibits, the same transcript, the same marks side by side on the same page—can produce a reversal under one standard and an affirmance under another. Nothing about the facts changes. Only the dial moves. That is why seasoned appellate lawyers will tell you that the first question in any appeal is not "Was the district court wrong?" but "How wrong does it have to be, and about what kind of thing?" Get that wrong and you can write a brilliant brief that loses on a ground you never engaged.

Trademark appeals carry a special twist that makes the standard of review even more decisive than usual. The marquee question in nearly every infringement case—likelihood of confusion—is a mixed question of law and fact whose ingredients and whose bottom line are reviewed under different standards. Worse (or better, depending on which side you are on), the procedural posture in which the case reaches the Second Circuit changes the standard that governs that single question. Lose the confusion issue on summary judgment and you get fresh, deferential-to-no-one review. Lose it after a jury verdict and you hit a wall most appellants never climb. The merits never moved. The posture did.

This guide is a working map of how the Second Circuit reviews trademark cases, written for practitioners, business owners, and the occasional curious judge. It connects throughout to the substantive and procedural topics we treat elsewhere—the Polaroid factors on summary judgment, the allocation of issues between judge and jury, and the discretionary rulings on Lanham Act attorney's fees, Rule 403 prejudice, and Daubert challenges to survey experts. The standard treatise on the subject is Wright & Miller, Federal Practice and Procedure §§ 2581 et seq., and the foundational Supreme Court cases run from United States Gypsum through U.S. Bank v. Lakeridge. None of it is mysterious once you see the architecture.

What an Appeal Is, and What It Is Not

Start with a clarification that saves clients a great deal of disappointment: an appeal is not a do-over. The court of appeals does not hear live witnesses, does not take new evidence, does not let you try the case you wish you had tried, and—as a default—does not entertain arguments you neglected to make below. It reads the record that already exists, in the cold, and it reviews specific rulings for specific kinds of error. The mythology of the appeal as a "second bite" is mostly that—mythology. What an appeal actually offers is a chance to show that a particular ruling, judged by the particular standard that governs it, cannot stand.

That framing matters because it foregrounds the dial. Every issue you raise on appeal is, in effect, a small motion: "This ruling was the kind of mistake that this standard of review does not tolerate." A pure legal ruling reviewed de novo tolerates almost no mistake—if the appellate court reads the statute differently, you win. A factual finding reviewed for clear error tolerates a great deal—you can be confident the finding was wrong and still lose, because "wrong" and "clearly erroneous" are not the same word. A discretionary ruling reviewed for abuse of discretion tolerates the most of all. The art of the appeal is matching your strongest arguments to the standards that give them room to breathe, and declining to spend your limited credibility on arguments the standard will smother.

The Three Core Standards

Almost every issue on appeal is reviewed under one of three standards. Internalize the differences and the rest of this guide—indeed, the rest of appellate practice—falls into place.

De novo review (also called plenary review) means the appellate court decides the question anew, owing no deference to the district court. It asks itself, simply, what the right answer is, as though answering for the first time. De novo review governs pure questions of law: the interpretation of the Lanham Act, the articulation of the governing legal test, the construction of a contract's plain terms, and—this one trips people up—the grant of summary judgment, which is itself a legal determination that there is nothing to try. Because no deference is owed, de novo issues are where appeals are won. If you can frame your strongest point as a question of law, you have done a large part of the strategic work of the appeal.

Clear-error review governs findings of fact, principally those a judge makes after a bench trial. Federal Rule of Civil Procedure 52(a)(6) is the operative text: a trial court's findings of fact "must not be set aside unless clearly erroneous, and the reviewing court must give due regard to the trial court's opportunity to judge the witnesses' credibility." The canonical gloss comes from United States v. United States Gypsum Co., 333 U.S. 364, 395 (1948): a finding is clearly erroneous only when the reviewing court, "on the entire evidence," is "left with the definite and firm conviction that a mistake has been committed." This is deference with teeth. The appellate court does not reweigh the evidence; where two views of it are each permissible, the trial judge's choice between them is by definition not clear error. Anderson v. City of Bessemer City, 470 U.S. 564, 573–74 (1985). The appellant must show not that another finding was possible but that the finding made was a mistake of which the court is firmly convinced. That is a much harder thing to prove, and most factual challenges die here.

Abuse-of-discretion review governs the long list of rulings the law entrusts to the trial judge's discretion: granting or denying injunctions, admitting or excluding evidence, awarding attorney's fees, imposing discovery sanctions, bifurcating a trial, managing the courtroom. A court abuses its discretion when it (1) rests its ruling on an error of law, (2) rests it on a clearly erroneous finding of fact, or (3) renders a decision that—though legally and factually sound—falls outside the range of permissible decisions. See Zervos v. Verizon N.Y., Inc., 252 F.3d 163, 168–69 (2d Cir. 2001) (the Second Circuit's standard formulation). This standard is deferential—two able judges can rule opposite ways on the same motion and both be affirmed—but notice the trapdoor built into it. Embedded inside abuse-of-discretion review is de novo review of any legal premise and clear-error review of any factual predicate. A discretionary ruling that rests on a misstatement of the law is reviewed, on that point, without deference. That trapdoor is the appellant's best friend, and we return to it repeatedly.

