# The Rival Trademark Filing Reference: Real Launch, Defensive, or Dead

*Given one trademark record, write down what it proves about a rival launch, its earliest and latest window under the statutory clocks, and how each field could mislead.*

- Canonical URL: https://www.refolk.ai/guides/rival-trademark-filing-launch-reference
- Pillar: Market and talent intelligence
- Format: Reference
- Published: 2026-11-17
- Last reviewed: 2026-11-17
- Reading time: 20 min
- Keywords: how to use trademark filings for competitive intelligence, intent to use trademark application meaning launch, find competitor trademark filed by shell company, trademark 44(d) foreign priority submarine filing, statement of use deadline extensions timeline, predict competitor product launch from trademark

## Key takeaways

- In a long-run study of published intent-to-use trademark applications, only 49.0% went on to register, against 97.7% of use-based ones, so an unweighted 'rival filed ITU' brief is wrong about half the time.
- The only hard launch clock on a US intent-to-use filing starts at the Notice of Allowance: six months to file a Statement of Use, extendable to a 36-month maximum, after which the filing cannot support a launch.
- In one eleven-day sample of USPTO daily files, 73.6% of Statements of Use came from applicants who never extended, so a record on its third to fifth extension is drifting, not launching.
- A §44(d) priority claim can move the real filing date up to six months before the US filing date, and Apple Pay was filed in the US 99 days after it was publicly announced.
- Attribute a trademark filing to a rival only on two independent links, such as attorney of record plus entity address or a later assignment to the parent; shared outside counsel alone proves nothing.
- In Refolk's index, half of the top sampled employers of US trademark paralegals are in-house brand owners, against 5% for UK trade mark attorneys, which suggests the filing team often sits inside the rival even when outside counsel is on the record.

A rival's trademark application is one of the few public documents that names a product before it ships, and one of the easiest to over-read. This reference is for strategy and talent-intelligence analysts holding a single record who must decide, in writing, whether it signals a real launch (and in which category, country, and window) or a defensive or dead filing that should stay out of a competitor brief. Each field below is defined with what it proves, the clock it starts, and the way it lies.

## What can a single trademark record prove about a rival launch?

A single record proves only that someone claimed a mark for described goods, on a date, under a stated legal basis. It becomes a launch read when you combine the basis, priority date, attribution and status history, and weigh them against a base rate that is harsh on intent-to-use alerts.

An intent-to-use (ITU) application is one filed on a sworn bona fide intention to use the mark, not on use that is already happening. Most launch alerts are ITU filings, and most of the noise is too.

**49.0% - Share of published intent-to-use applications that went on to register**

In a long-run academic study covering nearly two decades of the ITU system, against 97.7% for published use-based applications.

Two more published findings point the same way. Research in INTA's Trademark Reporter (Gerhardt and Lee) finds the overall registration rate has stayed around 20 percentage points lower since ITU was introduced. Among unsuccessful applications, failing to file a Statement of Use was the second most common cause of failure, at 40%. ITU filings are also common. In an eleven-day sample of the USPTO daily files, 49.2% of 16,720 new applications claimed ITU only.

The result: about half of new filings are intent-only, and about 51% of the ITU filings that reach publication never register (derived). A brief that says "rival filed ITU, launch coming" without weighting is wrong about half the time.

A record can still give you four useful things:

- **Category:** the goods text names what the applicant was willing to swear intent for.
- **Geography:** the US filing, plus any foreign priority office or international route.
- **Outer timing:** once a Notice of Allowance issues, a hard latest date for this filing.
- **Intent strength:** the extension count shows whether the applicant is converting or stalling.

