Ultimate DIY Guide to Nice Classification for Founders
Choosing the wrong trademark class can leave your real business unprotected
A common trademark filing mistake looks minor at the start and expensive later.
A founder files the brand for “online retail services” because the business sells through a webshop, then discovers the registration does not properly cover the products themselves. Or a SaaS company files for downloadable software in Class 9 but forgets the non-downloadable platform service in Class 42.
In both cases, the application may still register. The problem is that the protection may not match the business that actually generates revenue.
Whether you are filing nationally, regionally, or internationally, class selection is one of the first decisions that affects cost, timing, scope, and future enforcement. It is also one of the easiest places to make a DIY mistake.
This article explains how the Nice Classification system works, how founders can choose classes step by step, and where self-filing often goes wrong across the EU, UK, US, and other major markets.
Why class selection matters more than most founders expect
When you file a trademark, you do not protect the name in the abstract. You protect it for the specific goods and services listed in the application.
That list shapes several commercial outcomes:
How broad your protection is
Whether your application may face objections or oppositions
How much you pay in official fees
Whether your filing still works when the business expands
How useful the registration will be in enforcement
In plain English: class selection is not admin. It is part of brand strategy.
It also matters because missing items are not always easy to fix later. In many trademark systems, including the EU, UK, and US, you can usually narrow a specification after filing, but you generally cannot broaden it to add new goods or services that were not originally included. If a core part of the business is missing, the usual fix is a new application, new fees, and a later filing date, making the DIY strategy far more expensive.
If your expansion plan includes other markets, filing timing matters as well. We cover that broader ownership point in our writing on first-to-file and first-to-use trademark systems, because the class decision and the filing timing decision often belong in the same conversation.
What the Nice Classification system actually does
The Nice Classification is the international system used to organize the goods and services covered by a trademark application.
Classes 1 to 34 cover goods
Classes 35 to 45 cover services
The system is global. It is used across major trademark offices and filing systems, including WIPO, the EUIPO, the UKIPO, the USPTO, and many national offices worldwide.
The key point for founders is simple: the class number is a filing tool, not the whole legal answer.
Two brands can still conflict even if they are not in the exact same class number. Authorities and courts look more broadly at whether the goods or services are commercially similar and whether consumers may think they come from the same business.
For example, clothing in Class 25 and retail services for clothing in Class 35 are different classes, but they may still be commercially connected in a way that matters in a conflict analysis.
Nice is global, but filing practice is not identical everywhere
This is where founders often get tripped up. The class framework is international, but examination practice, specification drafting rules, fee structures, and use requirements are still local.
EU: founders often work with broad but clear specifications, and non-use vulnerability becomes important if the registration covers more than the business actually uses.
UK: the Nice system also applies, but founders should still check UKIPO guidance and accepted wording rather than assuming an EU-focused draft will always transfer cleanly.
US: the Nice classes are used, but filing practice is more use-driven. Applicants often need to think carefully about filing basis, identification wording, and proof-of-use requirements.
Other markets: many countries follow Nice closely, but local offices can still differ on acceptable wording, subclass practices, translation issues, and procedural formalities.
The practical takeaway is simple: do not assume one specification copied from another filing will work equally well in every jurisdiction.
Founders should protect what they sell now and what they will realistically launch next
Before opening a class database, write down the business in plain language.
What do customers pay you for today?
What do you deliver yourself?
What is provided through software?
What is downloadable, if anything?
What do customers access online without downloading?
What will you realistically launch in the next 24-36 months?
This sounds basic, but it is where many filings go off track. Founders often think in brand or product language. Trademark offices need you to translate that into goods and services.
A useful rule of thumb is this: focus mainly on current offerings and near-term launches. Very remote ideas can make the filing broader, more expensive, and harder to defend later.
This matters internationally, but in slightly different ways. In the EU and UK, claiming far more than you use can create later vulnerability. In the US, over-claiming can also become problematic because use and proof of use play such an important role in the filing lifecycle.
How to choose trademark classes step by step
1. List your real-world offerings
Create three simple buckets:
Current offerings
Near-term launches
Possible future ideas
Your filing will usually be built mainly around the first two.
