# How to classify a trademark: 45 Nice classes and the 4-part distinctiveness gradient

Ryan Walker · November 17, 2026

> Classify trademarks using the USPTO's 45 Nice classes and measure strength on the 4-part distinctiveness gradient, from generic to fanciful.

| Takeaway | Detail |
| --- | --- |
| The USPTO uses 45 Nice classes, not 5 mark types | Goods and services are classified into 45 Nice classes; there is no USPTO system of five mark types. |
| Mark strength is measured on a 4-part distinctiveness gradient | Marks are evaluated as generic, descriptive, suggestive, or arbitrary/fanciful. |
| Generic marks are not registrable | If your mark is generic, stop before filing because it cannot function as a trademark. |
| Descriptive marks require acquired-distinctiveness evidence or a reframe | If your mark is descriptive, plan acquired-distinctiveness evidence or reframe it to be suggestive before selecting Nice class(es). |

The USPTO classifies goods and services into 45 Nice classes and evaluates mark strength on a 4-part distinctiveness gradient, not five mark types.

Filing correctly depends on placing your mark on that gradient first—stopping if generic or preparing acquired-distinctiveness evidence if descriptive—before selecting the right Nice class(es).

![How to classify a trademark](https://static.mm-ais.com/article-images-ai/how-to-classify-a-trademark-45-nice-clas-ai-79ab58ab.jpg)

## Two axes: mark type vs. Nice class

Most filing mistakes start with a single bad assumption: that “mark type” and “class” are the same choice. They are two separate axes, and this section is where I separate them cleanly. The USPTO does not sort applications into a short menu of “5 types” of marks for strength purposes. It examines how distinctive the wording or design is on a four-part gradient, and it separately asks which Nice class or classes cover what you actually sell. A complete filing has to satisfy both axes at once.

Start with the distinctiveness gradient. In the “Easily Register a Trademark” framework from Udemy, mark strength runs from generic, to descriptive, to suggestive, to arbitrary or fanciful. Generic terms are common, everyday names for the thing itself and cannot function as a source identifier, so if your proposed name lands there, stop and pick another. Descriptive terms need a different plan: either build evidence that buyers have come to recognize the term as yours, or reframe it into something suggestive. Suggestive and arbitrary or fanciful marks sit on the stronger end because they require a step of imagination rather than just naming the product.

Only after you have placed the mark on that gradient should you turn to the Nice side. The Nice Classification, explained by Trademark Factory, contains 45 classes: 34 for goods and 11 for services. That structure is not a US-only quirk. StartupFino notes that India’s Trademark Act uses the same 45-class structure, which is a useful confirmation of how widely the Nice framework has converged internationally. The class list tells the Office what category your goods or services fall into; it says nothing about whether your brand name is inherently distinctive.

Here is the practical check I use before filing: write one line for the gradient result and one line for the class result. For the first line, ask what the term means to a buyer who has never seen your brand. If it answers “what the product is,” you are in generic or descriptive territory. For the second line, ask what the buyer is actually receiving from you — a physical item, a downloadable item, or an activity performed for them — and match that to the goods or services side of the Nice list. Keeping those two lines separate prevents the most common conflation: treating a strong-sounding name as a substitute for a correct class, or treating a correct class as a cure for a weak name.

To verify the class side, search the USPTO Trademark ID Manual rather than relying on memory or a generic category label. Enter your proposed goods or services wording, review the returned identifications, and use the wording that matches your real offering. Then go back to the gradient line and confirm you are not relying on a generic term just because the ID Manual gave you a clean class match. The two checks are independent, and a clean result on one does not carry over to the other.

![Two axes: mark type vs. Nice class — How to classify a trademark](https://static.mm-ais.com/article-images-ai/how-to-classify-a-trademark-45-nice-clas-ai-a2ec9bbf.jpg)

## The 5 filing types USPTO actually offers

When filers talk about “types” of trademarks, they often mean the choices made at intake — but there is no official USPTO menu of five mark types, and this article’s sources do not enumerate one. What intake actually involves is a set of separate decisions: which filing option and basis to select, and how to present the mark’s format. Treat any tidy “five types” list you encounter as a convenience label, and verify the current intake options directly on uspto.gov before filing so your application matches how you actually use, or plan to use, your mark.

