Every year, tens of thousands of trademark applicants receive office actions from the USPTO. These refusals are governed by the Lanham Act (15 U.S.C. 1051 et seq.), and examining attorneys follow the Trademark Manual of Examining Procedure (TMEP) when evaluating applications. For likelihood of confusion refusals under Section 2(d), the legal standard is the DuPont factors framework (In re E.I. du Pont de Nemours & Co., 476 F.2d 1357, C.C.P.A. 1973), which weighs up to thirteen factors -- though in practice, examiners focus on a few key factors. Other refusal types have their own TMEP sections and legal standards: descriptiveness (TMEP 1209), specimens (TMEP 904), disclaimers (TMEP 1213), identification of goods/services (TMEP 1402), and geographic significance (TMEP 1210).

Some applicants respond and register their marks. Others respond and don't. The difference between the two groups has never been studied at scale -- until now.

We analyzed 13,888 verified winning trademark office action cases to answer a straightforward question: which response strategies actually lead to registration on the Principal Register? Not which strategies practitioners think work, or which textbooks recommend, but which ones succeeded in real cases with real outcomes.

That analysis covered 14 distinct refusal types, extracted and categorized 18,896 individual arguments, and compared winning responses against 11,570 losing cases to identify the patterns that separate success from failure.

The results reveal clear patterns in how successful applicants structure their responses, which argument types dominate by refusal category, and how many arguments are enough.

Data current as of April 2026.

Methodology

Data Source and Scope

These findings come from the winning_arguments pipeline, which tracks trademark office action cases from initial refusal through final disposition. Every case in the dataset has a verified terminal outcome -- we can confirm whether the applicant ultimately achieved registration on the Principal Register.

What Counts as a Win

A case qualifies as a "verified win" only if the mark achieved registration on the Principal Register. This is a deliberate threshold. Cases where the applicant did not achieve Principal Register status -- including 1,956 that were reclassified as defeats because the outcome fell short of full registration -- were excluded from the winning dataset.

The dataset also includes 4,252 cases still pending verification and 20 partial wins, neither of which are counted in the 13,888 verified wins.

Argument Taxonomy

Each response was broken down into individual arguments, which were classified into two broad categories:

Argue strategies: Substantive arguments challenging the examiner's legal conclusion. These include arguments about commercial impression, mark distinction, goods coexistence, weakness of cited marks, third-party registrations, secondary meaning, and others.
Comply strategies: Amendments or concessions that narrow the application to avoid the refusal. These include goods/services amendments, class deletions, disclaimers, consent agreements, and others.

A single response can contain both argue and comply arguments. Most winning responses do.

Losing Cases

The 11,570 losing cases were analyzed separately to identify strategies that appear frequently in failed responses. These comparisons help distinguish between strategies that genuinely contribute to success and strategies that are merely popular.

A Note on Interpretation

These statistics represent real outcomes from real trademark prosecution cases -- they are not legal advice. The DuPont framework requires a case-by-case analysis, and no dataset can substitute for evaluating the specific facts, evidence, and circumstances of an individual application. Practitioners should treat these patterns as informative benchmarks, not prescriptive rules, and evaluate each case on its own merits.

Argue vs. Comply by Refusal Type

The following table shows all 14 refusal types in the dataset, sorted by case volume. For each, the argue percentage indicates how much of the winning argument mix was substantive argumentation versus compliance-based amendments.

Note: Because a single case can involve multiple refusal types, the case counts below sum to more than 13,888. The total reflects unique case-refusal combinations.