Example:
Today: downloadable fitness app
Next 12 months: subscription coaching platform
Possible later: branded supplements and workshops
This already tells you that one Nice class is unlikely to be enough.
2. Separate goods from services
This is one of the most common DIY failure points.
Goods are usually physical products or downloadable digital products. Services are activities you perform for customers, including SaaS access, consulting, training, retail, design, or maintenance.
Examples:
Downloadable mobile app may fall in Class 9
Non-downloadable software as a service may fall in Class 42
Business consultancy may fall in Class 35
Training courses or workshops may fall in Class 41
Modern startups often need more than one class because they combine software, services, education, content, and commerce under one brand.
3. Check official accepted terminology
Do not rely only on your own wording. Trademark offices care about whether the goods and services are described clearly enough.
Start with official tools relevant to where you plan to file:
Using accepted wording can reduce the risk of clarity objections and classification issues.
For example, “technology services” is often too vague to be commercially useful. “Software as a service featuring budgeting tools” is much clearer.
4. Map each offering to the most accurate class
Do this line by line rather than trying to guess the whole filing in one go.
Example for a fintech startup:
Downloadable budgeting app - Class 9
Non-downloadable financial planning software - Class 42
Business advisory services in finance operations - Class 35
Financial services - Class 36
Online training in financial literacy - Class 41
The goal is not to collect classes. The goal is to match the filing to the actual commercial offer.
5. Draft a sensible specification
The specification is the list of goods and services that appears in the application.
Good specifications are:
Clear enough that someone can understand what the business offers
Broad enough to cover realistic growth
Consistent with accepted office terminology
Limited to items you genuinely plan to use or file for
Weak specifications are often copied from another company, padded with irrelevant items, or written so vaguely that they create more comfort than real protection.
If you remove the brand name and read only the specification, it should still sound like a coherent business.
If you are filing in several countries, this is the stage where local differences begin to matter more. A phrase that is acceptable in one office may be questioned, narrowed, or rejected in another.
6. Pressure-test the filing before submission
Ask practical questions:
If we launch our core offer next month, is it covered?
If we slightly pivot, is the business model still covered?
Are we paying for classes we are unlikely to use?
Would an investor, acquirer, or future legal reviewer spot obvious gaps?
If enforcement becomes necessary, does this filing actually support the argument we would want to make?
Does this wording work in each jurisdiction where we plan to file?
This last point matters. A registration is much more useful when it clearly matches the products or services you are trying to defend in the market.
Common class patterns founders often need
These are examples, not shortcuts. Always match the filing to your specific business model and target jurisdictions.
SaaS startup
Class 9 for downloadable software or apps (e.g. iOS and/or Android), if offered
Class 42 for software as a service or platform as a service
Class 35 for business analytics or advisory services, if those are real offerings
Class 41 for training, webinars, or educational services, if monetized or genuinely provided
A frequent mistake is filing only Class 9 because the company thinks “we are a software business.” If the main product is accessed online, Class 42 may be central.
E-commerce brand
The relevant product classes for the goods sold
Class 35 for online retail services, where appropriate
A classic mistake is filing only the retail service and forgetting the products themselves.
Consumer app brand
Class 9 for the downloadable app
Class 42 for non-downloadable software or backend platform services
Class 38 only if telecommunications features are genuinely part of the offer
Founders sometimes add Class 38 because the app has messaging features. That can be right in some cases, but not every app needs it.
Consultancy or agency
Class 35 for business consulting, commercial strategy, marketing, or similar services
Class 42 for technical consultancy or software-related services
Class 41 for training, workshops, or courses
The right mix depends on what clients are actually buying from you.
DIY mistakes that often lead to refusal, weak protection, or re-filing
Filing in the wrong class because it “sounds right”
Intuition is not enough. A founder may think “we are in tech, so Class 9,” while the real revenue comes from subscription access to non-downloadable software in Class 42.
Practical consequence: the registration may not line up with the core offer you are trying to protect.
