A Broken Link Is Not an Invitation to Pitch Your Content

Broken-link building is worth doing when a maintained third-party page points to a genuinely dead resource and your replacement serves the same audience, performs the same job and improves the publisher’s page. If the entire argument is “their link is broken and our page is live,” do not send the email.

Most workflows start too far downstream. A crawler finds a 404, somebody appends authority metrics, an email address is scraped and the row becomes a prospect. The missing work is editorial reconstruction: why was the link placed in that sentence, and what does the reader lose now it has failed?

A dead link is not abandoned real estate. The publisher may remove it, find the original resource at a new address, cite a primary source, choose another independent resource or rewrite the passage. Your asset is one candidate in that comparison. My rule is that the replacement must earn the recommendation before outreach exists. Discovery of a broken URL gives you permission to investigate, not permission to pitch.

First prove the reference is actually broken

One failed request is evidence to inspect, not a campaign brief. Record the exact source page, linked URL, anchor text, surrounding passage, observed response and check date. Test the destination in a normal browser and with an HTTP check where appropriate. Look for redirects, consent walls, bot blocks, authentication, temporary server errors and geographic restrictions.

HTTP 404 Not Found means the origin server did not find a current representation for the target resource, or was unwilling to disclose one. RFC 9110 notes that the status does not reveal whether the condition is temporary or permanent. A timeout, 403 response or soft error returning 200 OK requires a different interpretation.

Then check whether the original organisation moved or replaced the resource. Search its website, documentation, navigation and obvious successor URLs. If an official successor exists, that may be the cleanest repair. You have discovered a publishing problem, not necessarily a customer-acquisition opportunity for your domain.

Separate sites you control. An internal dead link is maintenance: correct the source, redirect an old valuable route where appropriate, replace the reference or remove it. Broken-link outreach begins only when another publisher controls the page.

Read the sentence that used to carry the link

The surrounding passage is the closest surviving evidence of the missing resource’s job. Read the anchor, sentence, heading, page purpose, audience and adjacent sources. Then write one statement:

Suppose the anchor says download the pre-tenancy checklist inside a guide for private landlords. The job is not “landlord content.” It is to give that reader a usable checklist during tenancy preparation. A commercially important landlord-services page may be topically related and still fail the job completely. This is why keyword similarity produces dreadful outreach. Cyber security, ISO 27001, data protection and incident response can share a semantic neighbourhood while answering different questions for different audiences.

Google’s link best practices explain that anchor text and surrounding context help people and Google understand a destination. That does not create an entitlement to replace the link. It gives the researcher evidence for reconstructing the intended route. If an archived version is available, use it to understand function and scope. Do not reproduce protected expression or assume an old capture is current. Historical access can clarify what the resource did. It does not make your substitute original.

Test functional equivalence before content quality

The first gate is not “is our article good?” It is “does our resource perform the same function?” Compare audience, immediate task, format, jurisdiction, date or version, access conditions, evidence type and the moment in the publisher’s argument.

A polished thought-leadership essay is not equivalent to a calculator. A new US guide is not a safe replacement for withdrawn UK regulatory guidance. A gated template does not replace a freely downloadable checklist without changing the reader’s journey. A video-production reel does not replace technical documentation merely because both mention the product.

I call this the source-sentence fit test. If the candidate cannot be substituted into the original sentence without changing its meaning or promise, it fails before domain metrics, branding or outreach copy are considered.

The replacement should work without the campaign

Judge the resource as if no dead link had been found. The candidate should:

• Serve the source page's actual audience
• Fulfil the same immediate intent and functional promise
• Work without an undisclosed sale, registration or consultation
• Support material claims with attributable evidence
• State jurisdiction, date and scope limits where relevant
• Add original analysis, data, examples, utility or synthesis
• Remain available at a stable public address
• Have a credible owner and maintenance trigger

These are gates, not points in an invented authority score. Beautiful website design cannot compensate for the wrong audience. Proprietary research cannot compensate for an inaccessible paywall if the previous link promised a free tool. A long article cannot compensate for the missing template named by the anchor. Google’s people-first content guidance asks whether content provides original information or analysis, offers substantial value and helps its intended audience achieve a goal. Those are useful editorial questions. They are not a backlink certification, and passing them does not oblige a publisher to cite the page.

Commercial pages are not automatically excluded. An ecommerce compatibility guide, a documented methodology, an original dataset or a useful calculator can be the strongest available resource even when the company also sells. The question is whether the asset completes the promised job without baiting the visitor into another one.