## Filing basis and priority claim: what each basis proves

The filing basis tells you whether the applicant claims the mark is already in use, only intends to use it, or is relying on a foreign filing or registration whose date may matter more than the US one. The TMEP (the USPTO's examination manual, Chapters 800 and 1100) lists five bases.

| Basis | What it proves | How it misleads |
|---|---|---|
| §1(a) use | Sworn claim the mark is already used in commerce | Use may not cover every listed good |
| §1(b) intent | Verified bona fide intent only | About 51% of published ITU never register |
| §44(d) priority | A foreign application exists, filed up to 6 months earlier | US filing date looks like the start of the plan |
| §44(e) foreign registration | Applicant owns a registration in its country of origin | No US use required before registration |
| §66(a) Madrid | An international registration was extended to the US | No US use required before registration |

For §1(b), §44(d) and §44(e), the applicant must file a verified statement of bona fide intent to use the mark in commerce. A §44 or §66(a) applicant must assert intent but does not have to show actual use before registration. That means a registered §44(e) or §66(a) mark tells you less about US launch timing than a registered §1(b) mark, which had to prove use first.

### The priority date field

The priority date is the most misread field on the record. Under Article 4 of the Paris Convention, the priority period is six months for trademarks (twelve for patents and utility models). A §44(d) applicant files in the US within six months of a foreign application and receives that earlier date. Section 44(d) gives only a priority filing date. It is not a basis for publication or registration, so the record will need another basis to register.

Section 44 does not require the foreign application to come from the applicant's home country. That gap creates the submarine filing: a first filing in a registry that does not publish online, followed by a US filing inside the six-month window. A WIPO working paper on submarine trademarks lists eight jurisdictions where applications were not published online during its study period:

- Honduras
- Jamaica
- Liechtenstein
- Mauritius
- Saint Lucia
- Swaziland
- Tonga
- Trinidad and Tobago

My sources conflict on whether Tonga is a Paris Convention member. Before you rely on a Tonga priority claim, check it against WIPO's Paris contracting-party list.

Apple Pay is the reference case. The mark was filed in Jamaica, announced publicly 82 days later, and filed in the US 99 days after the announcement. That is 181 days after the Jamaica filing, just inside the six-month window (derived from published dates). An analyst watching only the US register saw the filing three months after the launch event. A §44(d) filing can be a lagging indicator.

> **Watch out:** Read the priority date, not the filing date
>
> On a §44(d) record the plan started at the foreign filing, up to six months before the US filing date, and the US filing can arrive after the product is announced. Put the priority date and office on the memo before you compute any window.

### Madrid designations

A §66(a) record exists because WIPO's International Bureau sent the USPTO a request to extend an international registration. The US is one designation among the countries the applicant chose, and the full list lives with the international registration at WIPO. I have not established how often the WIPO Global Brand Database or EUIPO eSearch update, or which fields they cover. So read the designation list at the source and record the date you checked it, rather than treating a missing country as a decision not to launch there.

## Status history fields and the clocks they start

On a §1(b) application, the Notice of Allowance (NOA) starts the only hard clock. The applicant then has six months to file a Statement of Use, extendable to a 36-month maximum, after which this filing cannot support a launch. Everything before the NOA is administrative pendency with averages, not deadlines.

A Notice of Allowance is the USPTO's notice that an ITU application has cleared publication and will register once use is shown. A Statement of Use (SOU) is the sworn filing, with a specimen (a sample showing the mark as actually used), that converts intent into registrable use.

| Milestone | Value | Source type |
|---|---|---|
| Foreign priority window | 6 months | Paris Convention Art. 4 |
| Filing to first action | 4.1 months (annual avg) | USPTO reporting |
| Filing to disposal | 9.5 months (annual avg) | USPTO reporting |
| Publication to NOA | About 8 weeks | Law-firm guidance |
| NOA to last SOU | 36 months max | 37 CFR §2.89 |

The two pendency rows change every fiscal year. The USPTO publishes them annually, and the figures above come from the reporting this guide was built on. The prior year's baseline was 5.6 months to first action and 11.7 months total. Re-check before you use them. I have not found a published figure for filing to publication on its own.