Protecting only one part of the business model
A DTC brand may file for online retail services but not for the cosmetics, clothing, or supplements it actually sells.
Practical consequence: the business can end up with a gap exactly where the product sits.
Using wording that is too vague
Terms like “digital solutions,” “innovation services,” or “technology platform” may sound broad, but they often do little useful work in a trademark filing.
Practical consequence: objections, delays, or a registration that looks broad but is harder to rely on.
Trying to cover everything
Founders sometimes copy a long specification from a larger company and file across many classes “just in case.”
Practical consequence: more fees, more complexity, and possible future vulnerability if important parts are never used.
Assuming class numbers decide conflicts
They do not. Similarity analysis is broader than whether two businesses filed in the same class number.
Practical consequence: false confidence during brand selection and clearance.
This is also why class selection should not replace a proper conflict check. A clearance review means checking whether earlier trademarks may block your filing or create enforcement risk later.
Assuming one jurisdiction’s wording will work everywhere
A founder may prepare a specification with EU wording, then assume the same text will file cleanly in the UK, US, and elsewhere.
Practical consequence: office objections, forced amendments, inconsistent coverage, or extra filing work across markets.
Ignoring realistic expansion
If your roadmap clearly includes training, downloadable tools, or branded goods in the near term, leaving them out may create avoidable re-filing work (and costs).
Practical consequence: fragmented protection and later filings with later priority dates.
Copying another company’s specification
Another brand’s filing may reflect a different business model, an older filing strategy, or advice tied to a different jurisdiction.
Practical consequence: you inherit their wording without inheriting their reasoning.
AI can help with the first draft, but it should not be the final classification decision
AI tools can be useful for turning a product list into draft goods and services wording or for spotting likely class options.
But AI can also flatten important distinctions, especially between downloadable goods, SaaS services, consultancy, education, and retail activity. It can also miss jurisdiction-specific drafting norms.
A practical workflow is:
Use AI to describe the business consistently
Use official databases such as TMclass, the USPTO ID Manual, UKIPO guidance, and PRV guidance to validate the wording
Use human review before filing if the business model is layered, the filing is multi-country, or the brand matters commercially
That approach is usually more reliable than treating automated suggestions as filing-ready.
When DIY is usually workable and when review is worth the cost
DIY is more realistic when
You have one clear product or one clear service
The filing sits in a small number of classes
Your wording matches accepted terminology cleanly
The commercial risk of getting it wrong is still relatively low
You are filing in one market with relatively straightforward facts
Professional review is usually worth it when
Your startup combines software, services, education, and content
You are filing in several countries or planning fast expansion
You are investing materially in brand launch, packaging, or paid acquisition
You already see similar marks in your market
Your roadmap matters to how the specification should be framed
You want stronger enforcement options later
If copycats, marketplace listings, or brand language online are part of the risk picture, it is also worth reading our analysis of trademark strategy around “dupe” language and online enforcement. Class choices and enforcement choices are often more connected than founders expect.
A practical checklist before you file
Have we described the business in plain English first?
Have we separated current offers from remote future ideas?
Have we split goods from services?
Do customers download anything?
Do customers access software online without downloading?
Do we also provide consulting, training, support, or retail services?
Have we checked official accepted terminology for each target office?
Are all core revenue lines covered?
Are we paying for classes we probably will not use?
Would missing something important force a new filing later?
Are we assuming a global filing strategy when we actually need jurisdiction-specific drafting?
Useful official resources for Nice Classification and major filing markets
The commercial takeaway for founders
The Nice Classification system is global and manageable for many founders, but only if you treat it as a business decision rather than a formality.
The aim is not to claim as many classes as possible. The aim is to build a filing that fits the business you actually run, the offers you are likely to launch next, the markets where you plan to operate, and the protection you may need if a conflict appears.
If the brand is important, the business model is layered, or the expansion plan crosses borders, class selection is one of those small-looking decisions that can become expensive later.
This article is for general information only and not legal advice. If you want a second pair of eyes on class selection, specification drafting, or a cross-border filing strategy, Abrande can help review the filing before submission.