These three words—law, fact, discretion—are the entire vocabulary. The craft lies in classifying each issue correctly, because the classification largely sets the odds.

When You Can Appeal at All: Finality and Its Exceptions

Before the dial even spins, there is a gate: is the ruling appealable yet? In the federal system the default is the final-judgment rule of 28 U.S.C. § 1291, which grants the courts of appeals jurisdiction over "final decisions"—those that end the litigation on the merits and leave the district court nothing to do but execute. The rule's purpose is anti-fragmentation: it spares the appellate courts a parade of mid-case detours and keeps litigation moving. Its cost is patience. A losing party usually cannot appeal an adverse interim ruling—a denied summary-judgment motion, an evidentiary call, a discovery order—until the whole case is over.

That default has a feature trademark lawyers exploit constantly. Because the injunction is the beating heart of trademark relief, the interlocutory appeal of injunctions under 28 U.S.C. § 1292(a)(1) looms large. A party may immediately appeal an order granting, continuing, modifying, refusing, or dissolving an injunction—including a preliminary injunction—without waiting for final judgment. This is why the most consequential event in a fast-moving brand dispute, the preliminary-injunction ruling, so often reaches the Second Circuit within months, reviewed (as we will see) for abuse of discretion. A second, narrower route is the certified interlocutory appeal under 28 U.S.C. § 1292(b): the district court may certify that an order "involves a controlling question of law as to which there is substantial ground for difference of opinion" and that immediate appeal "may materially advance the ultimate termination of the litigation," after which the court of appeals must agree to hear it. And a sliver of cases qualify for the judge-made collateral order doctrine of Cohen v. Beneficial Industrial Loan Corp., 337 U.S. 541 (1949), for the rare ruling that conclusively decides an important issue separate from the merits and effectively unreviewable later.

Two practical corollaries follow. First, identify for every adverse ruling both whether and when it can be challenged. A preliminary-injunction loss can be tested at once; almost everything else waits for final judgment. Second, beware a quiet trap of the finality rule: a denial of summary judgment ordinarily is not reviewable after a full trial, because the trial record supersedes the summary-judgment record. See Ortiz v. Jordan, 562 U.S. 180 (2011). The party who loses a summary-judgment motion, then loses the trial, generally cannot resurrect the summary-judgment ruling on appeal; it must have renewed its arguments through a Rule 50 motion on the trial evidence. The lesson is one you will hear again: the appeal is built at trial, not after it.

How Summary Judgment Is Reviewed (and Why It Tilts)

A grant of summary judgment is reviewed de novo, and understanding the reason illuminates a surprising amount of trademark appellate practice. When a district court grants summary judgment, it does not find facts. It holds, as a matter of law, that there is no genuine dispute of material fact and that the moving party is entitled to judgment as a matter of law under Rule 56(a). Because that is a legal conclusion rather than a factual finding, the court of appeals owes it no deference and applies the same standard the district court was bound to apply.

And that standard has a thumb on the scale. On summary judgment—and therefore on de novo review of summary judgment—the court views the evidence in the light most favorable to the non-moving party and draws all reasonable inferences in that party's favor, asking only whether a reasonable jury could return a verdict for the non-movant. Anderson v. Liberty Lobby, Inc., 477 U.S. 242, 248, 255 (1986); Celotex Corp. v. Catrett, 477 U.S. 317 (1986). The court does not decide whom to believe or weigh competing accounts; it asks whether the dispute is genuine enough that it must go to a factfinder. The Second Circuit applies exactly this lens to trademark summary judgments. See Star Industries, Inc. v. Bacardi & Co., 412 F.3d 373, 379, 383 (2d Cir. 2005) (reviewing the grant of summary judgment de novo while reviewing the trial court's findings for clear error).

The consequence is significant and counterintuitive: an appellant challenging a grant of summary judgment is not fighting deference at all. It stands on equal footing with the district court, arguing that the record contained a genuine dispute that should have gone to trial. That is structurally easier than persuading a court that a finding was clearly erroneous or that a discretion call was out of bounds. It also explains why the confusion question, when resolved on summary judgment, is reviewed without deference—the subject of the next section, and of our companion piece on the Polaroid factors on summary judgment.