Compare the repairs an editor should actually choose

“Better than a 404” is a meaningless standard. The real alternatives are live. For every candidate source, compare them in this order:

If another destination wins, reject your own placement. That is not timid growth marketing. It prevents time being spent on a pitch an editor can dismiss in ten seconds. I would rather identify the correct official successor than manufacture a reason to recommend a weaker company asset. A link that makes the source page worse is not a performance-marketing win worth reporting.

The same rule protects local outreach. A chamber, council, trade body, university or community site is not valuable merely because a third-party tool likes its metrics. The page must still be maintained, the audience must matter and the replacement must belong in the exact passage.

Decide whether outreach deserves human attention

An editorially defensible fit can remain a poor prospect. Research and outreach consume time from both teams. Set an effort ceiling before hunting contact data.

Broken-link building does not become safely scalable merely because crawling and email discovery can be automated. The judgement lives at the source-candidate pair. The more varied the publishers and resources, the less safely that judgement can be templated. Google’s spam policies define link spam around links created primarily to manipulate rankings. Broken-link building receives no magical exemption. Editorial usefulness, compensation, arrangement and resulting link practice still matter. Do not buy an unqualified editorial link, demand prescribed anchor text or disguise an agreement as neutral repair.

Write a note the editor can verify quickly

The outreach message should make the publisher’s work easier without pretending a relationship exists. Include the exact source-page URL; the section, sentence or anchor carrying the broken reference; the dead destination and observed result date; the candidate URL; one sentence explaining how it performs the old resource’s job; and any commercial relationship or limitation the editor should know. That is enough. Do not lead with a fake compliment, explain your company history or lecture the recipient on how broken links damage SEO.

A useful structure is:

Your guide at [source URL] links from [anchor/context] to [dead URL], which returned [observed result] when checked on [date]. We maintain [candidate URL], a [resource type] for [same audience/task]. It may fit that passage because [functional reason]. If another source is a better editorial fit, no action is needed.

The language is deliberately undramatic. The strength should come from fit, not persuasion tricks. Use a legitimate contact route and handle personal data proportionately. Do not scrape every employee ever associated with the domain. One relevant editorial contact is better than five simultaneous messages. Follow up within a restrained, pre-agreed limit, then stop. Silence is a decision.

Measure rejections before links

Campaign reports usually begin with emails sent and links won. Those numbers hide whether the research improved. Track the decision funnel instead:

• Suspected dead references found
• Failures verified
• Official successors identified
• Functional mismatches rejected
• Weak assets returned for improvement
• Abandoned publishers declined
• Defensible editorial approaches sent
• Responses, corrections and placements observed

This exposes the value of self-rejection. If 200 crawler rows become twelve defensible messages, that is not automatically a 94% failure rate. It may show that the gate prevented 188 pieces of noise.

Treat a placement as an observed editorial outcome, not a guaranteed ranking, traffic or business-growth result. Links can support discovery and relevance, but one acquired link does not carry a dependable revenue promise. Measure referral visits and suitable commercial outcomes where enough evidence exists, while admitting that many factors move simultaneously. The most useful performance metric may be the percentage of prospects rejected for a documented reason before outreach. It shows whether judgement is protecting the brand from low-quality automation.

Build assets worth citing before looking for vacancies

The best preparation happens before prospecting. Publish something with a reason to exist independently: original data with a disclosed method, a maintained calculator, a practical template, a primary-source comparison, a local dataset, a decision tool or a documented process that resolves genuine ambiguity. Then look for failed references whose former job matches it.

This order reduces the temptation to reverse-engineer thin content around whatever broken URLs a tool discovers. It also connects link acquisition to a broader content and digital-growth strategy: the company creates an asset that helps buyers, publishers and sales conversations whether or not outreach succeeds.

Close the row when the replacement does not win

Start with the source sentence. Prove the failure. Reconstruct the reader’s job. Test functional equivalence. Compare removal, official successors, primary evidence and independent alternatives. Only then decide whether the company’s resource deserves an editor’s attention.

I am comfortable closing a prospect row when somebody else’s source is better. That decision protects the publisher, the recipient and the brand doing the outreach. It also concentrates effort on the small number of placements where the recommendation can be defended without SEO theatre. A broken link is not an invitation to pitch your content. It is an invitation to solve an editorial problem. If replacing it with your page would not improve the publisher’s page for the publisher’s reader, close the row and do not press Send.