#### The intent-to-use clock, filing to forced abandonment

1. **Filing** - Serial number issues at once; record visible in TSDR and the daily XML
2. **First action** - Examiner's first review, about 4.1 months on recent averages
3. **Publication** - Mark opened to opposition; date drives the NOA
4. **Notice of Allowance** - About 8 weeks after publication; six-month SOU clock starts
5. **Extensions** - Up to five six-month extensions, good cause sworn after the first
6. **SOU or abandonment** - Use shown and registered, or the filing dies at NOA plus 36 months

*Only the post-allowance leg is a statutory deadline; every earlier leg is an average that moves year to year.*

Pendency does not mean the filing is hidden. A US filing gets a serial number immediately and appears in TSDR and the daily XML. The real dark period is the time before a §44(d) priority claim surfaces, not the months spent in examination.

### Status fields after allowance

| Field | What it proves | How it lies |
|---|---|---|
| NOA issued | Publication cleared; 36-month clock running | Says nothing about launch intent |
| Extension count | 0 to 1 is on track; 3 to 5 is drifting | Good cause can be thin, such as market research |
| SOU with specimen | Use claimed on the goods kept | Use may not span the original goods list |
| Abandonment | This filing is dead | The name died; the category may not have |

The rules are set out in 37 CFR §2.89. The applicant may request five extensions. After the first, each extension is granted only in six-month increments, and the total may not exceed 24 months. Each extension after the first needs a showing of good cause: a statement of ongoing efforts to use the mark, such as market research or steps to find distributors. Five extensions together give the 36-month ceiling.

The extension count is the strongest single sign that a filing is dying. In the same eleven-day sample of USPTO daily files, 73.6% of the 2,792 SOUs came from applicants who never extended, and only 6.1% came from applicants in their final window. The same sample recorded 3,630 abandonments for failure to file a use statement against 4,565 NOAs issued. Most real launches convert fast, and a record on its third extension is starting to look like most of the dead ones.

## Class and goods text: what the wording commits to

The class numbers and the goods identification name the product category the applicant was prepared to swear intent for. Read them as a ceiling on scope, not a product specification.

Applicants routinely file broad identifications. One law-firm source cites audit data showing about 50% of audited registrations could not prove use on every listed good. So the original goods text is a menu, and the useful read is what survives SOU deletions. Before the SOU, I name the category in one phrase taken from the most specific wording in the goods text. I then mark the rest as possibly defensive breadth. That second judgment is my own, not a sourced rule.

What to write down for this field:

- The class numbers.
- The goods wording, verbatim and trimmed to the specific terms.
- One phrase naming the category the wording commits to.
- After any SOU, which goods were deleted and which remain.

## Attributing an applicant to a rival: entity, counsel, address, assignment

Attribute a filing to a rival only when at least two independent links tie it to them, such as attorney of record plus entity address, or a later assignment to the parent. Shared outside counsel alone proves nothing, and an unfamiliar applicant name alone rules nothing out.

Rivals who care about secrecy combine three concealment patterns:

- **Freshly formed applicant entities.** One docketing-industry write-up describes Apple creating shell companies to file its foreign applications. Apple-linked shells filed realityOS and xrOS.
- **Shared outside counsel.** One firm files for many clients, so a familiar attorney of record is a weak link on its own.
- **Priority from a registry with no searchable online register.** This keeps the plan dark for up to six months before the US filing.

#### Layers between a filing and the rival behind it

1. **Foreign priority filing** - Filed first in a registry that may not publish online
2. **Applicant entity** - Freshly formed company with its own name and address
3. **Attorney of record** - Outside firm that also files for unrelated clients
4. **In-house filing team** - Paralegals and counsel who docket the rival's portfolio
5. **Parent brand owner** - The rival, visible only through later assignment or matched links

*Each layer can hide the owner, so attribution works inward from the visible record to the parent one link at a time.*

For each record, check four things:

1. **Entity formation date.** A company formed shortly before the filing deserves a closer look.
2. **Address.** Compare it with the rival's known registered and office addresses.
3. **Attorney of record.** Compare it with the rival's known counsel on its other marks. Read this in TSDR. I have not confirmed that the bulk case-file data includes attorney fields.
4. **Later assignment.** A transfer of the mark to the parent is the strongest single link. It is recorded in the USPTO assignment files (TRTDXFAG).