The Crux: Reviewing Likelihood of Confusion and the Polaroid Factors

Now to the heart of the matter. The single most important and most misunderstood question in trademark appeals is how the likelihood-of-confusion determination, and the underlying Polaroid factors, are reviewed. Polaroid Corp. v. Polarad Electronics Corp., 287 F.2d 492 (2d Cir. 1961), gave the Second Circuit its eight-factor framework for confusion: the strength of the senior mark, the similarity of the two marks, the proximity of the products, the likelihood the senior user will bridge the gap, actual confusion, the defendant's good faith, the quality of the junior product, and the sophistication of the relevant consumers. The appellate question is what kind of question each of those is—and what kind of question the bottom line is.

The Second Circuit's answer is a bifurcated standard, and getting it right is the whole game on a confusion appeal.

The individual Polaroid factor findings are treated as findings of fact. How strong is the senior mark? How similar are the marks to the eye and ear? How close are the products in the marketplace? Was there evidence of real-world confusion? Did the defendant act in good faith or set out to free-ride? Each of those is a factual determination, and after a bench trial each is reviewed for clear error. They rest on documents, testimony, surveys, and—crucially—the trial judge's read of the witnesses, which is precisely the kind of work Rule 52(a) shields.

The ultimate balancing, however—the bottom-line judgment of whether a likelihood of confusion exists—is treated as a legal conclusion, reviewed de novo. Weighing the factors against one another to reach the legal result is a legal act, and the appellate court reviews that weighing without deference, even while it defers to the subsidiary factual findings that feed into it. The settled formulation: the factor findings are facts (clear error); the balance is law (de novo). Bristol-Myers Squibb Co. v. McNeil-P.P.C., Inc., 973 F.2d 1033, 1043–46 (2d Cir. 1992).

Why split a single inquiry down the middle this way? The logic is conceptual, and it is worth internalizing because it predicts how other trademark questions will be reviewed. The individual factors are anchored in this record—what these marks look like, what these consumers think, what this defendant intended. Resolving them is fact-intensive work the trial court is uniquely positioned to do, having lived with the case. The balance, by contrast, is about how the law tells courts to combine those facts into a legal conclusion, and keeping that legal judgment consistent across cases is the appellate court's distinctive job. The doctrine, in other words, sorts the case-specific (defer) from the law-declaring (review freshly). Hold onto that sorting principle; it is the engine of the whole "mixed question" framework we unpack below.

The practical payoff for the appellant is concrete. After a bench trial, you usually must concede the factual findings (or show they were clearly erroneous, a steep climb) while arguing that the trial court misweighed them—gave too much weight to the absence of actual confusion, say, or too little to the marks' similarity. That misweighing argument is a de novo argument, and it is where a confusion appeal has its best chance.

A Detour Into History: The De Novo-Versus-Clear-Error Split

The bifurcated standard did not arrive fully formed, and the history is more than antiquarian. For decades the circuits—and panels within the Second Circuit—disagreed about whether the ultimate confusion conclusion was a finding of fact reviewed for clear error or a legal conclusion reviewed de novo. Some early decisions treated likelihood of confusion as essentially factual, reviewable only for clear error, on the theory that confusion is a real-world phenomenon and the trial court is closest to it. Others insisted that, because confusion is the legal standard for liability, the ultimate determination must be a legal one reviewed without deference, lest the governing rule fracture case by case.

Bristol-Myers Squibb (1992) is the decision that clarified the modern Second Circuit position: the underlying factor findings are reviewed for clear error, but the ultimate balancing is reviewed de novo. The court reasoned that while the predicate facts deserve Rule 52(a) deference, the weighing of those facts to reach the legal conclusion of confusion is a question on which appellate courts owe no deference, precisely so that the legal standard remains uniform. The Second Circuit has applied that two-tier approach consistently since. See, e.g., Star Industries, 412 F.3d at 383–84; Cadbury Beverages, Inc. v. Cott Corp., 73 F.3d 474, 477–78 (2d Cir. 1996).

The split is not merely a Second Circuit memory; it persists across the circuits today, which is why a forum's location can change the odds on the very same record. The Federal Circuit, reviewing Trademark Trial and Appeal Board decisions and applying the DuPont factors, treats likelihood of confusion as a legal conclusion reviewed de novo, with the underlying factor findings reviewed for substantial evidence. The Ninth Circuit, by contrast, has historically reviewed the entire confusion determination after a bench trial for clear error, treating the ultimate question as predominantly factual. The upshot is that an appellant who could not move a confusion finding in San Francisco might have a real shot at the balance in New York—same facts, different dial—because the circuits classify the mixed question differently. That divergence is exactly the kind of thing the Supreme Court's later guidance on mixed questions, discussed below, was meant to discipline.

The Procedural-Posture Overlay: One Question, Three Standards

Here is the insight that ties everything together and that a surprising number of litigants miss: the same likelihood-of-confusion question is reviewed under three different standards depending on the procedural posture in which the case reached the Second Circuit. Confusion does not carry one fixed standard of review. It carries three, and which one applies depends entirely on how the case was decided below.

