Off-Page SEO: The Complete Guide to Building Authority, Trust and Visibility

Search engines have never been able to take a website's word for itself. Any page can claim expertise, quality, or market leadership. What a search engine can observe is whether anyone else independently corroborates the claim — who links to the page, who cites the organisation, who reviews the business, who quotes its research, and how it is described across the rest of the web.

That body of external evidence is what off-page SEO manages. It is the discipline of earning the third-party signals that sit outside your own domain and influence how search systems, AI assistants, and human buyers judge whether you are worth surfacing.

Most guides collapse off-page SEO into link building. That is a serious under-description. Links remain the most measurable component, and they are still the most direct mechanism by which authority propagates across the web, but they now sit alongside brand citations, unlinked mentions, structured local data, review corpora, journalist coverage, expert commentary, and entity recognition in knowledge systems. Google's own representatives have spent the past several years reducing, not increasing, the emphasis they place on links relative to other signals — a shift documented in Section 3 and worth understanding before committing a budget.

This guide is organised to be used, not just read. It covers what off-page SEO is and why it resists direct control; how link-based ranking developed from citation analysis through spam-resistant models; what actually distinguishes a valuable backlink from a worthless one; the legitimate techniques available and their effort, risk and scalability profiles; digital PR as a discipline in its own right; brand mentions, reviews and local citations; competitor analysis and auditing workflows; how Google's spam policies treat manipulation; measurement frameworks; and a clearly separated section on off-page authority in AI-assisted search.

Throughout, claims are graded. Google's documented guidance is labelled as such. Academic research is described as what it proposed, not as how Google works today. Correlational industry studies are reported as correlations. Where the evidence is genuinely unsettled — and in several areas it is — the guide says so rather than manufacturing confidence.

Section 1: What is off-page SEO?

Off-page SEO is the practice of building and managing signals that exist outside your own website but influence how search engines and users evaluate it. Those signals include backlinks from other domains, brand mentions, business citations, customer reviews, media coverage, and third-party references to your organisation and its people. The unifying characteristic is that someone else controls publication.

Off-page or off-site?

The terms are interchangeable in practice. "Off-site SEO" is arguably the more literal description, since the defining feature is that the activity happens on other sites. "Off-page SEO" is the more common industry usage and the more heavily searched term. Both refer to the same discipline, and no meaningful distinction exists between them.

Off-page, on-page and technical SEO compared

[VISUAL SUGGESTION: On-page vs technical vs off-page comparison — three-column diagram showing control boundary]

The three are interdependent rather than parallel. Off-page authority cannot rescue a page search engines cannot index. Flawless technical hygiene will not make an unremarkable page competitive in a contested market. And earning external references is dramatically harder when there is nothing on the site worth referencing.

[Internal Link: Technical SEO Guide] · [Internal Link: On-Page SEO Guide]

Why off-page SEO cannot be fully controlled

This is the defining constraint of the discipline, and it has practical consequences.

You can publish a page whenever you like. You cannot make a journalist cover your research, a professor cite your dataset, or a customer leave a review. You can only make those outcomes more likely — by producing something worth referencing, making it discoverable by the people who would reference it, and building the relationships through which discovery happens.

Three implications follow. First, off-page work is probabilistic: campaigns produce distributions of outcomes, not guaranteed placements. Second, timelines are longer and less predictable than on-page work, which makes forecasting harder and patience a genuine requirement. Third, any technique that promises deterministic control over third-party publication — guaranteed link placements, purchased editorial coverage, bulk directory submission — is either low value or explicitly against Google's guidelines, and usually both. The absence of control is not a flaw in off-page SEO. It is precisely why the signals carry information.

Section 2: Why off-page SEO matters

Off-page SEO matters because third-party validation is the only evidence a search engine has that is not authored by the site being evaluated. Self-published claims are cheap; independent corroboration is not. External signals therefore carry disproportionate weight in distinguishing genuinely useful sources from merely well-optimised ones.

Beyond ranking mechanics, off-page work produces effects that show up across the business:

• Authority. A relevant editorial link from a recognised source in your field is a signal both search systems and human readers interpret as endorsement.
• Discovery. Links remain a primary path by which crawlers find new pages. An orphaned page with no external or internal links may simply never be found.
• Trust. Reviews, professional accreditations, media coverage and association memberships form the evidence base a cautious buyer uses before contacting you.
• Competitive differentiation. On-page and technical practices converge across a market — everyone eventually fixes their titles and their Core Web Vitals. Reputation, relationships and original research do not converge, because they cannot be copied quickly.
• Referral traffic. A well-placed link in a relevant publication delivers qualified visitors regardless of any ranking effect, and those visitors often convert better than search traffic because they arrive with third-party endorsement attached.
• Brand visibility and search demand. Coverage creates awareness, awareness creates branded search, and branded search is among the most commercially valuable demand a business can hold.
• Local visibility. For businesses serving a geography, citations, reviews and local relationships are not a supplementary tactic. They are the core of the discipline.
• AI discovery. Assistants that summarise and recommend draw heavily on third-party sources. What others publish about you increasingly shapes what an AI system says about you. This is covered carefully in Section 29.

The organising principle is simple enough to state in one line: what other websites and people say about you matters, and you cannot say it about yourself.

Section 3: The history of link-based search

Understanding how link analysis developed is not academic trivia. Nearly every modern debate in off-page SEO — relevance versus authority, quantity versus quality, why manipulation gets detected — was first framed in published research between 1998 and 2008. Knowing that literature tells you which arguments are settled and which never were.

Important caveat applied throughout this section: these papers describe what researchers proposed at the time. They do not describe how Google's current ranking systems work. Google's systems are proprietary, heavily machine-learned, and have evolved for over two decades. Historical research explains the concepts and the problems, not the implementation.

Before links: content-only retrieval

Early web search engines ranked largely on page content — term frequency, keyword placement, meta keywords. The weakness was structural rather than incidental: every input was controlled by the person who wanted to rank. Repeating a phrase was free, so ranking became a function of willingness to repeat phrases. Search quality degraded accordingly.

PageRank: links as citations

Brin and Page's 1998 paper "The Anatomy of a Large-Scale Hypertextual Web Search Engine," alongside the PageRank technical report by Page, Brin, Motwani and Winograd, introduced the idea that transformed the field. Borrowing from academic citation analysis, they treated a hyperlink as a vote — and, crucially, weighted each vote by the authority of the page casting it.

The insight was recursive: a page is important if important pages link to it. Because the signal came from third parties rather than the page owner, it was substantially harder to manufacture than keyword repetition. PageRank was query-independent — computed over the whole web graph, not per search — which made it a global measure of importance.

HITS: hubs and authorities

Jon Kleinberg's "Authoritative Sources in a Hyperlinked Environment" (Journal of the ACM, 1999) proposed a different decomposition. Kleinberg distinguished authorities (pages with valuable content on a topic) from hubs (pages that link to many good authorities), with the two reinforcing each other: good hubs point to good authorities, and good authorities are pointed to by good hubs. Unlike PageRank, HITS was computed on a query-specific subgraph, making it relevance-sensitive by construction.

Lempel and Moran's SALSA (2000) combined elements of both approaches using a stochastic model, and was notably more resistant to the "tightly knit community" effect — the tendency of a densely interlinked cluster of pages to inflate its own scores. That resistance turned out to matter enormously, because a tightly knit community that inflates its own scores is, functionally, a link scheme.

Hilltop: authority conditioned on expertise

Bharat and Mihaila's Hilltop (2001) attacked a specific weakness. Global authority scores could allow a very popular but topically unrelated page to influence rankings for a query it had no business influencing. Hilltop restricted the calculation to "expert documents" — pages judged to be authoritative about a topic and reasonably independent of the target — and derived authority from links originating in that expert set.

The practical lesson has aged perfectly: a link's value depends on the linking source's relevance to the subject, not only on its general prominence. Every practitioner intuition about topical relevance in link building traces back to this line of research.

Topic-Sensitive PageRank

Haveliwala's Topic-Sensitive PageRank (2002) generalised the idea by computing multiple PageRank vectors biased toward different topic categories, then blending them according to query context. Authority stopped being a single number and became something conditional. Chakrabarti and colleagues' work on the structure of broad topics on the web (2002) supplied supporting evidence that the link graph is meaningfully organised by subject rather than randomly connected.