Refusal Type	Cases	Argue %	Top Argue Strategies	Top Comply Strategies
Section 2(d) -- Likelihood of Confusion	3,459	58%	Commercial impression (1,393), Mark distinction (1,245), Goods coexistence (859)	Goods amendment (3,110), Class deletion (523)
Identification of Goods/Services	2,206	36%	Commercial impression (605), Mark distinction (421), Goods coexistence (353)	Goods amendment (3,275), Class deletion (406)
Disclaimer Requirement	1,842	59%	Commercial impression (1,157), Mark distinction (1,037), Weakness of cited mark (453)	Goods amendment (1,367), Disclaimer (947)
Specimen Refusal	1,590	55%	Commercial impression (596), Mark distinction (422), Specimen support (399)	Goods amendment (1,532), Class deletion (454)
Section 2(e)(1) -- Mere Descriptiveness	1,264	62%	Commercial impression (526), Mark distinction (476), Weakness of cited mark (227)	Goods amendment (569), Disclaimer (508)
Section 2(e)(3) -- Geographic Descriptiveness	116	70%	Commercial impression (53), Mark distinction (48), Secondary meaning (32)	Goods amendment (42), Disclaimer (31)
Section 2(e)(2) -- Deceptively Misdescriptive	89	71%	Commercial impression (35), Secondary meaning (27), Mark distinction (25)	Goods amendment (28), Disclaimer (24)
Ornamental Refusal	59	79%	Commercial impression (38), Goods coexistence (19), Third-party registrations (15)	Goods amendment (24), Class deletion (5)
Section 2(e)(4) -- Primarily a Surname	42	61%	Secondary meaning (19), Commercial impression (10), Goods coexistence (9)	Goods amendment (22), Disclaimer (11)
Section 2(a) -- Immoral/Scandalous/Deceptive	32	83%	Mark distinction (32), Commercial impression (23), Weakness of cited mark (6)	Goods amendment (11), Consent agreement (6)
Other Refusals	32	70%	Commercial impression (20), Mark distinction (19), Weakness of cited mark (5)	Goods amendment (13), Class deletion (6)
Section 2(e)(5) -- Functional Matter	31	67%	Mark distinction (17), Weakness of cited mark (13), Commercial impression (13)	Goods amendment (27), Disclaimer (2)
Failure to Function as a Mark	6	50%	Goods coexistence (2), Mark distinction (2), Secondary meaning (1)	Class deletion (3), Goods amendment (3)
Section 2(c) -- Name/Portrait of Living Person	3	50%	Cited mark abandoned (1), Foreign language (1), Mark distinction (1)	Consent agreement (2), Goods amendment (1)

What the Table Reveals

Several patterns emerge from this breakdown.

Argue-heavy refusal types have fewer cases. The refusal types where argue strategies dominate (Section 2(a) at 83%, ornamental at 79%, Section 2(e)(2) at 71%, geographic descriptiveness at 70%) all have relatively low case counts. This makes intuitive sense: these refusals often involve clear-cut factual disputes where the examiner may have misjudged the mark's nature, and a well-reasoned argument can resolve the issue.

High-volume refusals show more compliance. Identification of goods/services refusals have the lowest argue rate at 36%, meaning nearly two-thirds of winning arguments in this category were compliance-based. When the examiner says your goods description is unacceptable, the most efficient path to registration is usually to fix the description.

Section 2(d) sits in the middle. The most common refusal type shows a 58/42 argue-comply split. This reflects the dual nature of likelihood of confusion responses: you can argue why the marks are different, but you can also narrow your goods to reduce overlap with the cited registration.

Goods amendment is the universal comply strategy. It appears as the top comply strategy in 13 of 14 refusal types. In identification refusals, goods amendment alone accounts for 3,275 arguments -- more than all argue strategies in that category combined.

The Argument Count Sweet Spot

Not all successful responses use the same number of arguments. We analyzed the argument count distribution across winning cases to identify whether there is an optimal number of arguments per response.

Arguments per Response	Winning Cases	Percentage
1	1,199	25.6%
2	1,454	31.0%
3	1,033	22.0%
4	544	11.6%
5	300	6.4%
6	118	2.5%
7	22	0.5%
8	12	0.3%
9	1	0.0%
10	6	0.1%

Reading the Distribution

1-2 arguments is the most common count. Together, one-argument and two-argument responses account for 2,653 winning cases (56.6% of the distribution). This does not mean fewer arguments are better -- it means most cases can be resolved with targeted precision. A single, well-supported goods amendment can eliminate an identification refusal entirely. A goods amendment paired with a focused commercial impression argument is often sufficient for a Section 2(d) refusal. The data suggests that the majority of office actions do not require elaborate, multi-pronged responses.

Two-argument responses are the single most common winning pattern at 1,454 cases (31.0%). This aligns with the argue-plus-comply model: one substantive argument addressing the examiner's core concern, and one compliance-based amendment removing a secondary objection. It is a lean approach, and the data shows it works.

3-5 arguments is the sweet spot for complex refusals. Responses with three to five arguments account for 1,877 winning cases (40.0%). This range represents the ideal balance between thoroughness and focus for cases that require more than basic compliance. Respondents in this range tend to combine argue and comply strategies -- challenging the examiner's core finding while also making strategic amendments to narrow the application.

Cases with three arguments are particularly common (1,033 cases, 22.0%), suggesting a natural pattern: one argue point, one comply point, and one supporting element. A typical three-argument Section 2(d) response might combine a commercial impression argument (the marks create different overall impressions), a goods amendment (narrowing the applicant's identification to reduce overlap), and a third-party registrations argument (showing that similar marks coexist on the register for related goods).

