A company is usually credited with coining inbound marketing in 2005 or 2006. In a sworn statement to the trademark office, that company gave a first-use date of 8 July 2008.

The archived record agrees with the later date rather than the earlier one. The phrase is absent from every capture of the company’s own homepage until December 2007, and the domain belonged to a different business entirely until 2005.

Meanwhile the two-word phrase appears in the book record from the 1970s, in what looks like a different sense. None of this diminishes what the company built. It does mean the story usually told about the term is not the one the documents support.

What the archived record shows

Four checkable points, in order.

The domain was somebody else’s until 2005. Captures from that period show an intranet and extranet product, under a different copyright line.

The company associated with the term appears from April 2006. And in the captures examined, does not use the phrase.

The phrase is absent on 30 November 2007 and present on 11 December 2007. The December capture shows an “Inbound Marketing System”. Every capture examined between May 2005 and that date lacks it.

And the company’s own sworn date is later still. A trademark application filed in May 2015 and registered in November 2015, for a related one-word mark, declares a first use in commerce of 8 July 2008.

The book followed a year after that. Published in October 2009.

One thing this evidence cannot do. A complete sweep of one website’s archived captures is strong evidence about that website. It is not proof that nobody at the company used the phrase in conversation earlier. The distinction is worth keeping.

But the sworn date is a different kind of evidence. A first-use declaration to a trademark office is a legal statement by the applicant about their own history, and the applicant chose 2008.

Documented chronology of the appearance of the inbound marketing termThe documented chronology of the appearance of the term inbound marketing on the website of the company usually credited with coining it, together with the company’s own sworn statement to the trademark office. Web archive captures of the domain from the period up to 2005 show an entirely different business operating an intranet and extranet software product under a different copyright line, so the domain did not belong to the company now associated with the term. The company now associated with the term appears on the domain from April 2006, and in the captures examined does not use the phrase at any point during that period. A capture of the homepage dated the thirtieth of November 2007 does not contain the phrase. A capture of the same homepage dated the eleventh of December 2007 does contain it, presenting an inbound marketing system. A sweep of every available capture of the domain between May 2005 and that December 2007 date found no earlier occurrence. Separately, in a trademark application filed in May 2015 and registered in November 2015 in respect of a related single word mark, the company declared a first use and first use in commerce date of the eighth of July 2008, a declaration constituting a legal statement by the applicant about its own commercial history. The book associated with the term was published in October 2009, approximately a year after that declared first use. One limitation applies to this evidence, namely that a complete sweep of one website’s archived captures constitutes strong evidence about that website but does not establish that no person at the company used the phrase in conversation or in unarchived material at an earlier date, a distinction worth preserving. The sworn first use declaration is however a different category of evidence, since it is the applicant’s own account and the applicant selected 2008.What the captures showUntil 2005: a different company on the domainAn intranet and extranet product, under a different copyright line.April 2006: the company appearsAnd does not use the phrase in any capture examined.30 November 2007Phrase absent from the homepage.11 December 2007Phrase present: “Inbound Marketing System”.And the date the company swore to the trademark office8 July 2008declared as first use in commerce, in a filing made in 2015.One limitation worth keeping: a complete capture sweep proves things about a website, not about conversations.The sworn date is a different kind of evidence, and it is the company’s own.
Absent in November 2007, present in December 2007, and sworn as July 2008 by the company itself. Source : Internet Archive captures of the company homepage; USPTO Trademark Status and Document Retrieval, serial 86631432 (2026)

The phrase is older than the company

The second half of the story is the one that changes how the term should be read.

Corpus data records it from the 1970s. In a large book digitisation corpus, the two-word phrase registers from 1977, with near-continuous presence from 1986 through 2004.

One honest limit on that evidence. A two-word frequency count cannot distinguish a technical term from two words that happen to sit next to each other. The signal is real; the meaning requires a citation nobody has produced here.

The likely earlier sense is not the current one. Reference works of the period carry entries for inbound telemarketing, which is the handling of incoming calls in a direct response operation. That is inbound in the sense of a call arriving, not a buyer discovering you.

And one document could not be verified. No specific pre-2005 book, page or quotation was confirmed. The frequency data is solid; the reference is missing.

Which is worth stating rather than glossing. The defensible claim is that the phrase existed in the marketing literature well before 2005, most likely in a direct-response sense, and that the current meaning was popularised rather than invented by the company associated with it.

The trademark record supports that reading. No registration of the two-word phrase alone was located. A related 2009 application carried a mandatory disclaimer on the second word, and was abandoned that December. An office requiring a disclaimer on a component of a mark is saying that component is descriptive and cannot be owned.

Which is an unusual outcome for a coinage. A genuinely invented term is registrable. A term the office treats as descriptive is one the language already had.