### Who actually files: the in-house signal

The attorney of record is often an outside firm, but the people who run the docket may sit inside the rival. In Refolk's index of professional profiles, the trademark filing workforce splits sharply between in-house and firm employers.

| Segment | Profiles | In-house share, top sampled employers |
|---|---|---|
| Trademark Attorney, US | 244 | 19% (IBM, eBay) |
| Trade Mark Attorney, UK | 148 | 5% (Flutter) |
| Trademark Paralegal, US | 563 | 50% (incl. Apple, UPS, Home Depot) |

**50% - Top sampled US trademark-paralegal employers that are in-house brand owners**

From Refolk's index, against 19% for US trademark attorneys and 5% for UK trade mark attorneys; employer lists come from a 25-profile sample, so treat them as directional.

In Refolk's index there are 2.31 US trademark paralegals for every US trademark attorney. Banner Witcoff appears in both the US attorney and the US paralegal samples, which shows how one firm's name can sit across many unrelated filings. My inference, which I have not verified, is that the rival's own filing team and its docketing patterns reveal more than the attorney of record does. Knowing who runs a rival's trademark docket in-house tells you whose habits to look for in an unattributed filing.

Building that map by hand means searching profile by profile across employers and titles. In [Refolk](/) you describe the people you want in plain English and get the matching profiles back.

I ran this search: `Trademark paralegals currently working in-house at consumer electronics companies in California` - [see the full result list](https://www.refolk.ai/s/8qts69dhkp).

*Returns named in-house trademark paralegals with their current employers, a starting map of which brand owners run their own filing dockets.*

## Procedure: from one record to a written launch read

Work through the fields in this order. It takes one record to a dated, attributed read with its limits stated, and it is quick enough to run on every alert. Some practitioners who focus on submarine filings put the priority check first. The USPTO's own ITU guidance is organised around the status clock. I put basis and priority second because they change how every later date is read.

#### Single-record launch read

1. **Pull the record** - Pull it from the USPTO daily applications XML or TSDR. Capture the serial number, filing date, current basis and status code.
2. **Read the basis and the priority claim** - Record the basis. If it is §44(d), record the foreign filing date and office, and note whether that office is on the WIPO eight-jurisdiction list.
3. **Attribute the applicant** - Check entity formation date, address, attorney of record against the rival's known counsel, and any later assignment. Mark it attributed (two independent links), plausible, or ruled out.
4. **Read the class and goods text** - Read the goods wording, not just the class numbers, and name in one phrase the product category it commits to.
5. **Plot the status history** - Mark first action, publication, NOA, extension count and any SOU, and locate the record on the 36-month post-allowance clock.
6. **Compute the launch window** - Earliest is now, or earlier for §1(a) records and §44(d) records with an old priority date. Latest is NOA plus 36 months. Write both dates down.
7. **Decide whether to brief it upward** - Weight the read against the roughly 51% of published ITU applications that never register, then brief, hold, or drop with the reason written.

Desk estimates, which are not sourced: pulling the record takes minutes, and the priority check about 10 minutes. Attribution takes 30 to 90 minutes. The goods read and the status plot take about 15 minutes each.

### Computing the window when there is no NOA yet

Before allowance there is no statutory latest date, only an estimate. Add the pendency averages to the filing date to project an NOA, then add 36 months. My derived outer bound is that a §1(b) filing reaches forced abandonment roughly four years after filing. That assumes 9 to 12 months to NOA, which I have not verified for ITU applications specifically. Label any pre-NOA window "projected", not "statutory".

The latest date is the latest date this filing can support a launch. A rival can still launch later under a new filing, so a passed deadline kills the record, not the product.