On appeal from summary judgment, the entire confusion question is reviewed de novo, with the evidence viewed in the light most favorable to the non-movant. The appellate court asks whether, on that favorable view, a reasonable jury could find confusion (or could not). The factor "findings" are not findings at all, because nothing was tried; they are the court's assessment of whether the record permits only one conclusion. The Second Circuit reviews such rulings without deference and has repeatedly cautioned that summary judgment on the fact-laden confusion question is appropriate only where the Polaroid factors point so decisively in one direction that a reasonable jury could reach but one result. Cadbury Beverages, 73 F.3d at 478 (reversing summary judgment and emphasizing the caution required before resolving confusion as a matter of law). This is the most appellant-friendly posture.

On appeal from a bench trial, the bifurcated standard governs: factor findings for clear error, ultimate balance de novo, per Bristol-Myers Squibb. The middle of the spectrum.

On appeal from a jury verdict, the most deferential standard of all applies. The jury's findings—including its confusion verdict—are reviewed under the reasonable-jury standard: the verdict stands unless no reasonable jury could have reached it on the evidence. A motion for judgment as a matter of law under Rule 50, and its denial, is itself reviewed de novo—but "de novo" here is almost a trap for the unwary, because the court applies that fresh review through the highly deferential reasonable-jury lens, drawing all inferences for the verdict winner and refusing to weigh credibility. Reeves v. Sanderson Plumbing Products, Inc., 530 U.S. 133, 150–51 (2000). In practice a confusion verdict is nearly bulletproof so long as the evidence permitted it.

The strategic implication is enormous, and it runs in a direction most clients find startling: the posture in which you win or lose below largely dictates your appellate odds on the confusion question, before anyone briefs the merits. Lose on summary judgment and you get de novo review and a genuine chance of reversal. Lose after a jury verdict and you hit the reasonable-jury wall. Lose after a bench trial and you occupy the middle—free to attack the balance de novo but stuck with the factual findings absent clear error. Which means appellate posture is something to think about before trial, not after. The road you take to the courthouse determines the road you can take through the appeal, a point that connects directly to the judge-versus-jury choice we develop at length elsewhere. A defendant confident in the equities might prefer a bench trial precisely to set up a de novo misweighing argument; a plaintiff confident in a sympathetic jury might prefer to lock its confusion finding behind the reasonable-jury wall.

The Clear-Error Standard, Up Close

Because so much of a trademark case turns on findings of fact—the Polaroid factors, the distinctiveness of a mark, genericness, secondary meaning, abandonment, actual confusion—the clear-error standard is the one an appellant most often confronts and most often loses under. It rewards a closer look.

Clear-error review is deferential by design, not by accident. The Supreme Court's reasoning in Anderson v. Bessemer City is worth absorbing because it explains why the deference is so stubborn. The trial judge saw the witnesses, handled the exhibits, and absorbed the case in three dimensions; the appellate court has only a cold paper record in two. Duplicating the trial court's work on appeal "would very likely contribute only negligibly to the accuracy of fact determination at a huge cost in diversion of judicial resources." Anderson, 470 U.S. at 574–75. The deference is at its absolute zenith for findings that rest on credibility, which an appellate court is in no position to second-guess; but even findings drawn from documents and inferences receive clear-error protection. Id. at 574.

The operative principle—commit it to memory, because it decides cases—is the two-permissible-views rule: "Where there are two permissible views of the evidence, the factfinder's choice between them cannot be clearly erroneous." Id. The appellant must show not that another finding was available, but that the finding actually made was a mistake the appellate court is firmly convinced of. Those are different burdens, and the gap between them is where most factual appeals expire.

In trademark, the implications ripple outward. A great many of the determinations that decide cases are factual and thus shielded:

  • Where a mark falls on the Abercrombie spectrum of distinctiveness (generic, descriptive, suggestive, arbitrary, or fanciful);
  • Whether a once-protectable term has become generic;
  • Whether a descriptive mark has acquired secondary meaning;
  • Whether there was actual confusion, and how much weight it deserves;
  • Whether the defendant acted in bad faith; and
  • The amount of the defendant's sales, the costs proven, and the existence and extent of damages.

After a bench trial, every one of those is reviewed for clear error; after a jury trial, under the even more forgiving reasonable-jury standard. An appellant hoping to dislodge such a finding has two honest paths and one dishonest one. The honest paths: identify a genuine clear error (a finding flatly contradicted by uncontested evidence, or resting on a misreading of the record), or recast the dispute as a legal error—a misstatement of the governing standard—which is reviewed de novo. The dishonest path, which courts see through immediately, is to dress up a plain disagreement with the facts in the borrowed clothes of "legal error." The recasting works only where the trial court truly misstated the law; it is futile, and faintly insulting to the panel, where the appellant simply wishes the judge had believed a different witness.