Mark format determines the scope of what you protect. A standard character mark claims the wording itself, regardless of font, color, or stylization, which generally gives the broadest coverage for future branding changes. By contrast, a design or stylized mark protects a specific visual presentation, and nontraditional formats such as sound, color, or trade dress protect distinctive sensory or product-configuration features. If your brand identity may evolve, filing the words in standard characters can preserve flexibility that a single logo filing does not.

| Mark format | What it protects | Filing check |
| --- | --- | --- |
| Standard character (word) | The wording in any font, size, or color | Best when you want broad protection for the name itself |
| Design or stylized | A specific logo, font treatment, or visual arrangement | Use when the visual appearance is central to the brand |
| Sound, color, or trade dress | A distinctive sound, color claim, or product/packaging look | Be prepared to clearly depict the mark and show how consumers recognize it as a source indicator |

The filing basis is a separate decision, and the available sources do not detail the basis options, so treat the following as a check rather than a rule: confirm with the USPTO whether your mark is already in use in commerce (which typically requires a specimen at filing) or whether you are filing before use (which typically adds a later step and fee before registration). Because any later step carries its own government fee, verify the current amounts on the USPTO fee schedule page at uspto.gov before filing.

Finally, choose the filing form that fits your application. The available sources do not specify the current form options or their requirements, so check uspto.gov directly for which form restricts you to pre-approved ID Manual identifications and which allows custom wording. Confirm the current form requirements and filing options with the USPTO so your intake choices align with your evidence, timeline, and mark format.

## Class 9 vs 42: the software trap

Once a name is ready for filing, software brands often stumble on a different question: which Nice class actually covers the product? UpCounsel’s Class 9 guide identifies the split between Class 9 (downloadable software goods) and Class 42 (software-as-a-service services) as the most common misclassification. The same product name can require protection in more than one class because trademark rights attach to how customers receive and use what you offer, not just to the brand name itself.

According to UpCounsel, Class 9 covers downloadable software, electronics, and scientific goods, while Class 42 covers software-as-a-service, hosting, and design services. A practical test resolves most cases: if users download and install a copy they control on their own device, the offering leans toward Class 9; if users access the functionality over the cloud, typically by subscription or login, it leans toward Class 42. When a product does both—such as a downloadable application paired with a cloud-based platform—filing in both classes is often necessary to reflect the full scope of use.

| Class | Covers | Quick check |
| --- | --- | --- |
| Class 9 | Downloadable software, electronics, and scientific goods | Users download, install, or own a copy on their device |
| Class 42 | Software-as-a-service, hosting, and design services | Users access the software remotely via the cloud or a subscription |
| Class 41 | Education and entertainment services, including training and sporting activities | The core offering is instruction, courses, or entertainment rather than software access |

Class 41 becomes relevant when a software platform also delivers education or entertainment as a service, such as live training, online courses, or coaching. In those cases, the software delivery may point to Class 9 or Class 42, but the instructional or entertainment component can call for a separate Class 41 filing under the Nice Classification framework. Mapping each revenue-generating activity to its corresponding service or goods category helps prevent gaps in coverage.

Before filing, document how customers interact with your product today and how you plan to deliver it next. List whether you distribute a downloadable program, provide cloud access, host data, or offer training, then match each mode to the appropriate class. This delivery-based check, grounded in UpCounsel’s guidance, keeps the application aligned with actual use and reduces the risk of protecting the right name in the wrong class.

## Options compared: DIY vs attorney

After you have placed your mark on the distinctiveness gradient and identified the right Nice class(es), the remaining decision is execution: who prepares and files the application. The four practical filing routes are self-filing with a pre-approved identification from the USPTO ID Manual, self-filing with a custom identification, using an AI classification tool as a first pass, and hiring a trademark attorney. Each route trades cost, control, and drafting risk differently, and the best choice depends on your filer profile.

**Route 1 — Self-file with an ID Manual pre-approved identification:** If your goods or services match an existing entry in the USPTO’s ID Manual, this is the cheapest route and carries the lowest refusal risk tied to identification wording. The examiner’s concerns about clarity and classification are reduced because the language is already accepted. **Route 2 — Self-file with a custom identification:** When no Manual wording fits, you must draft your own description. This route is necessary for novel or narrowly defined offerings, but it requires more drafting skill; vague or overly broad wording can trigger objections and delay prosecution.