Four- and five-argument responses (544 and 300 cases, respectively) tend to appear in cases with multiple refusal types. When an office action raises both a likelihood of confusion issue and a disclaimer requirement, the response naturally requires more arguments to address each ground.

6+ arguments correlate with failure, not thoroughness. Only 159 winning cases (3.4%) used six or more arguments. The drop-off from five to six arguments is dramatic: a 60.7% decline in winning cases. From six to seven, it drops another 81.4%. By the time a response reaches eight or more arguments, it represents fewer than 20 winning cases total.

This pattern is not coincidental. The losing case data reinforces it convincingly. Failed responses tend to throw every available argument at the wall, often stacking five, six, or seven variations on the same theme. Each additional argument beyond the core few dilutes the overall response. An examiner reading a seven-argument commercial impression brief may conclude that if the applicant truly had a strong case, three arguments would have sufficed. More arguments can actually signal a weaker position -- if you need seven distinct reasons why your mark is different from the cited registration, perhaps none of those reasons is particularly strong on its own.

There is an important exception to this pattern: cases involving multiple distinct refusal types (such as a Section 2(d) refusal combined with a specimen refusal and a disclaimer requirement) naturally require more arguments because each refusal needs its own response. The diminishing returns apply to arguments aimed at a single refusal, not to responses addressing multiple independent issues.

The Takeaway

Draft your response with discipline. Identify the two or three strongest arguments for each refusal, support them thoroughly, and make any necessary amendments. Resist the temptation to include every conceivable point. A focused response with strong evidence outperforms a comprehensive response with thin evidence every time.

Top Strategies Across All Refusal Types

When we aggregate the argument data across all 14 refusal types, clear winners emerge in both the argue and comply categories.

Top Argue Strategies

Strategy	Total Uses	Appears in Top 3 For
Commercial impression	4,469	12 of 14 refusal types
Mark distinction	3,745	12 of 14 refusal types
Goods coexistence	1,242	5 of 14 refusal types
Weakness of cited mark	704	5 of 14 refusal types
Specimen support	399	1 of 14 refusal types
Secondary meaning	79	3 of 14 refusal types
Third-party registrations	15	1 of 14 refusal types

Commercial impression is the dominant argue strategy with 4,469 uses across the dataset. It appears in the top three argue strategies for 12 of 14 refusal types. This makes sense: regardless of the specific refusal, arguing that your mark creates a distinct overall impression in the minds of consumers is almost always relevant.

Mark distinction is a close second at 3,745 uses, also appearing in the top three for 12 of 14 refusal types. Where commercial impression focuses on the meaning and connotation of the mark, mark distinction emphasizes structural, visual, and phonetic differences. The two strategies are complementary and often appear together in successful responses.

Goods coexistence ranks third with 1,242 uses. This strategy argues that even if the marks are similar, the goods or services are sufficiently different to avoid consumer confusion. It is most prominent in Section 2(d) cases (859 uses) where the relatedness of goods is a key factor in the likelihood of confusion analysis.

Weakness of cited mark appears 704 times and is particularly effective in disclaimer and descriptiveness refusals, where arguing that the cited mark or term is weak (generic, descriptive, or commonly used in the industry) can undercut the examiner's basis for refusal.

Secondary meaning is notable not for its volume (79 uses in the top-three data) but for its strategic importance in specific refusal types. It is the leading argue strategy for Section 2(e)(4) surname refusals (19 uses) and ranks in the top three for geographic descriptiveness and deceptively misdescriptive refusals. When the refusal is about whether your mark has acquired distinctiveness, secondary meaning evidence is the direct answer.

Top Comply Strategies

Strategy	Total Uses	Appears in Top 2 For
Goods amendment	10,024	14 of 14 refusal types
Disclaimer	1,523	7 of 14 refusal types
Class deletion	1,397	6 of 14 refusal types
Consent agreement	8	2 of 14 refusal types

Goods amendment dominates the comply category with 10,024 uses -- more than all other comply strategies combined. It appears in the top two comply strategies for every single refusal type. This underscores a fundamental reality of office action practice: when in doubt, tighten your goods description. A well-crafted amendment can resolve multiple refusals simultaneously.

Disclaimer is the second most common comply strategy at 1,523 uses. It is most significant in descriptiveness-related refusals (Section 2(e)(1), 2(e)(2), 2(e)(3), 2(e)(4), 2(e)(5)) and disclaimer requirement refusals, where disclaiming a descriptive or generic component of the mark can satisfy the examiner without giving up the overall registration.