The adversarial turn: web spam research

By the mid-2000s the link graph was under sustained commercial attack, and a substantial research literature emerged in response. Twelve of the papers commonly cited in this area converge on three mechanisms rather than twelve separate ideas:

1. Trust and distrust propagation from seed sets. Gyöngyi, Garcia-Molina and Pedersen's TrustRank (VLDB, 2004) began with a small set of manually vetted reputable pages and propagated trust outward through links, on the reasoning that good pages rarely link to spam. Krishnan and Raj's Anti-Trust Rank (2006) inverted it — propagating distrust backwards from known spam, on the reasoning that pages linking to spam are themselves suspect. Benczúr and colleagues' SpamRank (2005) examined whether a page's incoming PageRank distribution looked statistically abnormal.

2. Statistical outlier detection on graph features. Gyöngyi and colleagues' mass estimation work (VLDB, 2006) quantified how much of a page's PageRank derived from spam-linked sources. Becchetti, Castillo, Donato, Leonardi and Baeza-Yates used rank propagation and probabilistic counting (2006), and later broader link-based characterisation (2008), to show that spam pages exhibit distinctive distributions across features like degree, reciprocity, and the shape of their neighbourhood graphs. Ntoulas, Najork, Manasse and Fetterly (WWW, 2006) demonstrated the same principle on content features rather than link features.

3. Topological clustering. Castillo, Donato, Gionis, Murdock and Silvestri's "Know Your Neighbors" (SIGIR, 2007) established that spam clusters — spam sites disproportionately link to and are linked from other spam sites — so a page's neighbourhood is itself predictive. Abernethy, Chapelle and Castillo's graph regularisation methods (2008) formalised this by treating link-connected pages as likely to share labels.

Gyöngyi and Garcia-Molina's "Web Spam Taxonomy" (2005) catalogued the techniques of the era: link farms, honeypots, expired-domain acquisition, comment injection, hidden links. Reading it today is instructive mainly for how little the tactics have changed.

The synthesis that matters: search engines have had, for two decades, published methods for identifying manipulated link patterns using trust propagation, statistical anomaly detection, and neighbourhood analysis. Any strategy premised on manipulation being undetectable is arguing against a mature adversarial literature.

Behavioural alternatives and stated limits

Henzinger, Motwani and Silverstein's "Challenges in Web Search Engines" (SIGIR Forum, 2002) named link spam explicitly as a major open problem — a notable admission from within the field. Liu and colleagues' BrowseRank (SIGIR, 2008) proposed a more radical response: build the importance graph from observed user browsing behaviour rather than from hyperlinks, on the grounds that links are cheap to create while genuine user visits are not.

From link scores to machine-learned systems

Modern search ranking is not a link algorithm with adjustments. It is a set of machine-learned systems drawing on many signal families, including relevance models built on natural language understanding, quality assessment, freshness, locality, and user-experience signals, of which link-derived signals are one input among many.

Google's public commentary tracks this shift. In 2016, Google search quality strategist Andrey Lipattsev described links as one of the top three ranking signals alongside content and RankBrain. By September 2023, speaking at Pubcon Pro in Austin, Google Search analyst Gary Illyes said he would not place links in the top three and that they had not been for some time, adding that ranking without links is possible. Duy Nguyen of Google's search quality team had said in a 2022 office-hours session that backlinks as a signal carry considerably less impact than in Google's early years.

Read these statements carefully. Google is not saying links are irrelevant — Illyes explicitly said they remain important and that their significance is overestimated. The claim is about relative weight, and it is a claim from Google representatives rather than an audited disclosure. Treat it as strong evidence about direction of travel, not as a precise coefficient.

Section 4: How links create authority

[VISUAL SUGGESTION: Link graph illustration — nodes, edges, and authority propagation across topically clustered clusters]

The web can be modelled as a directed graph: pages are nodes, links are edges. Link-based authority models exploit the observation that this graph is not random. People link to things they find useful, and that behaviour, aggregated across millions of publishers, encodes information no individual publisher could fake alone.

Authority propagates along edges. A page receiving links from pages that themselves receive many links accumulates more than a page receiving links from isolated pages. This is the recursive property PageRank formalised, and it is why "link equity" — the informal term for transferable authority — is a reasonable mental model even though no public formula describes what Google actually computes.

Several factors modulate how much a link plausibly conveys:

• Source strength. A link from a widely referenced page carries more than one from a page nothing references.
• Link volume on the source page. A page linking to thousands of destinations distributes whatever it conveys more thinly than one linking to three.
• Topical relevance. Hilltop and Topic-Sensitive PageRank established the principle that authority is more meaningful when the source is relevant to the subject.
• Placement and context. A link inside the body of an article, surrounded by relevant discussion, sits in a different category from one in a sitewide footer.
• Editorial discretion. Whether a human chose to place the link, in the reader's interest, is the distinction Google's link spam policy turns on.
• Destination relevance. A link pointing at a page that genuinely serves the linking context is more coherent than one pointing at an unrelated commercial page.

Why 100 weak links do not equal one strong editorial link

Practitioners repeat this because it holds up, and the reasoning is worth stating precisely rather than asserting.

A weak link is usually weak on several dimensions simultaneously. It sits on a page with little authority of its own, so there is little to propagate. It may sit among hundreds of outbound links, diluting whatever remains. It typically has no topical connection to the destination. It was not placed by editorial judgement. And it frequently sits in a neighbourhood that the spam-detection literature describes as characteristic of manipulation. These are compounding deficits, not a single small one repeated.

Meanwhile a strong editorial link from a relevant, well-referenced source is positive on every dimension at once — and it carries effects the link graph doesn't capture: real referral visitors, brand exposure to a relevant audience, and the possibility that other publishers see the coverage and reference you independently.

A note on arithmetic: you will see PageRank presented as a tidy formula for dividing authority among outbound links. The original damping-factor model is real and published, but Google has not operated a naive PageRank calculation for many years, and translating the 1998 formula into a prediction about a link's present-day value is unsound. Use the concept; distrust the arithmetic.

Section 5: What makes a high-quality backlink?

A high-quality backlink is editorially placed on a topically relevant page, from a site with genuine authority and real readership, within content where the link makes sense to a human reader. Every other consideration is secondary to those.

Here is a fuller evaluation framework, roughly in order of importance.

Backlink Quality Evaluation Checklist

Run any prospective or acquired link through this. Failures on the first four are usually disqualifying.

• Placed by editorial choice, not purchased or exchanged
• Linking site is topically related to our subject area
• Linking page is specifically relevant, not just the domain
• Link would make sense to a reader who knows nothing about SEO
• Site has identifiable ownership, authors and editorial standards
• Site's outbound links point to legitimate destinations
• Linking page is indexed and appears in search results
• Link is a standard crawlable <a href> element
• Anchor text is descriptive and not over-optimised
• Link sits in main content, not in a footer or sidebar block
• Page plausibly has readers who might click through
• Site does not sell links or advertise guest-post placements
• Destination page genuinely serves the linking context
• This link is not trivially available to every competitor
• Nothing about the acquisition would embarrass us if published

That last item is the most efficient single test in link evaluation. If you would not be comfortable seeing the acquisition method described in a trade publication, reconsider.

Section 6: Referring domains versus total backlinks

Referring domains — the number of distinct websites linking to you — is generally the more informative metric. Total backlink count includes every individual link, so a single site with a sitewide footer link can contribute thousands of backlinks while representing exactly one editorial decision.

Compare two profiles:

Site B represents a hundred separate publishers independently deciding you were worth referencing. Site A represents one publisher and a template.

On diminishing returns: it is reasonable to expect that the tenth link from the same domain conveys less than the first, and this is consistent with how the underlying models were described. But there is no public formula for the rate of decay, and claiming a specific one would be inventing a ranking factor. Treat the principle as sound and the parameters as unknown.

Useful distinctions:

• Domain diversity — links from many distinct sites rather than a few, repeatedly.
• Link diversity — a natural mix of link types: editorial mentions, resource listings, coverage, citations, profile references, user-generated discussion.
• Sitewide links — appearing on every page of a source. Usually template-driven, and commonly a footprint of paid or reciprocal arrangements. Not automatically harmful; rarely valuable in proportion to their volume.
• Unique referring domains — the count most reporting should lead with.

For context on the base rate: Ahrefs' analysis of roughly a billion pages in its Content Explorer index found that around 66% had no external backlinks at all. The practical reading is that link acquisition is genuinely difficult and genuinely uncommon, not that most pages are failing at something easy.

Section 7: Anchor text

Anchor text is the clickable text of a link, and it gives search engines a description of the destination written by someone other than the destination's owner. That is precisely why it was valuable to early search engines and precisely why it became a target for manipulation.

What natural distributions look like

Genuinely earned link profiles skew heavily toward brand names, URLs, article titles and generic phrases, because that is how writers actually reference sources. A profile dominated by commercially valuable exact-match phrases is unusual precisely because writers do not naturally write that way.