**Route 3 — AI classification tool:** Classifiers in the style of trademarkclassification.app can provide a useful first-pass suggestion of likely classes and phrasing. Treat the output as a research aid, not a filing decision: verify every suggested class and description against the current USPTO ID Manual before you submit. **Route 4 — Attorney drafting:** Professional drafting is justified for multi-class filings, portfolios, or situations where the description must be carefully scoped to match commercial use without creating avoidable limitations or refusals.

| Filer profile | Recommended route | Why it wins |
| --- | --- | --- |
| Single class, exact ID Manual wording available | Route 1: ID-Manual-only DIY | Lowest cost and lowest identification-related refusal risk. |
| Single class, no exact Manual match, straightforward offering | Route 2: Custom-identification DIY | Avoids attorney fees when you can draft a clear, accurate description and accept the added review burden. |
| Uncertain scope or exploring possible classes | Route 3: AI tool, then verify in ID Manual | Speeds up initial research, but only wins when every suggestion is manually confirmed before filing. |
| Multi-class, complex, or high-stakes brand launch | Route 4: Attorney | Strategic drafting and coordination across classes reduce costly mistakes that are harder to fix after filing. |

Whichever route you choose, run the same final check: confirm that each identification appears in, or aligns closely with, the USPTO ID Manual and that the selected class reflects how the mark is actually used. A cheaper route is only a win if it produces a clear, defensible application.

## Costs and numbers that matter

Budgeting starts with a simple mechanical fact: USPTO government fees are charged per class, not per application. If you file one mark across two Nice classes, the government fee doubles. That is true even when the mark is identical and the application arrives as a single filing. Before you set a budget, go to uspto.gov and verify the current per-class TEAS Plus and TEAS Standard amounts on the official fee schedule. I am deliberately not printing dollar figures here, because fee amounts are volatile and the only reliable number is the one on the USPTO schedule on the day you file. The check is mechanical: count your classes, multiply by the current per-class amount for the filing option you selected, and only then add any professional or search costs.

The cost asymmetry is the part most filers miss. A refused class and an over-filed class hurt in different ways. If you misclassify goods or services, the problem usually surfaces as an examiner amendment or refusal, and you pay for a response cycle: drafting time, additional filing-related expense, and delay while the application sits unresolved. If you over-file, you pay the government fee for a class you may never actually use, and you still carry the burden of proving use in that class to maintain the registration later. One error buys you a correction cycle; the other buys you an ongoing obligation. Neither is free, and the second can keep costing after registration.

That asymmetry is why class selection deserves the same discipline you apply to the mark itself. As detailed above, the Nice system’s 45-class structure (34 goods, 11 services, per Trademark Factory) defines what the protection attaches to, and FasterCapital frames class choice as one of the important steps in registration. Treat each class as a separate commitment: identify the actual goods or services, match them to the correct class language, and confirm that you are prepared to use the mark in that class — not merely that the class sounds adjacent to your business.

Here is the verification method I use before budgeting. First, list every good or service the mark will be used on now or in the near term. Second, assign each item to a Nice class and check the identification wording against the USPTO’s accepted descriptions. Third, count the distinct classes, not the number of products. Fourth, pull the current per-class TEAS Plus and TEAS Standard amounts from the USPTO fee schedule and multiply by that class count. Fifth, compare the result against your actual launch plan. If a class exists only because you thought it might be useful someday, remove it from the filing and revisit it when real use exists.

The practical rule is to spend money on precision, not coverage theater. A narrower filing with correctly matched classes is easier to price, easier to prosecute, and easier to maintain than a broad filing padded with speculative classes. A broader filing is not automatically better protection; it is simply more classes, more fees, and more use obligations. Verify the fee, count the classes, and make each class earn its place before the application goes in.

## What the evidence does NOT establish

This guide is built to help you make two filing decisions in the right order, but it does not establish every fact you may need before you submit an application. The most important limit is definitional: the “5 USPTO types” framing used in headlines is editorial shorthand, not a statutory USPTO category. Treat it as a convenience label rather than a legal classification, and rely on the distinctiveness analysis and Nice class selection described elsewhere in this guide as the operative framework.

No current USPTO fee amounts appear in the available sources for this article. Because filing, per-class, and maintenance costs can change, every dollar figure you encounter here or on third-party sites must be verified against the live USPTO fee schedule before you budget or file. Do not rely on cached blog posts, calculators, or marketplace listings for pricing; confirm the amount, the filing option it applies to, and when it takes effect directly with the USPTO.

The available sources also provides no data on refusal rates by class or by mark type. Any claim that a particular Nice class is “harder,” “more crowded,” or “faces more refusals” is unverified in this guide. To assess difficulty for your own mark, search the USPTO’s Trademark Status and Document Retrieval (TSDR) system and the USPTO Trademark Search database for your specific class and comparable wording or designs. Use those records to evaluate actual office actions and outcomes rather than relying on generalized assumptions.

