Lookalike onchain TLDs: which variants of your name matter

The version of your brand name that sits at the top level of a namespace is not interchangeable with a second-level domain. Whoever holds .yourbrand holds every name that could ever be issued under it: pay.yourbrand, login.yourbrand, id.yourbrand. The whole address space beneath that string follows the TLD. One string carries an entire namespace, and that is why the variant question feels sharper onchain than it does in the conventional DNS, and why the instincts an IP team has built around domain monitoring transfer only in part.

This article gives you a method for building a variant list, cutting it to a manageable size, and deciding what to do with what remains. It does not tell you to hold every plausible form of your name. It tells you which forms are worth your attention, and why the discipline of choosing matters as much as the list itself.

The method borrows from the way a defensive domain portfolio is run, with two adjustments. The unit you assess is a whole TLD, not a single domain. And on a registry without renewal fees, nothing forces you to prune, so the pruning has to be a decision you take on purpose.

Why the unit of exposure is the whole namespace

In the conventional DNS, a lookalike registration concerns one domain under an existing extension. The exposure is real but bounded, and an established procedure exists to deal with it. Onchain, the holder of a TLD controls a namespace: only the holder can issue names under that string, and the holder decides how those names are configured where they resolve. On Freename, the holder can also receive a share of each registration fee for names registered under the TLD, under the registry’s royalty program and subject to its current terms and activation conditions. A single string therefore carries every name beneath it.

For an audit, this changes what you count. A domain watch report counts individual registrations, often in large numbers, most of them harmless. A TLD-level review counts far fewer strings, but each one stands for an open-ended set of names that you cannot enumerate in advance. You are not asking “is this one address confusing?” You are asking “if names were issued under this string, could one of them be taken for us in a setting where it matters?”

The setting that matters most is a transfer of value. A name is a readable label for a wallet address. When a payer enters a name as the recipient, the funds go to the address that name points to, and an onchain payment is generally not reversible by an intermediary the way a card payment can be. Where the TLD is not yours, the name under it is not yours either. This is the scenario a variant review exists to recognize early.

The exposure is also impossible to close by registration alone. No organization can hold every combination, script and spelling of its name, across every naming system. That is the honest starting point for a variant analysis: the goal is not completeness, it is triage.

The families of variants: a reference map

Before ranking anything, list the families. The exposure differs from one family to the next, and the families overlap at the edges, so it helps to hold them in one place.

Exact name

The TLD that matches your primary mark exactly is the anchor. Everything else is measured against it. If you do not hold it, the analysis starts there, and the page on trademark TLDs explains what such a TLD is and includes a free Trademark Name Check.

Group and holding names

Many organizations trade under a consumer-facing mark and contract under a group or parent name. The group form appears in corporate filings, in procurement and on invoices. That is why it deserves a place on the list even though customers rarely say it: the people who do read it are finance teams and suppliers, and the transfers they handle tend to be the larger ones. Only the forms your organization really uses belong here. A “holdings” or “international” suffix that appears in none of your documents is a much weaker candidate than the legal name printed on your invoices.

Plurals

Whether a plural matters depends on how people refer to you. Some brands are routinely spoken of in the plural, for instance when the name is also the name of the product people buy. Others are never pluralized. Check customer correspondence and support logs before deciding; do not assume.

Hyphenated forms

A hyphenated form is visually very close to the mark, closer than most misspellings. In markets where compound names are commonly written with a hyphen, it can be the form a customer expects. In markets where they are not, it is simply a near neighbor.

Common misspellings

Your own data is the source here. Support tickets, search logs on your site and mistyped addresses in incoming mail show which errors people really make with your name. Record the few that recur and leave the rest. At TLD level this list is shorter than in a domain review, because there is only one label to get wrong.

Product and sub-brand names

A product with recognition of its own is a name customers may look for without thinking of the parent brand. Each such product is a separate entry on the list, assessed as its own mark. A product that exists only as a line in a catalog under the parent name usually is not.

Transliterations

Where you operate in a market that uses another script, the name as written in that script is a distinct string. It matters most where customers in that market meet the local form first. Your trademark filing strategy already answers most of this: the scripts in which you have filed marks are the scripts to consider. Your own counsel will know which are material.