Over-optimisation is the failure mode. Google's link spam policy identifies keyword-rich anchor text used in manipulative link building as a spam signal, and anchor manipulation has been a documented target of enforcement since the early Penguin era.

There is no perfect anchor text ratio

This bears stating plainly, because the myth is persistent and monetised. No universal ideal anchor text distribution exists. Nobody outside Google knows what distributions its systems evaluate, natural distributions vary enormously by industry and content type, and any published percentage target is inference presented as fact.

The workable approach: don't dictate anchor text for links you earn. When a publisher asks what to use, offer something descriptive and natural. Audit your profile periodically for implausible concentrations of commercial phrases. And treat any consultant offering a precise ratio as offering a guess with a decimal point attached.

Section 8: Standard, nofollow, sponsored and UGC links

[VISUAL SUGGESTION: Link attribute decision diagram — is the link paid? user-submitted? endorsed?]

Google's link attributes let a publisher describe the relationship a link represents. The current framework was established in a September 2019 announcement from Google's Danny Sullivan and Gary Illyes, which introduced two new values alongside the long-standing nofollow.

<!-- Standard link: implies editorial endorsement --> <a href="https://example.com/">descriptive anchor text</a> <!-- Nofollow: no endorsement implied --> <a href="https://example.com/" rel="nofollow">anchor text</a> <!-- Sponsored: advertising, sponsorship, affiliate or paid placement --> <a href="https://example.com/" rel="sponsored">anchor text</a> <!-- UGC: links within user-generated content --> <a href="https://example.com/" rel="ugc">anchor text</a> <!-- Attributes can be combined --> <a href="https://example.com/" rel="ugc sponsored">anchor text</a>

Directives became hints

The significant change: Google announced that these attributes function as hints rather than absolute directives. They took effect as hints for ranking purposes from the 2019 announcement and for crawling and indexing from 1 March 2020. Google's stated reasoning was that discarding all information from nofollowed links lost useful signal — including information about unnatural linking patterns — so a hint model lets Google retain the information while allowing publishers to withhold endorsement.

Practical consequences:

• You do not need to retrofit existing nofollow links. Google said explicitly that no changes are required. Use the more specific attributes on new links where convenient.
• Sponsored is the one attribute where accuracy matters most. Paid links that pass ranking credit violate Google's link spam policy. Marking genuinely paid links as sponsored (or nofollow) is the compliant approach.
• Nofollowed links are not worthless. They still deliver referral traffic, brand exposure, and discovery, and under a hint model Google reserves the option of using them. Refusing a relevant, high-visibility placement because it will be nofollowed is usually a mistake.
• Attributes are for publishers to apply, not requests to make. Asking a site to remove a nofollow so you get more ranking credit is asking them to misrepresent the relationship.

For scale context, Ahrefs' analysis of backlinks to the top 110,000 sites found roughly 10.6% carried the nofollow attribute, with the newer ugc and sponsored values used on a very small fraction of referring domains — adoption of the specific attributes remains limited years after introduction.

Section 9: Link earning versus link building

The distinction is partly semantic and partly strategic. "Link building" describes deliberate campaigns to acquire links. "Link earning" describes creating conditions where links arrive because something merits reference. In practice, most effective programmes combine them: you build assets that can be earned, then actively promote them so the right people find them.

Limitations worth stating. Pure link earning is slow and demands something genuinely worth referencing, which not every business has yet. Pure link building is labour-intensive, doesn't compound, and drifts toward transactional tactics under pressure to hit targets. Digital PR can consume significant budget and produce nothing when the angle doesn't land.

The mature position: build the asset, then work to get it in front of people with reason to reference it. Neither half works nearly as well alone.

Section 10: The best off-page SEO techniques

The table below profiles legitimate techniques. Effort, risk and scalability are practitioner assessments, not measured values.

A caution on scholarships. Scholarship link building was widely abused as a route to .edu links and is now heavily discounted by both universities and practitioners. If you fund a scholarship, fund it because you want to fund it. Do not build a campaign on the assumption of link returns.

Selection principle. Choose techniques matched to what your organisation actually has. A business with proprietary transaction data should run data studies. A firm with credentialed specialists should pursue expert commentary. A local trades business should invest in community relationships and citations, not attempt data journalism. The most common strategic failure in off-page SEO is copying the techniques of businesses with entirely different assets.

Section 11: Digital PR and SEO

Digital PR is the most reliable route to genuinely high-authority links, and it is the discipline most SEO teams are least equipped to run — because it is a media relations skill, not a search skill.

Digital PR is the practice of creating newsworthy stories and placing them with journalists and publishers, producing coverage that generates links, brand mentions, awareness and search demand simultaneously. It differs from traditional PR mainly in that outcomes are measured partly in link and search terms, and from link building in that the unit of value is a story rather than a placement.

What makes something newsworthy

This is where most campaigns fail, and no amount of outreach volume compensates. Journalists respond to stories, not to companies. A story generally needs at least one of:

• Novel data. Something nobody has published, especially if it can be broken down regionally, by sector, or by demographic so multiple outlets find their own angle.
• Timeliness. A connection to something already in the news cycle.
• Conflict or surprise. Findings that contradict what the audience assumes.
• Human consequence. A clear answer to "why should a reader care?"
• Authority. A credible source with standing to comment.

A product launch is not a story. A price change is not a story. A rebrand is definitively not a story. Research on how a widely-felt problem actually affects people, with numbers, is a story.

Digital PR formats that work

• Original surveys. Commission a representative panel, ask questions nobody has asked, publish the methodology openly. Regional breakdowns multiply coverage.
• Data journalism from proprietary data. If your business generates data in the ordinary course of operating, you may be holding a story nobody else can tell.
• Public dataset analysis. Government and regulator open data, analysed properly, produces defensible findings at modest cost.
• Reactive commentary. Fast, quotable expert reaction to breaking developments in your field. The highest return-per-hour tactic in digital PR, and it depends entirely on being fast and available.
• Thought leadership. Substantive positions on contested industry questions, published under a named expert.
• Newsjacking. Attaching genuine expertise to a live story. Works when the connection is real; embarrasses you when it isn't.

Digital PR workflow

[VISUAL SUGGESTION: Digital PR workflow — linear diagram from research through measurement with feedback loop to reclamation]

1. Research. Identify a question your sector's audience cares about and no one has answered with data. Verify it hasn't been covered recently.
2. Angle. Write the headline you want a journalist to write, before building anything. If you can't write a compelling headline, the story isn't there.
3. Asset. Build the underlying work: survey, analysis, dataset, or report. Document methodology fully — journalists will check, and reproducibility is what separates citable research from marketing.
4. Media list. Identify specific reporters who cover this beat, not publications. Read their recent work. A list of forty relevant journalists beats a list of four hundred generic addresses.
5. Pitch. Short, subject line as the headline, the finding in the first sentence, data and methodology accessible, contact details for a named expert. No attachments.
6. Follow-up. Once, politely, after a few days. Not more.
7. Coverage. Track placements as they appear, including syndication and pickup you didn't pitch.
8. Link monitoring. Identify which coverage linked and which mentioned you without linking.
9. Reclamation. Politely request links on unlinked coverage. Success rates vary widely; the ask is cheap and worth making.
10. Measurement. Coverage volume, referring domains earned, publication relevance, referral traffic, branded search movement, and where possible assisted conversions.

Why digital PR compounds

A successful campaign produces authority signals from sources that cannot be bought, brand awareness among a relevant audience, links from publications no outreach campaign would obtain, measurable referral traffic, increased branded search demand, journalist relationships that make the next campaign easier, and reinforcement of your organisation as a recognised entity in its field.

That last point connects directly to Section 29. Systems that summarise and recommend draw on published third-party sources. Being the organisation whose research gets cited is a durable position across both traditional and AI-assisted search.

Section 12: Brand mentions and entity authority

A linked mention references your brand with a hyperlink. An unlinked mention references it without one. Both are common in real coverage; unlinked mentions are often more common, because many publications link sparingly by policy.

Important accuracy note: Google has not confirmed that unlinked brand mentions function as a direct ranking signal, and the frequently-cited "implied links" patent describes a possible approach rather than a confirmed live system. Anyone telling you unlinked mentions are a confirmed ranking factor is overstating the evidence.

What can be said with more confidence is that unlinked mentions contribute to outcomes that matter:

• Reputation. Cumulative third-party references shape how a cautious buyer assesses you.
• Visibility and discovery. People find you through coverage regardless of hyperlinks.
• Entity understanding. Consistent references to an organisation, its people, its location and its field across many independent sources are how knowledge systems build confident entity representations.
• Reclamation inventory. Every unlinked mention is a candidate link, obtainable with a short email.
• Branded search demand. Coverage drives people to search your name, which is commercially valuable independent of any algorithmic effect.