These limits do not change the filing sequence set out above: place your mark on the distinctiveness gradient first, and if it is generic, stop; if it is descriptive, plan acquired-distinctiveness evidence or consider a suggestive reframe before you invest in filing. Only then select the appropriate Nice class or classes, as covered in the two-axes section. Treat class selection as a factual match between your goods or services and the official classifications, not as a proxy for approval odds.

Before you file, run a final verification pass: confirm current fees on the USPTO fee schedule, review live TSDR records for marks in your class, and document why your chosen class accurately covers your offering. That closes out the checks this guide has already walked through — gradient placement, ID Manual matching, and per-class fee counting — and keeps your filing decisions grounded in verifiable evidence.

## + filing worksheet

This section alone walks a concrete SaaS-plus-app product through both axes and delivers the copy-usable classification worksheet. Take “DataLens”: a downloadable analytics app, a subscription dashboard, and free training videos. The name is coined, so it sits at the arbitrary end of the distinctiveness gradient described in Udemy’s trademark course materials, which place mark strength on a four-part run from generic to descriptive to suggestive to arbitrary/fanciful. Because “DataLens” does not merely describe analytics software, the gradient check comes back clean and there is no descriptiveness problem to cure before class selection begins.

Now run the class checks in order. First, search the USPTO ID Manual for “downloadable analytics software” and confirm the Class 9 wording exists for the downloadable app. Class 9 is the goods-side home for software and electronics, as UpCounsel’s Class 9 guide notes. Second, search “software as a service featuring analytics” and confirm the Class 42 wording for the subscription dashboard. Third, search the ID Manual for the free training videos and treat Class 41 as a candidate only if the videos are a real service you provide, not a throwaway marketing extra; the Nice framework divides 45 classes into goods and services, with Class 41 covering education and training services, per Trademark Factory’s explanation of the Nice Agreement.

The gradient check is the gate, not the class list. If the name had been something like “FastAna,” you would stop before filing and either assemble acquired-distinctiveness evidence or reframe toward a suggestive name, because a descriptive term on the gradient needs a different plan than a coined one. FasterCapital’s classification introduction makes the same practical point from the other direction: choosing the right classes is a core filing step, but it only matters once the mark itself is worth pursuing. 99Logos likewise separates screening from registration — screening tells you whether the name is worth taking forward; registration is the formal process that follows.

Use this worksheet as your pre-filing record, one row per offering:

| Offering | Gradient result | ID Manual query | Class candidate | Evidence to keep |
| --- | --- | --- | --- | --- |
| Downloadable analytics app | Arbitrary/coined — passes | “downloadable analytics software” | Class 9 (goods) | Confirmed ID Manual wording |
| Subscription analytics dashboard | Arbitrary/coined — passes | “software as a service featuring analytics” | Class 42 (services) | Confirmed ID Manual wording |
| Free training videos | Arbitrary/coined — passes | Training/education wording for the videos | Class 41 candidate | Proof the videos are an actual service |

Fill the table before you touch the application form. If a row’s gradient result is generic, stop that row; if it is descriptive, attach the acquired-distinctiveness or reframe plan; if it is suggestive or arbitrary/fanciful, move to the class query. Only then carry the confirmed class wording forward. That order — mark strength first, Nice class second — is what keeps a product like DataLens from being filed against the wrong axis.

## What to do next

| Step | Action | Why it matters |
| --- | --- | --- |
| 1 | Place your mark on the 4-part distinctiveness gradient before touching any class list: generic, descriptive, suggestive, or arbitrary/fanciful. | The USPTO evaluates mark strength on that gradient, not on five mark types. |
| 2 | If your mark lands on generic, stop before filing. | A generic mark cannot function as a trademark and is not registrable. |
| 3 | If your mark lands on descriptive, either assemble acquired-distinctiveness evidence or reframe it to be suggestive before going further. | Descriptive marks need that evidence or a suggestive reframe to clear the gradient. |
| 4 | Only after the gradient check, open the USPTO ID Manual and select your Nice class(es) from the 45 Nice classes. | Goods and services are classified into 45 Nice classes; the gradient comes first, never the reverse. |
| 5 | Confirm the selected class(es) cover the actual goods and services in your filing, then file right. | Filing correctly depends on matching the ID Manual class(es) to the real goods and services after the strength check. |

## Frequently Asked Questions

**Can a generic mark be registered if it is filed under the correct Nice class?**

No, if your mark is generic, you should stop before filing because it cannot function as a trademark.

**What options are available if a mark is determined to be descriptive?**

If your mark is descriptive, you must plan acquired-distinctiveness evidence or reframe it to be suggestive before selecting Nice class(es).