How customers meet a name onchain, and why that shifts the risk map

People do not read a familiar name character by character. They recognize its shape. This is well understood from conventional fraud prevention, and it applies onchain with one difference: the settings in which a name is seen are fewer and more specific, which makes them easier to audit.

A name reaches a person in three main ways.

As a payment recipient. A payer types or pastes a name into the recipient field of a wallet, and the wallet resolves it to an address. The name is often the only thing the payer checks.

As an identity. An application shows a name next to an account, a message or an AI agent, and other parties rely on it to know whom they are dealing with.

As a website address. A name leads to pages published under it, in browsers that support the registry.

Support is partial in all three and it changes over time. An ordinary browser does not open an onchain name without added support, and not every wallet resolves every registry. The overview of what works lists each capability with its sources. This limit is part of the assessment, not a footnote: a variant can only mislead where it resolves, so the places where names of a given registry do resolve are the places to think about first.

These settings do not weight variants equally. In a payment field, a name is seen alone, without a logo, a page design or a sender to compare it with. Closeness to the mark carries most of the weight there. On a website, the surrounding content does more of the work than the string, and your existing content monitoring already covers part of the ground. As an identity label, what matters is whether the form looks like something your organization would use for a department, a product or a regional office.

Two consequences follow for the list. First, variants that are plausible on their own, with nothing around them, rank above variants that need a supporting context. Second, variants tied to something you really take payment for, such as a product with its own checkout or a group entity that issues invoices, rank above variants tied to marketing language only.

This is one of the points where the DNS is simply better equipped: a trademark owner there has a known procedure with predictable steps. Onchain, a trademark still carries its legal weight, but there is no single administrative route. That raises the value of sorting variants in advance, because acting on each one takes more effort.

The ranking method: three filters, one tier

A workable ranking asks three questions of each variant: how close it is to the mark, how plausible it is where money or identity is at stake, and whether the form is natural in a market where you operate. Each question is a filter with a yes or no answer, not a score to be averaged. Averages hide the reason a variant matters; a filter keeps it visible.

1. Closeness to the markThe variant passes if it contains the mark whole, differs from it by one or two characters, or sounds the same when spoken. A variant that passes none of these is too far from the mark to be confused with it and leaves the list here.
2. Plausibility in a payment or login contextAsk whether a person who met this string alone, in a recipient field, a sign-in prompt or an invoice line, could reasonably take it for your organization. If it only looks credible when other brand elements surround it, it does not pass.
3. Natural in a market you serveAsk whether this is a form your organization or its customers really use in at least one of your markets: a name printed on your documents, a product people ask for, a spelling convention of the local language. A form nobody uses does not pass.
4. Assign the tierThree filters passed is Tier 1. Closeness plus one of the other two is Tier 2. Closeness alone is Tier 3. Write the failed filter next to each Tier 2 and Tier 3 entry.

The order of the filters is deliberate. Closeness is cheap to judge and removes most candidates. Plausibility needs a person who knows how your customers pay and sign in. The third filter needs evidence from the business: invoices, product names, local usage. Running them in this order means the expensive questions are asked of a short list.

Answer each filter from evidence, not from imagination. The question is never “could someone conceive of a use for this string?”, because the answer to that is always yes and the list never ends. The question is “does this string correspond to something we do, somewhere we do it?”

Writing the failed filter next to each entry is what makes the list maintainable. A Tier 2 entry marked “not natural in any market” tells whoever reviews it next exactly which change in the business would move it up.

Worked example: ranking twelve variants

This is a worked example with an invented company. The assumptions are these. The company is a mid-sized payments business whose primary mark is the word “yourbrand”. It operates in three markets. In one of them, compound names are commonly written with a hyphen. It has one consumer product with its own checkout, called “yourbrandpay”. Its invoices are issued by a legal entity named “YourBrand Group”, always written without a hyphen. It has no entity or product whose name uses “corp”, “holdings” or “international”. Its customers do not refer to it in the plural. Its support logs show one recurring misspelling, with the letter “o” dropped.