Class deletion appears 1,397 times and is the primary alternative to goods amendment. Where a goods amendment narrows the description within a class, class deletion removes an entire class of goods or services from the application. This is a more drastic step but can be strategically sound when the refusal stems primarily from overlap in one specific class.

How Winners and Losers Differ

The comparison between winning and losing cases in Section 2(d) refusals is particularly instructive. The losing case data shows that many of the same strategies appear in both winning and losing responses:

Commercial impression was used in 4,046 losing cases with an average effectiveness score of 0.47
Goods amendment appeared in 3,021 losing cases with a 0.50 effectiveness score
Mark distinction was used in 3,087 losing cases with a 0.43 effectiveness score

The strategy itself is rarely the problem. What separates winners from losers is execution: the quality of the evidence, the precision of the legal analysis, and the strategic combination of arguments. A commercial impression argument with detailed side-by-side comparisons and consumer perception evidence succeeds where a conclusory assertion of "different impression" fails.

Consider the effectiveness scores from the Section 2(d) losing data more closely. Third-party registrations had the highest effectiveness score among losing strategies at 0.57, meaning even in losing cases, this argument moved the needle more than others. Consumer sophistication (0.51) and goods amendment (0.50) also scored above the midpoint. At the bottom, cited mark abandoned (0.41) and mark distinction (0.43) were the least effective in losing cases -- suggesting that when these arguments fail, they fail badly.

The implications are clear: choosing the right strategy matters, but developing it with sufficient evidence and precision matters more. Two applicants can both argue commercial impression in a Section 2(d) case. The one who provides a structured DuPont factor analysis with specific evidence addressing each factor wins. The one who states a conclusion without supporting it loses.

The same pattern holds for descriptiveness refusals (Section 2(e)(1)). The "suggestive not descriptive" argument -- the most common losing strategy at 390 uses -- has only a 43% effectiveness score. The argument is not inherently bad, but it requires substantial evidence to overcome the examiner's presumption that the mark is descriptive. Respondents who rely on it without that evidence lose more often than they win. By contrast, specimen support evidence in descriptiveness cases scored 0.48 -- not a guaranteed winner, but consistently more effective because it provides the examiner with concrete evidence of how consumers perceive the mark in the marketplace rather than abstract legal characterizations.

Secondary meaning claims in descriptiveness refusals scored 0.42, reinforcing that acquired distinctiveness is genuinely difficult to prove without substantial evidence of advertising expenditures, sales volume, consumer surveys, or five years of substantially exclusive use. Applicants who invoke secondary meaning as a throwaway argument rather than a supported evidentiary claim almost always lose.

Comparing Refusal Categories: The Argue-Comply Spectrum

The 14 refusal types fall into a spectrum from argue-heavy to comply-heavy:

Argue-dominant (65%+ argue):

Section 2(a) Immoral/Scandalous (83%)
Ornamental (79%)
Section 2(e)(2) Deceptively Misdescriptive (71%)
Section 2(e)(3) Geographic Descriptiveness (70%)
Other Refusals (70%)
Section 2(e)(5) Functional Matter (67%)

Balanced (50-64% argue):

Section 2(e)(1) Mere Descriptiveness (62%)
Section 2(e)(4) Primarily a Surname (61%)
Disclaimer Requirement (59%)
Section 2(d) Likelihood of Confusion (58%)
Specimen Refusal (55%)
Failure to Function (50%)
Section 2(c) Name/Portrait (50%)

Comply-dominant (<50% argue):

Identification of Goods/Services (36%)

This distribution tells a practical story. Refusals that challenge the fundamental nature of your mark (is it scandalous? is it ornamental? is it deceptively misdescriptive?) almost always require substantive argumentation. You cannot amend your way out of an examiner's conclusion that your mark is immoral or that your design is merely decorative.

Refusals that challenge how you described your goods and services are best resolved through compliance. When the examiner says your identification is too broad or uses unacceptable terminology, the most efficient path forward is to revise the language.

The middle ground -- likelihood of confusion, descriptiveness, disclaimer, and specimen refusals -- calls for a blended approach. The most successful responses in these categories combine substantive arguments with strategic amendments.