#### Brief, hold, or drop a rival filing

Horizontal axis runs from Attribution weak or single-link to Attribution on two independent links. Vertical axis runs from Drifting (late extensions or abandoned) to On track (pre-NOA, or NOA with zero extensions).

| Quadrant | What it means |
| --- | --- |
| Live mark, owner unclear | Keep attributing; do not brief yet |
| Rival launch signal | Brief the category and window upward |
| Noise | Log the serial number and drop |
| Rival's parked name | Brief as defensive or dead, not as a launch |

*Attribution and conversion are separate questions, and only a record strong on both belongs in a launch brief.*

> **Rule:** Brief the category and the window, never the name
>
> A rival brief built from a trademark record states the product category from the goods text and the earliest and latest dates from the clock. The word mark goes in a footnote, if anywhere.

**Single-record launch read memo**

```
Serial number:
Word mark:
Source and date pulled:
Basis at filing / current basis:
Priority claim: none | office ______, foreign date ______, on WIPO eight-jurisdiction list: yes / no
Classes:
Goods text (specific terms, verbatim):
Category the wording commits to (one phrase):
Applicant entity / formation date / address:
Attorney of record:
Assignment history:
Attribution: attributed | plausible | ruled out
Independent links (two needed for attributed): 1) ______ 2) ______
Status position: pre-publication | published | NOA on ______ | extension __ of 5 | SOU filed | abandoned
Earliest plausible launch:
Latest date this filing supports: ______ (statutory: NOA + 36 months | projected: no NOA yet)
What this record proves:
How it could mislead:
Decision: brief | hold | drop, because ______
```

*One memo per serial number; write "not established" rather than guessing any field.*

## How trademark-based launch reads go wrong

Most bad trademark intelligence comes from eight predictable misreads. Each one has a mimic, a filing that looks like a launch signal and is not, and a specific check that catches it. Run every record past this list before it reaches a brief.

1. **Reading §1(b) as "launch coming".** About 51% of published ITU applications never register. The mimic is a defensive name reservation. Check: the extension count. With 73.6% of SOUs coming with zero extensions, a record on its third to fifth extension is drifting.
2. **Reading the §44(d) US filing date as the start of the plan.** The real date is the foreign filing, up to six months earlier, and the US filing can trail the launch. Apple Pay reached the US register 181 days after its Jamaica filing and 99 days after its announcement. Check: always read the priority date field, never only the filing date.
3. **Treating an unfamiliar applicant as a non-rival.** A freshly formed entity can be a shell, and dismissing it is a false negative. Check: the attorney of record, the address, and any later assignment to the parent in the USPTO assignment files.
4. **Treating an attributed shell's mark as the product name.** realityOS and xrOS were real Apple-linked filings, and the product launched as visionOS (reported through a secondary forum source quoting MacRumors). Check: brief the category and timing, never the name.
5. **Treating shared outside counsel as proof.** One firm files for many clients. Banner Witcoff, for example, appears in both of Refolk's US trademark samples. Check: require at least two independent links, such as counsel plus entity or address.
6. **Reading goods text as a product spec.** Applicants file broad identifications, and audit data cited by a law-firm source shows about 50% of audited registrations could not prove use on all goods. Check: what remains after SOU deletions.
7. **Briefing a fraudulent or junk filing.** The USPTO reported removing 14,000 invalid trademark applications and registrations in a single fiscal year. Check: whether the record or its filer has been removed or sanctioned. I have not established a dependable public method for this check, so treat it as a manual read of the prosecution history and say so in the memo.
8. **Assuming "unsearchable" registries stay hidden.** One law firm observes that search providers increasingly keep even obscure registry databases updated, and the WIPO list is a snapshot of its study period. Check: search the foreign office directly. A missing hit is not evidence that no filing exists.

> A trademark record that is strong on attribution but drifting on the clock is a parked name, not a launch.

The two failures that cost the most are opposites. Over-reading §1(b) fills briefs with dead names. Dismissing a shell applicant misses the one filing that mattered. The first is controlled by the base rate and the extension count, the second by the two-link attribution rule. Neither is fixed by reading more records, only by reading each one more carefully.

## Keeping a trademark watch current

A trademark watch goes stale in three places: the bulk feed, the clock figures, and the list of registries treated as hidden. Re-check each on its own cadence rather than trusting the numbers in this guide indefinitely.