Four variants reach Tier 1, four Tier 2 and four Tier 3. The even split is a coincidence of the example, not a target.

The exact mark is in Tier 1 by definition. The group name and the product name join it despite being longer strings, because each corresponds to something the company really does: one issues invoices, the other takes payments. The hyphenated form is in Tier 1 only because of the assumption about one market. Remove that assumption and it fails the third filter and drops to Tier 2. This is the filter doing its job: the same string ranks differently for two companies with different footprints.

The Tier 2 entries all look credible when seen alone, and all fail the third filter. The plural is not a form customers use. The hyphenated group name is not how the entity is written. The abbreviation belongs to informal notes, not to documents. The recurring misspelling is close enough to pass at a glance, but it is nobody’s intended form of the name, so it is watched, not acquired.

The Tier 3 entries contain the mark and so pass the first filter, but they correspond to nothing in the company. A person who knows the company would have no reason to expect a “corp”, “holdings” or “international” form. The transposed spelling looks wrong at first sight. They are recorded, with the reason, and left alone.

One limit of the example should be stated. The third market uses another script, and the company has not yet decided how its name is written there. Until it does, there is no string to rank. The list carries an open line for it, and the question goes to the team that handles trademark filings in that market.

What to do with each tier

The tier gives the kind of action, not the sequence. Sequence depends on what is held and what is not, and on your organization’s tolerance for risk.

Tier 1: acquire or watch closely

For a Tier 1 string you do not hold, the question is whether acquiring it is possible and proportionate. If nobody has registered it, you register it with the registry in the ordinary way. If it is already held, the registry record is public and anyone can check it on a block explorer, but approaching the holder is a negotiation, and usually one that is better conducted discreetly. For a Tier 1 string in someone else’s hands, acquiring the onchain TLD that matches a brand is what the brand namespace recovery service covers.

Where acquisition is not possible or not proportionate, the string goes under close watch. In practice that means someone looks at it on a set rhythm and knows what they are looking for: names issued under the TLD that refer to your brand, and signs that the TLD is being put to use.

Tier 2: watch, with a defined trigger

A Tier 2 entry needs a trigger, the event that would move it to Tier 1 handling. Typical triggers are a name under the TLD that refers to your brand or to one of your products, a customer or supplier report that mentions the string, or a change in your own business that makes the form natural, such as entering a market where the hyphenated form is standard. Without a written trigger, a Tier 2 list goes stale: nobody knows what would justify acting, so nobody acts or everybody overreacts. Namespace watch is the service for monitoring the names of a brand onchain.

Tier 3: record and leave

The record matters because circumstances change. A string that corresponds to nothing in your business may come to correspond to something if you launch a product or create an entity whose name matches it. Tier 3 is not a permanent verdict. It is a decision to set the string aside, with a note saying why.

Why trying to hold everything is a mistake

Permutation tools can produce very long lists of name variants. Such a list is raw material for the filters, not a shopping list.

In the conventional DNS, the discipline of pruning a defensive portfolio is enforced by annual renewal fees. A name that is not worth renewing is dropped, and the portfolio contracts toward what the organization values. There is no annual renewal fee on a Freename TLD, so that forcing function is absent.

The absence of a renewal cost is a real advantage for the names you need to hold. It also removes the yearly moment at which someone asks whether a name still earns its place. An organization that acquires every plausible variant has created a portfolio it must govern. Each TLD is a token held in a wallet, and each wallet needs an owner, access controls, a succession plan when people leave and a line in the asset register. Each TLD must be either put to use or deliberately left inactive, and looked at again when the brand changes. That work grows with the number of TLDs held. A short portfolio in which every entry was justified is easier to run, and easier to explain to an auditor, than a long one assembled out of caution.

There is also a proportionality argument that your counsel will recognize. Holding a TLD is not the same as preventing all use of a string onchain. It closes one avenue of confusion on one registry. If a variant fails the plausibility filter, or is natural in no market you serve, the price of acquiring it may exceed the benefit even though the price is paid once. A defensive holding should be justified by a real exposure, not by the fact that a string can exist.