Refusal-Specific Strategy Profiles

For the six highest-volume refusal types, we have published dedicated strategy guides with deeper analysis, winning case studies, and losing-case comparisons:

Section 2(d) -- Likelihood of Confusion (3,459 cases): Governed by TMEP 1207.01, these refusals are evaluated under the DuPont factors, weighing the similarity of marks, relatedness of goods/services, and conditions of the marketplace. The dominant argue strategies (commercial impression and mark distinction) and why goods coexistence is the swing factor.
Identification of Goods/Services (2,206 cases): Under TMEP 1402, the identification must be "specific, definite, clear, accurate, and concise." The most compliance-heavy category and when to argue instead of amend.
Disclaimer Requirement (1,842 cases): TMEP 1213 requires applicants to disclaim unregistrable components of a composite mark that do not create a separate commercial impression. When to disclaim, when to fight, and how weakness-of-cited-mark arguments change the calculus.
Specimen Refusals (1,590 cases): Under TMEP 904, applicants must submit specimens showing the mark as actually used in commerce in connection with the identified goods or services. The unique role of specimen support evidence and why commercial impression still dominates.
Section 2(e)(1) -- Mere Descriptiveness (1,264 cases): TMEP 1209 provides that a mark is "merely descriptive" if it immediately conveys knowledge of a quality, feature, function, or characteristic of the goods or services. Why "suggestive not descriptive" fails more often than it succeeds and what works instead.
Section 2(e)(3) -- Geographic Descriptiveness (116 cases): Under TMEP 1210, the refusal requires that the mark's primary significance is a generally known geographic location and that consumers would make a goods/place association. How secondary meaning evidence and geographic incongruity arguments win at high rates.

The remaining eight refusal types have smaller sample sizes but reveal distinctive patterns worth noting:

Section 2(a) -- Immoral/Scandalous (32 cases, 83% argue): Governed by TMEP 1203, these refusals assess whether a substantial composite of the general public would find the mark scandalous or immoral. The highest argue rate in the dataset. Mark distinction is the leading strategy (32 uses), reflecting that these refusals are often factual disputes about what the mark actually conveys. You cannot amend your way out of a scandalousness finding -- you must convince the examiner that the mark does not carry the meaning they attributed to it.
Ornamental (59 cases, 79% argue): Under TMEP 1202.03, a proposed mark that is merely ornamental or decorative does not function as a trademark to indicate source. Commercial impression dominates here (38 uses), with applicants arguing that their design functions as a source identifier, not mere decoration. Third-party registrations (15 uses) appear as a top-three strategy only in this category, as showing that similar ornamental designs have been registered as trademarks can rebut the examiner's characterization.
Section 2(e)(4) -- Primarily a Surname (42 cases, 61% argue): Under TMEP 1211, the test is whether the primary significance of the mark to the purchasing public is that of a surname. The only refusal type where secondary meaning is the leading argue strategy (19 uses). This is the textbook application of acquired distinctiveness evidence -- proving that consumers associate the surname with the applicant's goods or services through years of use, advertising, and sales.
Section 2(e)(5) -- Functional Matter (31 cases, 67% argue): TMEP 1202.02 addresses functionality, where a feature is essential to the use or purpose of the article or affects the cost or quality. Mark distinction leads (17 uses), often arguing that the design has non-functional elements that serve as source identifiers independent of any utilitarian purpose.

These lower-volume categories share a common thread: when the refusal challenges what your mark is (scandalous, ornamental, a surname, functional), substantive argumentation is almost always required. The comply strategies play a supporting role at best.

A Note on the Supplemental Register

When all substantive arguments fail, some applicants choose the Supplemental Register as a fallback. This provides some trademark protection but without the full benefits of the Principal Register -- no presumption of validity, no right to the registered trademark symbol with Principal Register standing, and no constructive notice. In our data, 1,956 cases took this path. We excluded these outcomes from the winning dataset because a Supplemental Register outcome typically represents the applicant conceding the substance of the refusal rather than overcoming it.

What This Means for Your Next Office Action

The data points to three actionable principles:

First, lead with commercial impression and mark distinction. These two strategies account for the vast majority of successful argue arguments across nearly every refusal type. They work because they address the core question in most refusals: is your mark sufficiently distinct?

Second, don't neglect compliance. Goods amendment is the single most common element in winning responses, appearing 10,024 times across the dataset. Even when you plan to argue, a strategic goods amendment can narrow the dispute and strengthen your position.

Third, focus your response. The 3-5 argument sweet spot is not arbitrary. It reflects the reality that examiners are more persuaded by a few well-supported arguments than by a scattershot approach. Identify your strongest two or three points, develop them thoroughly, and let the weaker arguments go.

GleanMark's office action response tools can help you identify which strategies are most effective for your specific refusal type and generate a response grounded in the patterns from this dataset. For a deeper dive into any specific refusal category, explore the dedicated strategy guides linked above.