Entities and the Knowledge Graph

Modern search systems reason about entities — people, organisations, places, products — as distinct things with attributes and relationships, not just as strings of text. Google's Knowledge Graph is the best-known implementation.

Being recognised as an entity is not something you can switch on. It emerges from consistency: the same organisation name, same location, same domain, same described specialism, repeated across independent authoritative sources — your own site, industry bodies, business registries, media coverage, review platforms, professional profiles. Contradictory information across those sources makes confident entity resolution harder.

Practical steps that are defensible without overclaiming: keep organisation naming consistent everywhere; maintain accurate structured data for organisation and location; ensure key people have consistent professional profiles; pursue references from sources already well-understood in your field; and correct inaccurate third-party information where you can.

Co-occurrence — your brand appearing alongside topic terms across many sources — is a related concept practitioners cite. It is a reasonable inference about how association is learned, and it is not a documented Google mechanism. Labelled accordingly.

Section 13: Online reviews as off-page SEO

Reviews sit in off-page SEO because they are third-party content about your business published where you don't control it. Their effects need careful separation, because conflating them produces bad decisions.

Three distinct effects:

1. Local ranking influence. Google's own local ranking guidance identifies prominence as a factor in local results, and review quantity and rating are among the inputs it references. This is the clearest documented ranking connection reviews have — and it applies to local results, not to web results generally.

2. Consumer influence. This is large, well-measured, and independent of ranking. BrightLocal's Local Consumer Review Survey found that about 75% of consumers "always" or "regularly" read online reviews when researching a business, while only around 3% say they never do. More recent editions report tightening expectations around star ratings and review recency. Whatever reviews do algorithmically, they substantially determine whether someone who has already found you contacts you.

3. Reputation and brand signals. Review corpora across platforms form part of the public record about your organisation, feeding both human judgement and the third-party sources that AI systems summarise.

Practical implications:

• Ask consistently and systematically, at the natural moment of satisfaction.
• Never incentivise, filter for positives, or fabricate. Platform policies prohibit it, Google has enforced against it, and fake reviews are increasingly detectable and legally risky in many jurisdictions.
• Respond to reviews, including negative ones, in a way a prospective customer would find reassuring.
• Diversify beyond Google onto the platforms your sector's buyers actually consult.
• Treat recency as part of the picture — a strong rating built entirely three years ago reads differently from an active one.

What not to claim: there is no confirmed evidence that reviews on third-party platforms influence general organic web rankings the way they influence local pack results. Sell reviews internally on conversion and local visibility, which are both well-supported, rather than on speculative organic ranking effects.

Section 14: Local citations

A local citation is any online reference to your business's core identifying information — name, address and phone number, commonly abbreviated NAP. Citations help search engines corroborate that a business exists, where it operates, and how to contact it.

Both matter. Structured citations are easier to build and audit; unstructured ones often carry more weight because they arise from genuine coverage.

Where citations come from

Major business directories and mapping platforms; industry-specific directories relevant to your trade; chambers of commerce and boards of trade; professional associations and licensing bodies; local media and community publications; supplier and dealer locator pages; event and sponsorship listings.

Consistency and cleanup

Consistency is the operative concern. If your business appears with three phone numbers, two suite numbers and an old trading name across various platforms, you are supplying contradictory evidence about a single entity.

A citation cleanup workflow:

1. Establish the canonical NAP — exactly how the business should be represented everywhere, including formatting.
2. Audit existing citations across major platforms and sector directories.
3. Identify inconsistencies: old addresses, disconnected numbers, name variants, wrong categories.
4. Identify duplicates — the most common and most damaging problem, since duplicate listings split reviews and confuse resolution.
5. Correct or claim listings, prioritising the platforms with the most visibility in your market.
6. Merge or remove duplicates through each platform's process.
7. Re-audit periodically, particularly after any move, rebrand or number change.

Where citations matter most

Citations are not equally important everywhere, and treating them as universally critical wastes budget.

They matter most for businesses serving a specific geography where customers search with local intent, in sectors with established directory ecosystems (trades, legal, medical, dental, hospitality, automotive), in competitive local markets where competitors have complete profiles, and in regulated professions where licensing directories carry genuine authority.

They matter least for purely online businesses with no service area, national e-commerce, and SaaS products sold without geographic constraint. A software company with customers in forty countries gains little from a local directory listing.

Section 15: Local link building

Local link building targets relevance to a place rather than volume. A link from a regional newspaper or a municipal business association can outperform a nationally larger but geographically irrelevant source for local visibility.

Productive sources, roughly by accessibility:

• Local and regional news. Genuinely newsworthy local stories: expansion, hiring, community initiatives, expert commentary on regional issues.
• Chambers of commerce and boards of trade. Membership typically includes a directory listing, and active participation produces more.
• Industry and trade associations. Member directories plus opportunities to contribute expertise.
• Community organisations and charities. Sponsorship and volunteering, where the involvement is real.
• Local events and festivals. Sponsor listings, speaker pages, exhibitor directories.
• Suppliers, distributors and partners. Dealer locators, partner pages, case studies.
• Universities and colleges. Guest lectures, co-op placements, research collaboration, career pages.
• Local resource pages. Municipal and community "local businesses" or "useful services" listings.
• Local blogs and interest publications. Regional food, culture, trade and lifestyle sites.

The governing rule: relevance beats volume decisively here. Ten links from organisations genuinely connected to your community and sector are worth more than two hundred from generic directories that list every business anywhere. If a directory would list any business that submitted a form, its endorsement carries approximately the information content of that fact.

Section 16: Competitor backlink analysis

Competitor analysis answers a specific question: which links exist in this market, which of them could realistically be obtained, and what does the pattern reveal about how competitors earn authority?

Workflow

1. Identify true SERP competitors. Not the companies your sales team names — the domains actually ranking for your priority queries. These differ more often than expected.
2. Export referring domains for each competitor, at domain level, using a backlink tool. Domain-level exports are more workable than link-level.
3. Categorise every referring domain. This is the step most teams skip and the step that creates the value:
   • Editorial coverage (media, publications)
   • Industry and association sources
   • Resource and curated listings
   • Directories and aggregators
   • Partner, supplier and customer sources
   • Guest contributions
   • User-generated and forum sources
   • Paid or clearly transactional placements
   • Local and geographic sources
4. Find link intersections — domains linking to two or more competitors but not to you. These are the highest-probability prospects, because the source has demonstrated willingness to link within your category more than once.
5. Separate replicable from unreplicable. Replicable: association memberships, resource listings, industry directories, partner pages, comparable expert commentary. Unreplicable: coverage of a funding round, a long-standing exclusive partnership, links to a competitor's proprietary dataset. Note the unreplicable ones as intelligence, not targets.
6. Reverse-engineer PR campaigns. Look for date clusters where a competitor gained many media links at once, then find the asset that earned them. Read it. Assess whether you can do something better or approach the question differently.
7. Identify their linkable assets. Sort competitor pages by referring domains. The top pages show what your market's publishers actually cite.
8. Find broken opportunities. Dead competitor pages that still hold links are reclamation targets if you can offer a genuinely better replacement.
9. Analyse anchor profiles for evidence of manipulation. A competitor whose profile is dominated by commercial exact-match anchors is not a model to copy.
10. Check geographic relevance, particularly for local and nationally-scoped businesses.
11. Prioritise using the scorecard in Section 32.

Competitor Link Gap Framework

[VISUAL SUGGESTION: Competitor link gap framework — four-quadrant matrix of replicability against value]

The most valuable output is usually not a prospect list. It is the realisation that a competitor's authority rests on one thing — a dataset, an annual survey, a genuine research programme — that you could match if you decided to.

Section 17: Backlink auditing

A backlink audit establishes what you have, what changed, and whether anything needs attention. Run one quarterly for most sites; monthly for large or actively campaigning ones.

Two disciplines matter more than the checklist.

First: lost and broken links are where the immediate value is. A link pointing at a 404 is authority you already earned and are discarding. Fixing it requires a redirect, not an outreach campaign. Most audits under-prioritise this because it's less interesting than finding threats.

Second: do not become obsessive about toxic links. This deserves its own section, immediately following.

Section 18: Toxic links — fact versus myth

"Toxic link" is an SEO industry term, not a Google classification. Google's documentation does not use it. It generally refers to links a practitioner or a third-party tool judges harmful — from spam networks, link farms, hacked sites, or irrelevant foreign-language domains.