Finally, a broad defensive posture sends a signal of its own. An organization known to acquire anything near its name invites offers of strings that it would never have ranked. A narrow, reasoned list is easier to defend internally and gives outsiders less to speculate on.

Before any variant is ranked, a brand’s onchain exposure can be scored at no charge with the Namespace Exposure Score tool. It is a useful first reading of how much of the list is likely to need action at all.

Keeping the list current when the brand portfolio changes

A variant list reflects the brand portfolio on the day it was built. Acquisitions, divestitures, product launches and rebrands all create new strings to assess, and can make held TLDs redundant.

The practical rule is that the variant review is a deliverable of every brand change, not an afterthought. When a new product name is approved, the matching TLD and its close variants go on the checklist at the same stage as the trademark filing and the domain registrations. When a sub-brand is retired, its TLDs move to reduced attention: still held if they were acquired, no longer generating new work, and noted as such so that nobody reopens the question by accident.

A merger or acquisition brings an incoming brand portfolio and, with it, a new set of strings. The review of the Tier 1 forms of each incoming mark belongs in due diligence, alongside the review of trademarks and domains, so that the position is known before the names of the parties are public. The page for brands and legal teams sets out how the desk works with general counsel and brand protection teams.

Changes in footprint count as much as changes in names. Entering a market with different spelling conventions or another script can move an entry from Tier 2 to Tier 1 without any new product at all. So can launching payments or sign-in under a product that used to be a marketing name only. This is where the failed filter recorded next to each entry pays for itself: the reviewer searches the list for entries that failed on “natural” and asks whether the new market or the new feature changes the answer.

The list also needs a route for signals from outside the IP team. Customer support, fraud prevention and community moderation sometimes meet a confusing string before any monitoring does. A defined path from the person who receives the signal to the person who owns the list is worth more than sophisticated monitoring without one.

Give the list a single owner and a version history. Each change records what moved, why, and which business event caused it. When colleagues who are new to the subject join the review, the glossary defines the terms, including the difference between a TLD and a name issued under it.

Getting to a first list

Start with what you already have. Take the list of registered marks, the legal names of the entities that issue invoices and the products with their own customer recognition. Apply the family table to each, then run every candidate through the three filters, writing down the failed filter each time. For a focused brand, the Tier 1 list that comes out is usually short, which is the point: it is a scope a team can act on.

Then establish the position of each Tier 1 string: held by you, held by someone else, or not registered. Where a Tier 1 TLD is held by another party and you decide to pursue it, brand namespace recovery is the route, and closing runs through onchain escrow on an aftermarket.

If you would prefer to talk the list through, the desk is reached through the contact form. One named partner follows a file from first note to closing, and every exchange is covered by a standard mutual NDA.

Questions people ask

Does holding a TLD stop similar names under a different TLD?

No. Holding .yourbrand means nobody else can issue names under that TLD. It has no effect on .yourbrands, .yourbrandpay or any other distinct string, each of which is an independent namespace. What you hold protects exactly the string you hold and nothing wider, which is why the variants are ranked one by one.

Does a trademark registration give a right to the matching onchain TLD?

Not automatically. A trademark protects a sign for given goods and services in the countries where it is registered or used. It can support a claim against a confusing use, but it does not by itself transfer a TLD that someone else holds. What can be achieved depends on the registry’s terms and on the law that applies. Your own counsel decides.

When does a product name become a priority variant?

When the product has recognition of its own, meaning customers deal with it under its own name, and when it has a payment, sign-in or identity use attached to it. A product marketed purely under the parent brand, with no transactional identity of its own, will usually fail the third filter and sit in Tier 2 or Tier 3.

Should the same variants be checked on more than one registry?

Yes, in principle. Onchain naming systems are separate from one another, and a string held on one is not thereby held on another. Apply the same tiers everywhere, but weight each registry by where its names resolve, since a name only matters where a person can meet it.

Who should own the variant list?

The team that owns the trademark portfolio, because the list is derived from it and changes when it changes. Finance and security should be able to read it, since they are the ones who see unusual payment instructions and sign-in reports first.