Mixed Questions and How They Get Sorted

The likelihood-of-confusion balance is one instance of a mixed question of law and fact, and the Supreme Court has given courts a general method for deciding how such questions are reviewed—a method that explains, in one stroke, why confusion's ingredients and its conclusion are treated differently.

The governing case is U.S. Bank National Ass'n v. Village at Lakeridge, LLC, 583 U.S. 387 (2018). The standard of review for a mixed question, the Court held, depends on whether answering it is predominantly factual or predominantly legal—on "what kind of work" the question requires. Id. at 395–96. Where the mixed question demands immersing in case-specific facts and drawing case-specific inferences, review is more deferential (clear error); where it demands expounding the legal standard or developing the governing rule, review is more plenary (de novo). The inquiry is functional, not formal: which sort of work does answering the question mainly involve?

Apply that lens to trademark and the bifurcation falls out automatically. Resolving each Polaroid factor is predominantly factual—what these marks and consumers and this defendant did—so the factor findings are reviewed for clear error. Combining the factors into the ultimate confusion conclusion is predominantly a matter of applying and elaborating the legal standard, so the balance is reviewed de novo. The same sorting recurs throughout the field: distinctiveness classification and genericness lean factual; the articulation of the legal test leans legal; the application of an equitable defense like laches blends both. Lakeridge is the master key. Once you can place a given mixed question on the fact-law spectrum, you can predict its standard of review—and brief it accordingly.

Abuse-of-Discretion Review: Injunctions, Fees, and Evidence

Many of the rulings that shape a trademark case are committed to the trial court's discretion and reviewed for abuse of discretion. An appellant attacking them must clear a high bar—and the smart appellant clears it by finding the legal error hiding inside the discretionary call.

Injunctions—again, the central trademark remedy—are reviewed for abuse of discretion. A grant or denial of preliminary or permanent injunctive relief is affirmed unless the trial court applied the wrong legal standard, rested on a clearly erroneous fact, or made a decision outside the range of permissible choices. The legal components are reviewed de novo within that framework: the proper articulation of the four-factor eBay test (eBay Inc. v. MercExchange, L.L.C., 547 U.S. 388 (2006)) and, in trademark specifically, the rebuttable presumption of irreparable harm that the Trademark Modernization Act of 2020 restored at 15 U.S.C. § 1116(a). If a district court applies a categorical irreparable-harm presumption that the statute does not support, or omits an eBay factor altogether, that is a legal error reviewed without deference even though the ultimate injunction call is discretionary. We unpack the equitable-relief machinery in our bench-versus-jury article.

Evidentiary rulings—including Rule 403 balancing and Daubert reliability determinations—are likewise reviewed for abuse of discretion, and the deference is real. The Supreme Court confirmed in General Electric Co. v. Joiner, 522 U.S. 136, 141–43 (1997), that a trial court's decision to admit or exclude expert testimony is reviewed deferentially, given the court's superior vantage on reliability—even as the legal framework (whether the court applied the right Daubert test at all) is reviewed de novo. The same deferential posture governs the Rule 403 weighing of probative value against unfair prejudice and the admissibility of consumer-survey experts, which are perennial battlegrounds in confusion cases. Surveys, in particular, are usually challenged on weight rather than admissibility, and weight is for the factfinder—another reason these rulings rarely fall on appeal.

Attorney's fee awards under § 1117(a) are reviewed for abuse of discretion, including the threshold "exceptional case" determination. The Supreme Court held as much in the patent context (the relevant statute is materially identical) in Highmark Inc. v. Allcare Health Management System, Inc., 572 U.S. 559, 563–64 (2014), rejecting de novo review of exceptionality and instructing appellate courts to defer. Because the Lanham Act's fee-shifting language tracks the Patent Act's after Octane Fitness, LLC v. ICON Health & Fitness, Inc., 572 U.S. 545 (2014), courts apply the same deferential review to trademark fee awards. The practical effect: a fee award is hard to overturn unless the district court used the wrong legal standard for exceptionality. We develop this in our article on Lanham Act attorney's fees.

Discovery sanctions—including preclusion of undisclosed evidence under Rule 37(c)(1), the subject of our Rule 37 article—and case-management decisions like bifurcation under Rule 42(b) are also reviewed for abuse of discretion. Across this whole category, the appellant's most reliable path is the trapdoor: identify an embedded legal error, a misstatement of the governing standard, because that single point is reviewed de novo and can topple the discretionary ruling that rests on it. Attack a discretionary ruling as merely unwise and you will lose; attack the legal premise it stands on and you may win.