Three facts to hold simultaneously:

1. Google says it handles most of this without your help. Google's documentation on link spam states that its systems are designed to ignore or neutralise unnatural links rather than penalise the targeted site, and Google representatives have repeatedly said the vast majority of sites never need the disavow tool. The reasoning is sound: if arbitrary third-party links reliably damaged a site, sabotage would be trivial and cheap.

2. A third-party toxicity score is not a Google judgement. Tools computing "toxicity", "spam score" or similar are running their own heuristics — anchor patterns, outbound link volume, site age, TLD, language. These can be a useful triage aid. They have no access to Google's evaluation. A high toxicity score does not mean Google considers a link harmful. Treating vendor scores as verdicts is the single most common cause of wasted effort in link auditing.

3. Manual actions do exist. Google's manual actions include unnatural links to a site, and Search Console reports them. This is the one case where the situation is unambiguous, because Google has told you.

When to investigate, when to ignore, when to disavow

On the disavow tool: it exists, Google maintains it, and it is genuinely useful in the narrow cases above. It is also capable of harm — disavowing legitimate links removes signals you earned, and the tool offers no undo of consequence. Google's own guidance positions it as an advanced feature most sites don't need. Use it surgically, or don't use it.

Section 19: Link spam and penalty risks

Google's spam policies for web search are public documentation and the authoritative statement of what constitutes manipulation. The link spam policy centres on a clear principle: links intended to manipulate rankings, rather than to serve readers, violate the guidelines.

Practices Google's link spam policy identifies, in plain terms:

On private blog networks: PBNs are a direct instance of automated and manipulative link creation. They are not a grey area, and this guide does not describe how to operate one.

On site reputation abuse: Google introduced a policy addressing third-party content published on a host site primarily to exploit that site's ranking signals — the practice sometimes called parasite SEO. If you place content on other domains, the question to ask is whether it is there because the host's audience benefits or because the host's domain ranks. The second is the policy violation.

Consequences, stated proportionately

Outcomes range from nothing at all — Google ignoring the links, which is the most common result — through algorithmic devaluation of the manipulated signals, to site-wide manual actions requiring removal and reconsideration.

Two errors to avoid in equal measure. The first is treating any link risk as catastrophic, which produces the compulsive disavowing described above. The second is treating the policies as unenforced, which produces link buying on the theory that everyone does it. Both are wrong. The realistic position: deliberate manipulation carries genuine risk that scales with how systematic and detectable it is, while ordinary web noise carries essentially none.

The honest strategic argument against link buying is not primarily fear of penalties. It is that purchased links tend to be poor links — from sites that sell to anyone, in content nobody reads, with no referral value — and that the money buys a depreciating asset you'd have to keep buying.

Section 20: Guest posting

Guest posting occupies contested ground because two very different activities share the name.

Legitimate editorial contribution: a genuine expert writes a substantive piece for a publication whose audience wants it, going through real editorial review, published under their name, with a link where it genuinely helps the reader. This is ordinary trade publishing and has existed since trade publishing existed.

Scaled guest-post link schemes: mass-produced articles placed across many sites primarily to embed keyword-rich links, often through networks selling placements, frequently on sites that exist to host such content. Google's link spam policy names large-scale article campaigns with links as a violation, and Google has enforced against guest-post networks repeatedly.

The differentiator is not the format. It is intent, quality, and the nature of the publication.

Quality checklist for guest contributions

• The publication has a real, identifiable audience in our field
• It publishes original editorial content beyond contributed pieces
• There is genuine editorial review and pieces get edited or rejected
• It does not advertise paid placements or "write for us — links included"
• Our named expert is genuinely qualified on the topic
• The piece would be worth publishing with no link at all
• Links are contextual and genuinely useful to the reader
• We are not producing near-identical pieces for multiple outlets
• No money changed hands for the placement
• We would be comfortable if the arrangement were public

"Guest posting is dead" is wrong, and so is treating it as a scalable link channel. Legitimate contribution works and produces good links — but slowly, at low volume, requiring real expertise and real writing. If your guest posting is scalable, it is probably the other kind.

Section 21: Broken link building

Broken link building finds dead pages that still have links pointing at them and offers a working replacement.

Workflow: identify resources in your field that have gone offline (defunct tools, closed publications, removed guides); pull the referring domains still linking to the dead URL; assess whether you have — or could build — a genuinely comparable replacement; then contact the linking sites, noting the broken link and offering your resource as a substitute.

Why it works: you are reporting a real problem. A broken outbound link is a defect the publisher would want to fix.

Limitations, stated honestly:

• It requires a genuinely comparable replacement. Offering a thin page as a substitute for a substantial dead resource fails and damages your standing with that publisher.
• Many linking pages are themselves abandoned, so nobody will act.
• Reply rates are low, and published "success rates" for this tactic vary so widely and depend so heavily on execution that citing a figure would be inventing one. Assess your own results across a first campaign and use those as your benchmark.
• It doesn't scale — each opportunity needs individual assessment.

Best used as a supplementary tactic against a specific list of high-value opportunities, not as a programme's foundation.

Section 22: Link reclamation

Reclamation recovers authority you already earned and lost. It is the highest return-on-effort activity in off-page SEO and the most consistently neglected.

Two operational notes. Run a full reclamation pass after every site migration, redesign or URL restructure — these are where links get silently orphaned at scale. And set up brand monitoring so unlinked mentions surface within days, while the piece is fresh and the writer still remembers it. A request made a week after publication succeeds considerably more often than one made a year later.

Section 23: Linkable assets

A linkable asset is content that earns references because it is genuinely useful to someone writing something else. The test is functional: would a writer, researcher or practitioner need to point at this?

The underlying mechanism. People link when linking makes their own work better — supporting a claim, giving readers something useful, or saving themselves an explanation. Assets that succeed do the linker a favour. Assets that fail are about the publisher rather than the reader.

Why most "linkable assets" earn nothing: they duplicate something that already exists and ranks. There is no reason to cite the eleventh guide to a topic when the first is better and already established. Originality is not a stylistic preference here. It is the mechanism.

Section 24: Relationship-based link building

Over any multi-year horizon, relationships produce more high-quality links than campaigns do — because relationships create repeated opportunity while campaigns create single events.

Where the durable relationships are:

• Journalists and editors covering your sector. Become the source they call when they need a comment on deadline. This is built by being fast, quotable, honest, and willing to say when something isn't your area.
• Industry publications. Regular, genuine contribution rather than one-off placements.
• Partners and integrators. Genuine commercial relationships produce genuine cross-references.
• Customers. Case studies and testimonials, which link naturally in both directions.
• Vendors and suppliers. Customer showcases and dealer listings.
• Professional associations. Committee work and event participation.
• Academic contacts. Research collaboration, guest teaching, data sharing.
• Peer communities. Practitioner groups where you contribute expertise rather than promotion.

The line to hold. Genuine relationships produce links as a by-product. Reciprocal arrangements — explicit or tacit — are what Google's policy on excessive link exchanges addresses. The distinction is whether the link exists because it helps a reader or because two parties agreed to trade. Two organisations that genuinely work together linking to each other is normal business. Two organisations linking to each other because they agreed to is a scheme, regardless of how it's framed.

Section 25: Social media and off-page SEO

Accurate position: links from major social platforms generally do not function as PageRank-equivalent ranking signals. Most are nofollowed or otherwise attributed, and Google representatives have consistently declined to characterise social signals — shares, likes, follower counts — as direct ranking factors.

That is a narrow statement, and it is routinely over-read in both directions.

What social media does do, indirectly but genuinely:

• Distribution. It puts content in front of people who might link to it — including journalists, who use social platforms professionally to find sources and stories.
• Discovery. Content circulating socially gets found by publishers who then reference it from their own sites, where the links do count.
• Journalist visibility. Being visible and responsive where reporters work makes you findable when they need a source.
• Brand awareness and search demand. Social exposure drives branded search.
• Profile presence. Consistent, complete social profiles contribute to the consistency of information about your organisation across the web, which supports entity resolution.

The distinction that matters: social activity is an amplification channel for off-page work, not an off-page signal in itself. Measure it as distribution — reach, referral traffic, and links subsequently earned from sites that discovered you socially — not as an SEO signal. Teams that report social shares as an SEO outcome are measuring the wrong thing.

Section 26: Forums and online communities

Communities — Reddit, Quora, Stack Exchange, sector forums, professional Slack and Discord groups — matter for reasons largely unrelated to links.

What they're genuinely good for: understanding how your market describes its problems in its own words, which is better keyword and content research than most keyword tools provide; building individual expert reputation; being discovered by people researching purchases; and increasingly, visibility in the third-party discussion that AI systems draw on when summarising a topic.