What corpus frequency data establishes about the age of the term and what it cannot establishWhat corpus frequency data from a large book digitisation project establishes about the age of the two word phrase inbound marketing, and the limits of what such data can establish. The corpus records non zero frequency for the phrase beginning in 1977, with further occurrences in 1981, 1982, and from 1986 onwards including 1987, 1988, 1989 and 1990, followed by near continuous presence from 1992 through 2004, which is the period immediately preceding the date at which the term is commonly said to have been coined. This establishes that the two word string existed in the digitised book record for approximately three decades before the attribution date. Three limits apply to that evidence. First, a two word frequency count cannot distinguish a term used in a technical sense from two words that happen to appear adjacently for unrelated grammatical reasons, so the signal establishes existence of the string rather than existence of the concept. Second, the likely sense in the earlier period differs from the current one, since reference works of that era carry entries for inbound telemarketing, describing the handling of incoming telephone calls in a direct response operation, which is inbound in the sense of a call arriving rather than in the sense of a buyer independently discovering a supplier. Third, no specific bibliographic reference from before 2005 was verified, meaning no title, author, page or quotation is available to support the frequency signal, because the relevant search interfaces returned quota errors, consent blocks or access refusals. The defensible claim is therefore that the phrase existed in the marketing literature well before the attribution date, most probably in a direct response sense, and that the current meaning was popularised rather than invented by the company associated with it.The phrase in the book record19401977198620042022the attribution datefirst occurrencenear-continuous presenceWhat this supportsThe two-word string existed in the bookrecord for about three decades beforethe attribution date.What it does notThe sense. A frequency count cannot tella technical term from an adjacency, andno specific reference was verified.The likely earlier sense is inbound telemarketing: a call arriving, not a buyer discovering you.So: popularised rather than invented, which is a defensible claim and a different one.
The frequency signal is real. The meaning is not established, and no specific reference was verified. Source : Google Books Ngram corpus, English 2019, two-word phrase frequency (2026)

The conceptual predecessor is a book

Separately from the phrase, the idea has a documented antecedent that is usually mentioned and rarely credited properly.

Permission Marketing, published in 1999. Its argument is that attention must be volunteered rather than taken, and that a business earns the right to a next message by making the previous one worth receiving.

That is the substance of the later framework. Publishing something worth finding, being found, and continuing the conversation on terms the reader controls.

What the later framework added was operational. A stack, a vocabulary, a certification programme and a set of measurements. Those are real contributions and they are what made the idea adoptable by companies that would never have read the book.

Which is a fair way to state the credit. Not the coinage, and not the concept. The productisation.

And the productisation is the part worth examining critically, because the framework’s central promise, that publishing enough will make paid acquisition unnecessary, is not something the underlying idea ever claimed.

Three distinct attribution claims and the documentary support for eachThree distinct claims commonly bundled together in the attribution of the inbound marketing term, and the documentary support available for each of them. The first claim is that the company coined the phrase. The documentary record does not support this, since the phrase is absent from every archived capture of the company’s own website before December 2007, the company itself declared a first use in commerce of July 2008 in a sworn trademark filing, and corpus frequency data from a large book digitisation project records the two word phrase from 1977 with near continuous presence from 1986 to 2004, although that frequency data cannot establish the sense in which the phrase was used and no specific pre 2005 bibliographic reference was verified. The trademark record supports the same reading, since no registration of the two word phrase alone was located and a related application filed in April 2009 and abandoned in December of that year carried a mandatory disclaimer on the second word, an office requiring such a disclaimer being one that treats the component as descriptive and therefore incapable of exclusive ownership, which is an unusual outcome for a genuine coinage. The second claim is that the company invented the underlying concept. A documented conceptual predecessor exists in a book published in 1999 arguing that attention must be volunteered rather than taken and that a business earns the right to send a further message by making the previous message worth receiving, which is the substance of the later framework. The third claim is that the company productised the concept, providing a software stack, a vocabulary, a certification programme and a set of measurements. This claim is supported, constitutes a genuine contribution, and is what rendered the idea adoptable by organisations that would never have encountered the book. The fair statement of credit is therefore neither the coinage nor the concept but the productisation.Three claims, one that survivesClaim 1: coined the phrasenot supportedAbsent from its own site before December 2007. Sworn first use: July 2008. In the book corpus from 1977.And the trademark office required a disclaimer on “marketing”, which means descriptive.Claim 2: invented the conceptpredecessor existsA 1999 book argues that attention must be volunteered, and that each message earns the right to the next.That is the substance of the later framework.Claim 3: productised itsupportedA stack, a vocabulary, a certification programme, a set of measurements. A real contribution.And what made the idea adoptable by companies that would never have read the book.Which is the fair statement of credit: not the coinage, not the concept, the productisation.And the productisation is the part worth examining, because it added a promise the original idea never made.
Coining the phrase, inventing the concept, and productising both. Only the third survives the documents. Source : Archived captures, USPTO filings, corpus frequency data, and publisher records (2026)
The two categories of evidence used to date the term and what each can establishThe two distinct categories of evidence used to establish when the term entered use, and the specific limits of what each can prove. The first category is a sweep of archived web captures. A complete sweep of every available capture of a company’s homepage across a defined period constitutes strong evidence about the content of that website during that period, and in this instance establishes that the term was absent from every capture examined between May 2005 and the thirtieth of November 2007 and present in the capture dated the eleventh of December 2007. What such a sweep cannot establish is that no person within the company used the phrase in conversation, in correspondence, in a presentation or in any material that was never published to that domain or never archived, so the absence of the phrase from a website is not proof of the absence of the phrase from a company. The second category is a sworn statement to a trademark office. A declaration of first use in commerce is a legal statement made by the applicant about the applicant’s own commercial history, submitted under penalty, and in this instance the company declared a first use in commerce date of the eighth of July 2008 in an application filed in May 2015 and registered in November 2015. What such a declaration establishes is what the company itself claimed about its own history at a point when a false statement would have carried consequences, which is a materially stronger form of evidence about the company’s own use than any third party account. Taken together, the two categories converge on a date substantially later than the attribution in common circulation, one through the absence of contrary evidence on the company’s own website and the other through the company’s own affirmative statement.Two kinds of evidence, two limitsA capture sweepProves: what was on that website,across every capture examined.Does not prove: that nobody usedthe phrase in a room somewhere.A sworn filingProves: what the company itselfclaimed about its own history.Made under penalty, so a strongerstatement than any third-party account.And they convergeOne shows no earlier use on the site. The other names July 2008. Both point away from the usual story.Which is why the article says “the record does not support 2005” rather than “the term was not used”.The first is provable. The second is not, and the difference is the whole discipline.
A capture sweep proves things about a website. A sworn filing proves what the company itself claimed. Source : Internet Archive captures and USPTO Trademark Status and Document Retrieval (2026)