Reviewing Damages and Profits Awards

Monetary awards get their own layered review—really, all three standards at once, applied issue by issue. Pull the award apart and review each piece under its own dial:

  • The legal availability of a remedy—whether an accounting of profits is available at all, whether a category of damages is permitted under the Lanham Act, how a statutory provision is construed—is a question of law reviewed de novo. The Supreme Court's holding in Romag Fasteners, Inc. v. Fossil, Inc., 590 U.S. 212 (2020), that willfulness is not an absolute precondition to a profits award under § 1117(a), is a pure legal ruling of this kind.
  • The factual findings underlying an award—the defendant's gross sales, the costs proven, the existence and amount of actual damages—are reviewed for clear error after a bench trial (and under the reasonable-jury standard after a jury trial).
  • The discretionary judgments inherent in an equitable monetary award—whether to order an accounting at all, how to exercise the apportionment and equitable-adjustment powers the statute confers, whether to enhance or reduce an award "according to the circumstances of the case"—are reviewed for abuse of discretion.

This means a challenge to a damages or profits award must be aimed, not sprayed. An appellant arguing that the district court applied the wrong legal rule for apportionment has a de novo argument with real traction. An appellant arguing merely that the court should have credited different sales or cost figures faces clear-error deference. An appellant arguing that the court should have exercised its equitable discretion differently faces the abuse-of-discretion wall. A disciplined appeal separates these threads and concentrates fire on the component carrying the least deference, rather than attacking "the damages" as an undifferentiated lump. And the discipline starts at trial: to preserve a strong appellate position, frame your objections below in terms of the legal standard wherever a genuine legal error exists, because building the legal objection into the record is what later opens the door to de novo review.

Preserving Issues: The Appeal Is Built at Trial

A point that decides more appeals than litigants expect is preservation, and it is mostly a matter of housekeeping done—or neglected—in the district court. An appellate court generally will not entertain an argument that was not raised below, and it reviews only the record that was actually made—no new evidence, no new theories, no second thoughts. A party that failed to object to evidence, to move under Rule 50, or to press a legal argument at trial may forfeit the issue, leaving at most demanding plain-error review, which in civil cases is often no meaningful review at all. See Fed. R. Civ. P. 51 (jury-instruction objections); Puckett v. United States, 556 U.S. 129, 135 (2009) (forfeiture and plain error).

The mechanics are specific and unforgiving:

  • Evidentiary objections must be made contemporaneously and, where a ruling is merely provisional, renewed when the evidence is offered. A definitive in-limine ruling, however, preserves the issue without renewal under Federal Rule of Evidence 103(b)—a distinction we develop in our Rule 403 article. Know which kind of ruling you got, because the difference is dispositive.
  • Sufficiency-of-the-evidence challenges after a jury trial generally require a Rule 50(a) motion before the case goes to the jury and a renewed Rule 50(b) motion after the verdict. Skip the renewal and you forfeit the sufficiency challenge—this is the most common preservation disaster in trademark trials, and it is fatal. Unitherm Food Systems, Inc. v. Swift-Eckrich, Inc., 546 U.S. 394 (2006).
  • Legal arguments and theories must be presented to the district court; an argument raised for the first time on appeal is ordinarily waived.

Even a preserved error will not produce reversal if it was harmless—if it did not affect substantial rights. Federal Rule of Civil Procedure 61 and 28 U.S.C. § 2111 codify the harmless-error principle, and it is the appellee's last and often best line of defense: yes, the ruling was wrong, but it did not change the outcome. An appellant must therefore be ready to show not only error but prejudice—that the error, more probably than not, affected the result.

The throughline could not be clearer: the issues available on appeal, and the standards under which they are reviewed, are fixed by what counsel did and said below. The appeal is built at trial. We treat the mechanics at length in our forthcoming piece on preserving error for appeal.

A Word on FRAP and the Clock

Standards of review do not operate in a vacuum; they ride on top of the Federal Rules of Appellate Procedure, and a missed deadline forecloses the appeal no matter how favorable the standard would have been. The threshold rule is jurisdictional and merciless: in a civil case the notice of appeal must be filed within 30 days after entry of the judgment or order, Fed. R. App. P. 4(a)(1)(A), and this deadline is treated as jurisdictional, Bowles v. Russell, 551 U.S. 205 (2007). Certain post-trial motions—a Rule 50(b) renewed JMOL, a Rule 52(b) motion to amend findings, a Rule 59 motion for new trial or to alter the judgment—toll the appeal clock until the district court disposes of them, Fed. R. App. P. 4(a)(4), which is one more reason those motions matter beyond their immediate stakes.

From there, FRAP governs the architecture of the appeal: Rule 3 (how an appeal is taken), Rule 10 (the record on appeal—a reminder that the appellate court sees only what was made below), Rule 28 (the contents of briefs, which in the Second Circuit by local rule require an explicit statement of the standard of review for each issue), and Rule 32 (form and length). That standard-of-review statement is not a formality. It is the brief's opening move on the dial, and a panel reads it to learn, before a word of argument, how much deference each issue carries. Frame it carelessly—concede clear error where you could have argued de novo—and you have given away ground you did not need to cede.