What they are not: a link channel. Community links are typically nofollowed or ugc-attributed, and dropping links into threads is comment spam under Google's policies and gets you removed by moderators regardless.

How to participate credibly: answer questions in your area of genuine expertise without linking; disclose your affiliation when relevant; link to your own material only when it directly answers the question and the community's norms allow it; accept that most contributions produce no measurable return; and treat the reputation of the individual expert, not the brand, as the asset.

Section 27: Podcasts, interviews and expert contributions

Appearing as a subject expert is among the more efficient off-page activities available to anyone with genuine expertise, largely because the supply of credible, articulate specialists willing to be interviewed is smaller than demand.

What an appearance produces: a link from the episode page, usually editorial and often long-lived; brand exposure to a self-selected relevant audience; quotable material that gets referenced elsewhere; positioning of a named individual as an authority, which supports the personal-entity signals discussed in Section 28; and syndication, since podcast episodes get written up, quoted and listed across multiple sites.

Practical notes: pitch a specific topic you can discuss with real depth, not a general offer to appear; target shows your buyers actually listen to rather than the largest available; ensure the episode page links to a relevant page, not just the homepage; and reuse the appearance — transcribe it, quote it, cite it — so one recording generates multiple references.

Expert commentary services and journalist request platforms operate on the same logic at higher frequency: a short, well-informed, timely response to a reporter's query can produce coverage in outlets no outreach campaign would reach.

Section 28: Off-page SEO and E-E-A-T

E-E-A-T — Experience, Expertise, Authoritativeness, Trustworthiness — comes from Google's Search Quality Rater Guidelines, the document human evaluators use when assessing search result quality.

Accuracy point that matters: E-E-A-T is not a direct ranking factor. There is no E-E-A-T score. The guidelines instruct human raters, whose assessments help Google evaluate whether ranking changes improve results; they do not describe a computed signal. Google has said its systems aim to identify signals that correlate with what raters recognise as quality. So E-E-A-T is best understood as a description of what Google is trying to reward, not a mechanism.

Off-page signals are heavily represented in how raters are asked to assess reputation, because reputation is by definition external. Third-party evidence that supports each component:

What to do with this. Ensure named authors have verifiable external footprints — professional profiles, published work, credentials that can be checked. Pursue references from sources already recognised in your field. Maintain accurate information in registries and licensing directories. Keep review profiles active. And make the connection between your site and these external signals explicit through consistent naming and structured data.

What not to do: buy awards, fabricate credentials, or add author bios claiming expertise that no external source corroborates. The whole point of these signals is that they are third-party. Manufacturing them internally defeats the mechanism and creates a discoverable inconsistency.

Section 29: Off-page SEO for AI search (evolving — expect this section to date fastest)

This section is deliberately isolated so it can be updated without touching the rest of the guide. It is also the section where the evidence is weakest, and it is written to say so.

What is reasonably well established

• AI answer systems cite sources. Google's AI features, ChatGPT with browsing or search, Perplexity, and Microsoft Copilot present answers with links to underlying sources. Being among the cited sources is a visibility outcome distinct from ranking.
• They draw on third-party content about you, not only your own. When these systems answer questions about a category or recommend providers, they draw on reviews, comparisons, industry coverage, forum discussion and listicles — content you don't control.
• Structured, unambiguous content is easier to extract. Clear definitions, explicit entity naming, factual statements and organised headings are easier for a retrieval system to lift accurately than allusive prose. This is a reasonable engineering inference and consistent with how retrieval works.
• Entity clarity helps. Systems that reason about organisations as entities benefit from consistent, corroborated information about who you are, where you operate, and what you do.

What is industry inference, not established fact

• That brand mention volume across the web influences AI recommendation frequency. Plausible, widely asserted, not demonstrated.
• That specific content structures reliably increase citation rates. Studies exist; methodologies vary; results are not stable across systems or over time.
• That "optimising for AI search" is a distinct discipline from producing authoritative, well-structured, well-referenced content. The evidence to date suggests substantial overlap rather than divergence.

What is unfounded speculation

• Any claim to know how a specific AI system weights sources. These are proprietary, they change frequently, and no public documentation describes ranking or citation selection in the way Google's documentation describes search.
• Any product promising guaranteed AI citations or "LLM ranking" positions.

What this implies for practice

The honest conclusion is that off-page fundamentals point in the same direction for AI visibility as for search visibility, which is convenient but should not be presented as a validated strategy:

1. Be the primary source. Systems that cite tend to cite whoever originated a finding. Original research is the strongest available position.
2. Ensure third-party content about you is accurate. Reviews, comparison pages, directory entries and industry listings are the substrate these systems summarise.
3. Maintain entity consistency. Contradictory information across sources produces uncertain answers about you.
4. Get referenced by sources these systems already draw on — established industry publications, reference sites, and well-regarded community discussion.
5. Write extractably. Definitions that stand alone, specific rather than vague claims, named entities rather than pronouns.
6. Measure what you can. Track brand mentions in AI responses to your priority queries manually or with emerging tools, and accept that measurement here is immature.

BrightLocal's recent local consumer research reports rapid growth in consumers using AI tools when researching businesses. Directionally this is consistent with other reporting; the specific figures are recent, single-source and moving quickly, so treat magnitudes as provisional.

[Internal Link: AI SEO Guide]

Section 30: Off-page SEO metrics

Reporting principle. Lead with referring domain growth, referral traffic, branded search and conversions. Those four connect off-page activity to business outcomes. Third-party authority scores make poor headline metrics for reasons covered next.

Section 31: Third-party authority metrics

These are not Google metrics. Google does not use Domain Authority, Domain Rating or Authority Score as ranking factors. Each is a proprietary model built by a software vendor from its own crawl, designed to approximate something Google does not disclose. Google has confirmed it has no public equivalent metric, and toolbar PageRank — the last public link score — was retired years ago.

Legitimate uses: rapid comparative triage across a prospect list; benchmarking your profile against actual SERP competitors; detecting large directional changes; communicating rough scale to non-specialist stakeholders.

Limitations: each depends on its vendor's crawl coverage, which differs; the scales are logarithmic, so a move from 20 to 30 is not comparable to 60 to 70; none captures topical relevance, which the research literature suggests matters more; and all are trivially inflated by tactics that add link volume without adding value.

The failure mode to avoid. When a score becomes a target, teams start acquiring links that raise the score rather than links that help the business — which is how link buying gets internally justified. Use these metrics as benchmarks, never as objectives, and never report them as the primary measure of an off-page programme.

Section 32: How to prioritise link prospects

A scorecard makes prioritisation explicit and comparable across a large prospect list. This is an internal planning tool. It does not represent Google's algorithm and does not predict ranking impact.

Scoring rules. Weight relevance and editorial quality most heavily. Treat spam risk as a veto rather than a component — any prospect scoring low there is removed regardless of its total. Set a threshold (a common approach is pursuing only prospects above roughly 60% of available points) and stop pursuing everything below it. The discipline of not chasing marginal prospects is where the time saving comes from.

Section 33: A 90-day off-page SEO workflow

[VISUAL SUGGESTION: 90-day off-page SEO roadmap — three-phase Gantt-style timeline]

Days 1–30: Research and audit

• Full backlink audit: referring domains, anchors, lost links, broken links, sitewide patterns
• Fix every backlink pointing at a 404 or a redirect chain — immediate recovered value
• Identify true SERP competitors for priority queries
• Export and categorise competitor referring domains; build the link gap analysis
• Audit local citations and NAP consistency, if geography is relevant
• Audit review profiles across relevant platforms
• Set up brand mention monitoring, including unlinked mentions
• Establish baselines for every metric in Section 30
• Score and prioritise an initial prospect list
• Decide which linkable asset to build, based on what your organisation uniquely holds

Days 31–60: Assets and outreach

• Build the primary linkable asset — research, tool, dataset or definitive resource
• Document methodology thoroughly enough for a journalist to verify
• Run the reclamation programme: unlinked mentions, lost links, incorrect links
• Execute citation cleanup and duplicate removal
• Launch the review generation process
• Begin Priority 1 outreach from the link gap analysis (proven, replicable prospects)
• Build the media list for the PR campaign: named journalists on the relevant beat
• Register with expert commentary and journalist request services
• Start supplier, partner and association outreach — usually the fastest wins available

Days 61–90: PR, relationships and measurement

• Launch the digital PR campaign around the completed asset
• Pitch, follow up once, and track coverage including syndication
• Reclaim links from unlinked coverage as it appears
• Pursue podcast and interview opportunities
• Continue Priority 2 and 3 prospects
• Re-audit citations and reviews for progress
• Report against baselines: referring domains, referral traffic, branded search, rankings, conversions
• Assess which channels produced results and reallocate accordingly
• Plan the next asset, informed by what actually earned references

Realistic expectation. Ninety days establishes the programme and delivers reclamation wins, citation fixes and early links. It does not deliver mature authority. Compounding effects arrive over quarters, and the plan's real output is a repeatable system rather than a finished result.