**What are the four categories used to evaluate mark strength on the distinctiveness gradient?**

Marks are evaluated as generic, descriptive, suggestive, or arbitrary/fanciful on the 4-part distinctiveness gradient.

**Are mark type and Nice class the same consideration in a trademark application?**

No, mark type and Nice class are two separate axes, and confusing them is a common source of filing mistakes.

**Does the USPTO use a system of five mark types to assess trademark strength?**

No, the USPTO does not sort applications into a short menu of five mark types for strength purposes but instead examines distinctiveness on the 4-part distinctiveness gradient.

**At what point in the filing process should the appropriate Nice class(es) be selected?**

The appropriate Nice class(es) should be selected only after placing your mark on the 4-part distinctiveness gradient and addressing any issues if it is generic or descriptive.

## Quick answers

| How many Nice classes does the USPTO use to classify goods and services? | The USPTO classifies goods and services into 45 Nice classes. |
| --- | --- |
| What are the four parts of the distinctiveness gradient used to evaluate mark strength? | Marks are evaluated as generic, descriptive, suggestive, or arbitrary/fanciful. |
| Can a generic mark be registered as a trademark? | No, if your mark is generic, it cannot function as a trademark and is not registrable. |
| What should be done if a mark is determined to be descriptive? | If your mark is descriptive, you should plan acquired-distinctiveness evidence or reframe it to be suggestive before selecting Nice class(es). |
| Are mark type and Nice class the same choice when filing a trademark? | No, mark type and class are two separate axes, and the USPTO does not sort applications into a short menu of five types of marks for strength purposes. |

Also worth reading: **2024 USPTO Trademark Filing Fees A Detailed Cost Breakdown for Logo and Business Name Registration**: [2024 USPTO Trademark Filing Fees](https://aitrademarkreview.com/blog/2024_uspto_trademark_filing_fees_a_detailed_cost_breakdown_f.php) · **2024 USPTO Trademark Registration Costs A Detailed Breakdown by Filing Class and Application Type**: [2024 USPTO Trademark Registration Costs](https://aitrademarkreview.com/blog/2024_uspto_trademark_registration_costs_a_detailed_breakdown.php) · **TEAS Plus vs

TEAS Standard A Detailed Comparison of USPTO's 2024 Trademark Filing Options**: [TEAS Plus vs TEAS Standard](https://aitrademarkreview.com/blog/teas_plus_vs_teas_standard_a_detailed_comparison_of_uspto_s.php)

### Related reading

- [Decoding USPTO's 45 International Classes A Deep Dive into Trademark Categorization](https://aitrademarkreview.com/blog/decoding_uspto_s_45_international_classes_a_deep_dive_into_t.php)
- [Navigating the Complexities of Trademark Classes A Guide to Efficient Searching in 2024](https://aitrademarkreview.com/blog/navigating_the_complexities_of_trademark_classes_a_guide_to.php)
- [Distinctiveness Matters Protecting Sweet Brand Trademarks](https://aitrademarkreview.com/blog/distinctiveness_matters_protecting_sweet_brand_trademarks.php)
- [Trademark search false positives: verify USPTO query syntax before filing](https://aitrademarkreview.com/blog/trademark-search-false-positives-verify-uspto-query-syntax-before-filing.php)
- [Trademark Filing Extensions: How a $125 Fee Changes Multi-Class AI Filings](https://aitrademarkreview.com/blog/trademark-filing-extensions-how-a-125-fee-changes-multi-class-ai-filings.php)
- [AI trademark similarity scores: screening aid, not a prediction](https://aitrademarkreview.com/blog/ai-trademark-similarity-scores-screening-aid-not-a-prediction.php)

### Latest

- [Trademark search false positives: verify USPTO query syntax before filing](https://aitrademarkreview.com/blog/trademark-search-false-positives-verify-uspto-query-syntax-before-filing.php)
- [Embedding trademark searches 2026: verify the 40% claim before filing](https://aitrademarkreview.com/blog/embedding-trademark-searches-2026-verify-the-40-claim-before-filing.php)
- [Trademark Filing Extensions: How a $125 Fee Changes Multi-Class AI Filings](https://aitrademarkreview.com/blog/trademark-filing-extensions-how-a-125-fee-changes-multi-class-ai-filings.php)

Canonical: https://aitrademarkreview.com/blog/how-to-classify-a-trademark-45-nice-classes-and-the-4-part-distinctiveness-gradient.php
Markdown: https://aitrademarkreview.com/blog/how-to-classify-a-trademark-45-nice-classes-and-the-4-part-distinctiveness-gradient.php/index.md