Why the genealogy changes the practice

Knowing where a framework came from tells you what it was designed to do, and what it was not.

The 1999 argument was about permission, not volume. Earning the right to send the next message. Nothing in it says that publishing enough removes the need to pay for reach.

The productised version added that promise. And it arrived at a moment when organic distribution was unusually cheap, which is a condition that has since changed and which no framework can restore.

Which explains a common disappointment. Companies adopting the framework in the 2010s inherited a promise made under distribution conditions that no longer hold, without inheriting the caveat, because the caveat was never written down.

The permission argument survives the change. Whether or not you can be found for free, the standard of making each message worth the next one is unaffected by anybody’s ranking system.

And the productisation survives too, minus the promise. A stack, a vocabulary and a measurement discipline are useful without the claim that they replace paid acquisition.

One practical consequence. Treat the framework as an operating method rather than a channel strategy. It tells you how to treat attention you have been given. It does not tell you where the attention comes from, and it never did.

Separation of the durable argument from the promise added during productisationSeparation of the durable underlying argument from the additional promise introduced when that argument was turned into a commercial framework, and the consequences of that separation for practice. The underlying argument, published in 1999, holds that attention must be volunteered rather than taken, and that a business earns the right to send a subsequent message by making the preceding message worth receiving. Nothing within that argument asserts that publishing a sufficient quantity of material removes the need to pay for reach. The promise added during productisation asserts precisely that, and it was introduced at a moment when organic distribution was unusually inexpensive, a condition that has since changed and that no framework is able to restore. This explains a widespread pattern of disappointment, since organisations adopting the framework during the following decade inherited a promise made under distribution conditions that no longer obtain, without inheriting any accompanying caveat, because no caveat was ever recorded. The permission argument survives the change in conditions entirely, since the standard of making each message worth the next one is unaffected by any ranking system operated by any platform. The productisation also survives, minus the promise, since a software stack, a shared vocabulary and a measurement discipline remain useful without the accompanying claim that they substitute for paid acquisition. The practical consequence is that the framework should be treated as an operating method governing how an organisation handles attention it has already been granted, rather than as a channel strategy explaining where that attention originates, which it does not address and never did.The argument, and the promise added to itThe 1999 argument, which survives everythingAttention must be volunteered rather than taken, and each message earns the right to the next.Unaffected by any platform’s ranking system, then or now.The promise added later, which does notThat publishing enough removes the need to pay for reach.Made when organic distribution was unusually cheap. That condition changed.Which explains the disappointment patternCompanies inherited the promise without the caveat, because the caveat was never written down.Treat it as an operating method for attention you have, not a strategy for where attention comes from.
The permission argument survives every platform change. The promise that publishing replaces paid reach does not. Source : Method, over the 1999 conceptual predecessor and the later productisation (2026)

What to do with this

When somebody presents inbound as a way to reduce paid spend, ask which part of the framework makes that claim. The permission argument does not, and the productised version made it under distribution conditions that have changed.

Keep the part that holds. Publish things worth finding, ask for permission rather than assuming it, and make each message earn the next. That standard is twenty-seven years old and unaffected by any platform decision.

Drop the arithmetic that assumes free distribution. Budget for reach separately, and judge published material on whether it survives being found rather than on whether it generates enough traffic to replace advertising. That separate budget has its own arithmetic, which is what our page on B2B paid acquisition describes: what a media budget buys, and how it is judged on cost per lead rather than on traffic.

And when the origin story comes up, the accurate version is short. The phrase existed before the company. The concept existed before the phrase. What the company built was the operating system around both, and that is a real thing to have built.

The related pieces are paid, earned, shared, owned and editorial line for B2B.