- **The daily feed.** The USPTO daily applications XML (TRTDXFAP) updates daily. It carries word mark, serial number, registration number, filing date, registration date, goods and services, class numbers and status codes. Access is through the USPTO Open Data Portal and requires sign-in.
- **The research dataset.** The USPTO Trademark Case Files Dataset carries a foreign-priority flag and basis-at-filing flags. Use it to pull every §44(d) record for a rival's known entities in one pass. Its documentation does not publish an ITU abandonment share, so do not quote one from it.
- **The pendency averages.** Replace the first-action and disposal figures each time the USPTO publishes its annual numbers, and note the source date in the memo.
- **The submarine list.** Registry digitisation keeps shrinking the concealment window, according to one trademark-watch firm. Before treating a priority office as dark, test whether it now publishes online.
- **The rival's filing team.** People move between in-house teams and firms, which changes whose habits show up in unattributed filings. Re-running the in-house paralegal and counsel searches in Refolk each quarter keeps the attribution map current without rebuilding it.

#### Before a trademark read goes into a brief

- [ ] The memo records the priority date and office, or states that no priority was claimed.
- [ ] The foreign office has been checked against the WIPO eight-jurisdiction list, and any Tonga claim against WIPO's Paris contracting-party list.
- [ ] Attribution rests on at least two independent links, or the record is marked plausible or ruled out.
- [ ] The category is named from the specific goods wording, and any SOU deletions are reflected.
- [ ] The extension count is recorded and read against the 73.6% zero-extension conversion share.
- [ ] The latest date is labelled statutory (NOA plus 36 months) or projected (no NOA yet).
- [ ] The brief states the category and window, and the word mark is not presented as the product name.
- [ ] The memo states how this record could mislead, in one sentence.

## Frequently asked questions

### Does an intent-to-use trademark application mean a competitor is launching a product?

No, it means the competitor swore a bona fide intention to use the mark, which is much weaker than a launch. Only 49.0% of published intent-to-use applications in a long-run study went on to register, against 97.7% of use-based ones. Treat an ITU filing as a candidate signal and look for conversion evidence, mainly a Notice of Allowance followed by a Statement of Use with few or no extensions.

### How long after a trademark filing does a competitor usually launch?

There is no reliable average lead time, and the commonly repeated six to twelve month figure has no source I can stand behind. What exists is a statutory ceiling: once a Notice of Allowance issues, the applicant has at most 36 months to file a Statement of Use. The earliest launch can be now, or even before the US filing when a foreign priority claim is involved.

### How do I find a competitor trademark filed by a shell company?

Look for freshly formed applicant entities, then test them with independent links: the attorney of record against the rival's known counsel, the entity's address, and any later assignment of the mark to the parent, which appears in the USPTO assignment files. Require at least two links before attributing, because one law firm files for many clients.

### What is a 44(d) submarine trademark filing?

It is a US application that claims priority from an earlier foreign application, often filed in a registry that did not publish online. Paris Convention priority for trademarks is six months, so the foreign filing stays effectively hidden until the US filing appears. Always read the priority date, because the real plan starts there, and the US filing can arrive after the public announcement.

### How many Statement of Use extensions can a trademark applicant file?

Five, each for six months, giving a maximum of 36 months from the Notice of Allowance to the last Statement of Use. Every extension after the first requires a sworn showing of good cause, such as ongoing market research or efforts to find distributors. Most real launches convert without any extension, so a high extension count is the strongest single sign of a drifting filing.

### Should I put the trademarked name in a competitor brief?

Brief the category and the timing window, not the name. Apple-linked shell filings for realityOS and xrOS were real, yet the product launched as visionOS. A name can be a placeholder, a decoy, or abandoned while the product ships under something else, so the goods text and the clock carry more durable information than the word mark.

---

*From the Refolk guide library. I revise these guides rather than replacing them, so the current version is always at https://www.refolk.ai/guides/rival-trademark-filing-launch-reference*