Section 34: Off-page SEO by business type

Section 35: Off-page SEO for local businesses

Local businesses deserve additional depth because their off-page priorities are genuinely different — for most, off-page SEO is local SEO.

Google Business Profile is the foundation. Complete every field; select the most specific primary category; keep hours accurate including holidays; add real photographs regularly; use Products and Services fields fully; answer questions. Note that the GBP Q&A section was retired in favour of AI-generated answers sourced from the business's website, which makes the content on your own site the input to what appears there — a direct link between on-page content and off-page presentation.

Reviews. Ask systematically at the point of satisfaction. Respond to all of them. Never incentivise or filter. Watch recency: consumers discount old reviews, and BrightLocal's research shows rising expectations around both rating thresholds and freshness.

Citations. Establish canonical NAP, build across major platforms and sector directories, and treat duplicate elimination as the priority — duplicates split reviews and confuse entity resolution.

Local press. Regional outlets need local stories and are considerably more accessible than national media. Expansion, hiring, community initiatives, and expert commentary on regional issues all qualify.

Community involvement. Sponsorships, event participation, charity work, local sports teams. These produce links, citations and word of mouth simultaneously, and they are the least copyable form of local authority.

Supplier and association relationships. Dealer locators, certified-installer lists, membership directories and trade body listings are typically the fastest legitimate local links available, and most businesses have several they've never claimed.

[Internal Link: Local SEO Guide]

Section 36: Twenty common off-page SEO mistakes

1. Buying links. Poor quality, depreciating, against policy, and rarely the cheapest route to the same outcome.
2. Chasing DA or DR alone. Optimising a vendor's estimate rather than your business.
3. Ignoring topical relevance. The factor the research literature suggests matters most, routinely traded away for a higher score.
4. Neglecting brand building. Links are easier to earn when people have heard of you; the causation runs both ways.
5. Over-using exact-match anchors. The clearest self-inflicted manipulation footprint.
6. Volume over value. Hundreds of weak links accumulate risk without accumulating authority.
7. Automated outreach at scale. Recognisable instantly, ignored routinely, and it burns prospects permanently.
8. No personalisation. "Dear Sir/Madam, I read your article" is a deletion trigger.
9. Ignoring lost links. Recovering an existing link is far cheaper than earning a new one.
10. Not tracking conversions. Off-page programmes get cancelled because nobody connected them to revenue.
11. Ignoring reputation. Reviews and complaints shape decisions regardless of ranking effects.
12. Irrelevant guest posts. A link from a site with no connection to your field on a site nobody reads.
13. Mass directory submission. Listings available to anyone convey the information that you filled in a form.
14. Systematic link exchanges. Explicitly identified in Google's link spam policy.
15. Dismissing nofollow links. They carry traffic, exposure and discovery, and Google treats attributes as hints.
16. Skipping local citations when geography genuinely matters to the business.
17. Treating PR as link acquisition. Journalists can tell, and it damages relationships worth more than the links.
18. Ignoring referral traffic. The clearest evidence a link has real value, generally unexamined.
19. Fake or incentivised reviews. Against platform policy, increasingly detected, and legally risky in several jurisdictions.
20. Publishing weak linkable assets. The eleventh guide to a well-covered topic earns nothing, however well written.

Bonus, and arguably the most expensive: copying the tactics of businesses with entirely different assets. A local plumber running data journalism, or a research-heavy B2B firm chasing directory listings, is spending in the wrong place.

Section 37: Fifteen off-page SEO myths

Section 38: The off-page SEO checklist

Strategy

• Priority queries and true SERP competitors identified
• Off-page objectives tied to business outcomes, not vendor scores
• Techniques matched to assets the organisation actually has
• Baselines recorded for every reported metric

Backlinks

• Referring domains audited for quality and relevance
• Every backlink pointing at a 404 or redirect chain fixed
• Lost links reviewed and reclamation attempted
• Anchor profile checked for implausible concentrations
• Prospect list scored and thresholded

Digital PR

• A story or dataset nobody else holds identified
• Methodology documented to a verifiable standard
• Media list built from named journalists on the beat
• Coverage tracked including syndication

Brand mentions

• Monitoring live for linked and unlinked mentions
• Unlinked mentions worked as a reclamation queue
• Organisation naming consistent across all properties

Local citations

• Canonical NAP defined and documented
• Major platforms and sector directories claimed
• Duplicates identified and merged or removed
• Re-audit scheduled after any move or rebrand

Reviews

• Systematic request process at the point of satisfaction
• All reviews responded to, including negative ones
• Presence on the platforms your buyers actually consult
• No incentives, filtering or fabrication anywhere in the process

Competitor research

• Referring domains exported and categorised
• Link intersections identified
• Replicable separated from unreplicable
• Competitor linkable assets analysed

Content assets

• At least one asset a writer would need to cite
• Genuinely original rather than duplicative
• Promoted actively to people with reason to reference it

Outreach

• Individually researched and personalised
• Value to the recipient stated first
• One follow-up maximum
• Nothing that requires a purchase or an exchange

Technical

• Links resolve to live 200-status pages
• Protocol and www redirects functioning
• Structured data for organisation and location accurate
• Post-migration redirect map rebuilt from the backlink inventory

Measurement

• Referring domains, referral traffic, branded search and conversions reported
• Third-party scores used as benchmarks only
• Attribution stated as correlation, not causation

Risk control

• No purchased links passing credit
• No systematic exchanges
• Advertising correctly attributed sponsored
• Search Console monitored for manual actions

AI visibility (evolving)

• Third-party information about the business verified accurate
• Key concepts written as standalone extractable statements
• Entity information consistent across authoritative sources
• AI responses to priority queries sampled periodically

Section 39: A recurring off-page SEO routine

Section 40: How long does off-page SEO take?

No fixed timeline exists, and any specific promise should be treated as a warning sign. Off-page outcomes depend on variables largely outside the practitioner's control.

What determines the timeline: competitive intensity in your market; your site's existing authority and history; whether you have content worth referencing yet; market size and the density of relevant publications; the rate at which you can realistically earn links; whether you have any newsworthy story or data; existing brand recognition and relationships; whether technical and on-page foundations are sound; the type of search intent you're competing for; and how established the incumbents are.

A realistic sequencing, offered as a pattern rather than a forecast:

• Weeks 1–4: reclamation and citation fixes can produce measurable recovery quickly, because you are recovering signals you already had.
• Months 2–3: first earned links from outreach and relationship work; early PR campaigns land or don't.
• Months 4–6: accumulating referring domains and, in less competitive markets, visible ranking movement.
• Months 6–12: compounding effects as assets accumulate references and relationships produce repeat coverage.
• Beyond 12 months: the position that actually differentiates — recognised authority in a field — which is why it can't be bought.

Claims to reject: guaranteed rankings, guaranteed timelines, guaranteed link counts at guaranteed authority levels. Nobody controls third-party publication decisions or search ranking outcomes, and anyone selling certainty about either is selling something else.

Special section: What has changed and what has not

The constant across all of it: search systems have always tried to identify sources that others independently recognise as worth referencing. Every technique in this guide is a way of becoming such a source. Every technique Google penalises is a way of faking having become one.

Special framework: The Off-Page SEO Authority Pyramid

[VISUAL SUGGESTION: Authority pyramid — seven ascending levels with the base labelled as prerequisite]

This is an educational strategic framework for sequencing work. It is not Google's ranking model and does not describe any search system's architecture.

How to use it. Work upward, and don't skip. Teams that attempt Level 5 without Level 2 run PR campaigns with nothing to point at. Teams that stall at Level 3 accumulate links while a competitor builds recognition they can't match. The pyramid's real argument is that the durable levels are the top ones, and they take the longest — which is exactly why they're defensible.

Special framework: The Link Opportunity Matrix

Exceptions that justify overriding the matrix:

• Local relevance overrides topical relevance for geographically-served businesses. A regional newspaper is low topical relevance and high value.
• Referral value can justify a low-authority link if the audience converts. A small community forum sending qualified buyers beats a large irrelevant site.
• Relationship potential can justify a marginal first placement if it opens a channel to a publication that matters.
• Prestige has non-SEO value. Coverage in a nationally recognised outlet supports sales, recruitment and credibility whatever it does for rankings.