A Worked Example: Meridian v. Meridia (Hypothetical)

To see how posture drives review, consider an invented dispute decided three different ways. This scenario is entirely hypothetical and illustrative; it is not a real case. "Meridian," an established watchmaker, sues "Meridia," a newer competitor, for trademark infringement in the Eastern District of New York. The central question is likelihood of confusion under the Polaroid factors. Suppose the case ends in three alternative ways, and Meridian appeals the adverse outcome in each.

Scenario one — summary judgment. The district court grants summary judgment to Meridia, holding no likelihood of confusion as a matter of law. On appeal, the Second Circuit reviews de novo, viewing the evidence in the light most favorable to Meridian and drawing all reasonable inferences in its favor, asking whether a reasonable jury could find confusion. Under Cadbury, the court is wary of resolving the fact-intensive confusion question on summary judgment; if the Polaroid factors do not point unmistakably one way—if, say, the marks are quite similar and the products overlap but actual-confusion evidence is thin—the grant will be reversed and the case sent to trial. Meridian's odds are at their best here: no deference, a favorable inference rule, and a circuit that prefers to try close confusion cases. Note what Meridian wins, though: a trial, not a judgment. (More on that in a moment.)

Scenario two — bench trial. The case is tried to the judge, who finds the marks only weakly similar, no actual confusion, and good faith on Meridia's part, and concludes there is no likelihood of confusion. On appeal, the Second Circuit reviews the factor findings (similarity, actual confusion, good faith) for clear error and the ultimate balance de novo, per Bristol-Myers Squibb. Meridian will struggle to overturn the factual findings—under the two-permissible-views rule, the judge's assessments stand if reasonable, and "the marks aren't that similar" is a quintessential permissible view. But Meridian can argue de novo that the judge misweighed the factors—for example, by treating the absence of actual confusion as dispositive when the Second Circuit has long held that actual confusion is not required and its absence is rarely conclusive. Meridian's odds are moderate, and they live entirely in the balance, not the findings.

Scenario three — jury verdict. The case goes to a jury, which returns a verdict of no confusion. Meridian moved under Rule 50(a) before submission and renewed under Rule 50(b) after the verdict (good lawyering—skip either and the challenge is forfeited). On appeal from the denial of the renewed motion, the court reviews under the reasonable-jury standard: the verdict stands unless no reasonable jury could have found as it did. If the evidence permitted a finding of no confusion—and with weak similarity and no actual-confusion proof, it plainly did—the verdict is essentially unassailable. Meridian's odds are lowest.

Same dispute. Same Polaroid factors. Same marks on the same page. Three different standards, three very different appellate prospects—all determined not by the merits but by the posture in which the case happened to be decided. That is the lesson of the standard of review distilled to a single hypothetical.

What Winning Actually Buys: The Remand Problem

A final point that routinely surprises non-lawyers: winning an appeal usually does not end the case in the winner's favor. When the Second Circuit reverses or vacates, it most often remands—sends the case back to the district court—rather than entering judgment itself. What the remand looks like depends on the standard that produced the reversal:

  • A reversal for legal error typically returns the case for application of the correct legal standard, which may mean a new trial, renewed proceedings, or reconsideration under the proper rule.
  • A reversal of a grant of summary judgment ordinarily means the case proceeds to trial, because the appellate court has held only that a genuine dispute exists—not that the appellant wins. (This is Scenario one above: Meridian's prize is a trial date, not a trophy.)
  • A reversal for clear error in a factual finding, or for abuse of discretion, commonly returns the matter for new findings or a renewed exercise of discretion under the corrected approach.

Only occasionally—where the undisputed facts admit of one reasonable conclusion—will the court direct entry of judgment for the appellant. More often, the reward for a successful appeal is the opportunity to litigate again under better rules. That reality should temper how a client values an appeal. An appellate victory is frequently a second chance, not a final win, and the cost, delay, and uncertainty of further proceedings on remand belong in the decision whether to appeal at all and how to frame the relief sought. The disciplined appellant asks for the specific remand it actually wants—new findings on a single factor, say, rather than a fresh trial—and shapes the argument to earn it.

Practical Takeaways

For the appellant, the first task is triage: classify every issue by its standard of review, then build the appeal around the issues carrying the least deference. Lead with de novo issues—legal errors and challenges to a grant of summary judgment—where the court owes nothing and reversal is realistic. When a finding is factual, ask yourself honestly whether it was clearly erroneous (a firm-conviction-of-mistake showing) or whether your real complaint is a legal error in the standard applied, which can be reframed for de novo review. Do not squander credibility attacking permissible factual findings under the guise of disagreement; panels notice. For confusion after a bench trial, target the balance (de novo), not the factor findings (clear error). Hunt for the legal error embedded in any discretionary ruling—the trapdoor. Confirm that each issue was preserved below, because a forfeited issue is usually a dead one. And mind the clock: the notice of appeal and the standard-of-review statement are the first two moves, and both are easy to fumble.