Special framework: The Off-Page SEO Moat

This is the most strategically important idea in the guide, and the one most off-page programmes never reach.

Most SEO advantages are temporary because they're copyable. A competitor can rewrite their titles, fix their site speed, restructure their internal links, and publish a comparable guide. Within a quarter or two, on-page and technical advantages converge across a competent market.

Off-page authority is different, because the most valuable forms of it cannot be copied quickly at any price.

The strategic conclusion. A link programme produces links. A moat produces links continuously, as a by-product, while also producing referral traffic, brand search, AI visibility, sales credibility and pricing power.

The practical question to ask in any off-page planning session is not "how do we get more links this quarter?" It is: what could we become the definitive source for, that a competitor could not become within two years? Then build that, and let the links follow — which they reliably do, because that is the condition link-based ranking was designed to detect in the first place.

Section 42: The final strategic framework

Modern off-page SEO is not the accumulation of links. It is the deliberate construction of reasons for reputable third parties to mention, cite, reference, review, recommend, link to, quote, include and discuss you.

Every legitimate technique in this guide is an instance of that. Every practice Google's spam policies prohibit is an attempt to skip it.

 Earn Attention ↓ Create Value ↓ Build Relationships ↓ Earn Mentions ↓ Earn Links ↓ Build Reputation ↓ Increase Authority ↓ Strengthen Search Visibility

Read it in reverse and the logic becomes clearer: search visibility follows authority, authority follows reputation, reputation follows mentions and links, mentions and links follow relationships and value, and all of it follows attention that was earned rather than purchased.

Off-page SEO done properly is not a marketing channel. It is what happens when an organisation becomes genuinely worth referencing, and then makes sure the right people know it exists.

Frequently asked questions

What is off-page SEO? Off-page SEO is the practice of building and managing signals outside your own website that influence how search engines and users evaluate it — backlinks, brand mentions, business citations, reviews, media coverage and third-party references. Its defining feature is that other people control publication.

What is the difference between on-page and off-page SEO? On-page SEO covers what you publish on your own site: content, headings, internal links and metadata. Off-page SEO covers signals published elsewhere: links, mentions, citations and reviews. You control on-page work fully; you can only influence off-page outcomes.

Are backlinks still important? Yes, but less dominantly than a decade ago. Google's Gary Illyes said in 2023 that links are not in the top three ranking signals and have not been for some time, while adding that they remain important and that their significance is often overestimated. Links still matter for authority and discovery; they are no longer the whole game.

How many backlinks do I need? There is no target number. Ahrefs found roughly two-thirds of pages have no backlinks at all, so a small number of relevant, authoritative links is often enough to compete. The useful benchmark is the profile of the pages currently ranking for your specific queries, not an absolute figure.

What makes a backlink high quality? Editorial placement, topical relevance to your subject, a source with genuine authority and readership, relevance at page level and not just domain level, coherent surrounding context, and natural anchor text. Editorial intent and relevance matter more than any authority score.

Does Domain Authority affect rankings? No. Domain Authority is a Moz metric, as Domain Rating is Ahrefs' and Authority Score is Semrush's. All three are third-party estimates built to approximate link strength. Google does not use them as ranking factors. They are useful for comparison, not as objectives.

Are nofollow backlinks useful? Yes. Since 2019 Google treats nofollow, sponsored and ugc as hints rather than directives, so it may still use the information. Independently, nofollowed links deliver referral traffic, brand exposure and discovery. Declining a relevant high-visibility placement because it will be nofollowed is usually a mistake.

Does social media help SEO? Indirectly. Google has not characterised social shares or likes as direct ranking factors, and most social links are attributed. But social platforms distribute content to people who may link to it, help journalists find sources, and drive branded search. Treat social as amplification for off-page work, not as an off-page signal.

Are brand mentions ranking signals? Unconfirmed. Google has not stated that unlinked brand mentions function as a direct ranking factor. What can be said is that mentions build reputation, drive discovery, support entity recognition, generate branded search and create reclamation opportunities — all valuable without the algorithmic claim.

What is link equity? Link equity, sometimes called link juice, is the informal term for the authority a link transfers from one page to another. The concept derives from PageRank. No public formula describes what Google actually computes today, so treat it as a useful mental model rather than a measurable quantity.

What is anchor text? Anchor text is the clickable text of a hyperlink. It describes the destination in words chosen by someone other than the destination's owner, which is why search engines have used it as a relevance signal and why it became a manipulation target. Descriptive, varied, naturally occurring anchors are the goal.

Can bad backlinks hurt my rankings? Usually not. Google's documentation states its systems are designed to ignore or neutralise unnatural links rather than penalise the target site. The real risk arises from links you created — bought links or schemes — which can attract algorithmic devaluation or a manual action. Ordinary spam and scraper links accumulate on every site and require no response.

Should I disavow spammy links? In most cases, no. Google positions the disavow tool as an advanced feature most sites never need. The genuine use cases are a manual action for unnatural links reported in Search Console, or a known history of link buying. A third-party toxicity score is not a reason to disavow, and disavowing legitimate links removes signals you earned.

Is guest posting safe? Legitimate editorial contribution to relevant publications is safe and useful. Scaled article campaigns placed primarily to embed keyword-rich links violate Google's link spam policy. The distinction is intent, quality and publication legitimacy — not the format. If your guest posting scales easily, it is probably the wrong kind.

What is digital PR? Digital PR is the practice of creating newsworthy stories — usually original research, data analysis or expert commentary — and placing them with journalists and publishers. It produces coverage, links, brand awareness and search demand at once, and is generally the most reliable route to genuinely high-authority links.

What are local citations? Local citations are online references to your business's name, address and phone number. Structured citations appear in directory formats; unstructured citations appear within ordinary content like news articles or event listings. Consistency and the elimination of duplicates matter more than raw citation volume.

Do reviews help SEO? For local search, Google's local ranking guidance identifies prominence as a factor and references review quantity and rating among its inputs. For general organic web rankings, there is no confirmed direct effect. Reviews' largest measurable impact is on consumer behaviour: BrightLocal found around 75% of consumers always or regularly read reviews when researching a business.

How do I find competitor backlinks? Identify the domains actually ranking for your priority queries, export their referring domains from a backlink tool, categorise each by type, then find domains linking to multiple competitors but not to you. Those intersections are your highest-probability prospects because the source has already shown willingness to link within your category.

How long does link building take? Reclamation and citation fixes can show results within weeks. First earned links typically arrive within two to three months. Compounding authority takes six to twelve months or longer, and recognised industry authority takes years. No credible practitioner guarantees a timeline, because nobody controls third-party publication decisions.

Does AI search change off-page SEO? It appears to increase its importance rather than reduce it. AI answer systems synthesise third-party sources — reviews, coverage, comparisons, community discussion — so what others publish about you shapes what these systems say about you. The fundamentals point the same direction: be the primary source, keep external information accurate, and maintain consistent entity signals. Specific mechanics remain undocumented, and claims to know how any AI system weights sources should be treated sceptically.

Can competitors hurt my site with bad links? Google states its systems handle unnatural inbound links without harming the target site, and documented cases of successful third-party link sabotage are rare. Monitor your profile, document anything unusual, and check for manual actions in Search Console — but the appropriate response is usually observation rather than intervention.

Key takeaways

• Off-page SEO is broader than link building. Links, brand mentions, citations, reviews, media coverage and entity signals all sit within it, and links are one component rather than the whole discipline.
• Its defining constraint is that you don't control it. That absence of control is what makes the signals informative and why techniques promising control are either low-value or non-compliant.
• Link quality is dominated by editorial intent and topical relevance. Both were established in research between 1998 and 2002 and have aged better than any metric built since.
• Links matter less than they did, on Google's own account. Illyes placed them outside the top three ranking signals in 2023 while affirming they remain important. Plan accordingly rather than at either extreme.
• Referring domains beat total backlinks as a measure of independent recognition, and roughly two-thirds of pages have no backlinks at all — the bar is lower than it appears.
• DA, DR and Authority Score are vendor estimates, not Google metrics. Use them as benchmarks; never as objectives.
• Most link risk is self-inflicted. Google says it ignores most unnatural inbound links. The genuine exposure comes from schemes you ran, not spam that found you.
• Digital PR is the most reliable route to high-authority links, and it fails when treated as link acquisition rather than as media relations.
• Reclamation is the highest return-on-effort activity available and the most consistently skipped.
• The durable advantage is a moat, not a link count. Original research, proprietary data, real expertise and genuine relationships cannot be copied within a planning cycle. Everything else can.