For the appellee, the strategy is the mirror image: lean into deference at every turn. Characterize the disputed rulings as factual (clear error) or discretionary (abuse of discretion) wherever legitimate; press the two-permissible-views rule and the trial court's superior vantage; and after a jury verdict, invoke the reasonable-jury wall and refuse to let the appellant relitigate credibility. Defend the trial court's legal framework as correct, so no de novo foothold opens inside a deferential standard. Argue harmless error as a fallback even where you concede nothing—wrong but immaterial is a winning posture. And press preservation: if the appellant failed to raise or renew an argument below, say so plainly and early.

For both sides, the unifying principle is that in the Second Circuit, as everywhere, the standard of review is the lens through which the merits are seen—and in trademark cases that lens shifts with the procedural posture. Summary judgment is reviewed de novo; bench-trial fact findings for clear error with the confusion balance de novo; jury verdicts under the reasonable-jury standard; discretionary rulings for abuse of discretion. The individual Polaroid factors are facts; their balance is law. Master those distinctions, build the record to preserve the issues that matter, and choose the trial posture with the appeal already in mind, and the appellate phase becomes a contest you have prepared for rather than one whose rules you discover, too late, in the opinion that affirms the other side.

Frequently Asked Questions

Is the likelihood-of-confusion question reviewed de novo or for clear error in the Second Circuit? Both—it depends on what you are challenging and the posture. After a bench trial, the individual Polaroid factor findings are reviewed for clear error and the ultimate balancing of those factors for de novo. Bristol-Myers Squibb Co. v. McNeil-P.P.C., Inc., 973 F.2d 1033 (2d Cir. 1992). On appeal from summary judgment, the whole question is de novo with inferences drawn for the non-movant. On appeal from a jury verdict, it is reviewed under the deferential reasonable-jury standard.

Why does the same confusion question get reviewed three different ways? Because the standard of review attaches to what the district court did, not to the abstract issue. Summary judgment is a legal determination (de novo); a bench-trial finding is fact-finding (clear error) topped by a legal balance (de novo); a jury verdict is protected by the reasonable-jury standard. Same question, three procedural vehicles, three dials.

Can I appeal a denial of summary judgment after I lose the trial? Generally no. Once the case is tried, the trial record supersedes the summary-judgment record, and the denial is ordinarily unreviewable. Ortiz v. Jordan, 562 U.S. 180 (2011). To preserve a sufficiency challenge you must renew your arguments through a Rule 50(b) motion on the trial evidence.

What is "clear error," exactly? A finding is clearly erroneous only when, on the whole record, the reviewing court is "left with the definite and firm conviction that a mistake has been committed." United States v. United States Gypsum Co., 333 U.S. 364, 395 (1948). Where two views of the evidence are each permissible, the trial court's choice between them cannot be clear error. Anderson v. City of Bessemer City, 470 U.S. 564 (1985).

Are trademark injunctions easy to overturn on appeal? No. Injunctions are reviewed for abuse of discretion, a deferential standard. The realistic path to reversal is to show the district court applied the wrong legal standard—for example, an irreparable-harm presumption inconsistent with the Trademark Modernization Act's § 1116(a)—because legal premises within a discretionary ruling are reviewed de novo.

How are attorney's fee awards under § 1117(a) reviewed? For abuse of discretion, including the "exceptional case" determination, which the Supreme Court held is entitled to deferential review in Highmark Inc. v. Allcare Health Management System, Inc., 572 U.S. 559 (2014). See our Lanham Act attorney's fees article.

What happens if I forget to renew my Rule 50 motion after the verdict? You likely forfeit any sufficiency-of-the-evidence challenge, and the appellate court may decline to review it at all. Unitherm Food Systems, Inc. v. Swift-Eckrich, Inc., 546 U.S. 394 (2006). The Rule 50(a) motion before submission and the Rule 50(b) renewal after the verdict are both required.

Does winning my appeal mean I win the case? Usually not. The Second Circuit most often remands for further proceedings under the corrected standard. A reversed summary judgment means a trial, not a judgment. Only where the undisputed facts admit one answer will the court direct entry of judgment in the winner's favor.

Does it matter that other circuits review confusion differently? Yes, if you have any choice of forum. The Federal Circuit treats likelihood of confusion as a legal conclusion reviewed de novo; the Ninth Circuit has historically reviewed the whole determination for clear error after a bench trial; the Second Circuit splits the difference with its bifurcated approach. The same record can carry different appellate odds in different circuits.

Related Articles

This article is provided for general informational purposes and does not constitute legal advice. Standards of review are nuanced and can turn on the precise posture and issue presented; consult qualified appellate counsel about any particular matter.