Link Building & Digital PR in 2026: How to Earn Links That Last

Administrator in Search Engine Optimization Jul 12 / 5 days ago

In February 2025, a B2B SaaS founder shared his content budget with me. One folder held invoices for 127 "guest posts" bought through 2024, priced between $150 and $400 each. Total spend came to just over $19,000. Ahrefs showed the links. Google Search Console showed the truth. Rankings for his money pages had not moved at all. Not one position.

Here is the part that stung. Nothing was penalized. There was no manual action, no warning, no dramatic crash. Google's spam systems had simply decided those links counted for nothing. He paid full price for zero effect. That is what modern link spam enforcement looks like, and it is why so many teams feel like their link building stopped working somewhere around 2023.

The good news is that earned links still move rankings, and in my experience they move them harder than ever, because fewer competitors are earning them. This guide covers what changed, what still works in 2026, and how to run link building campaigns that produce links Google actually counts.

What you will get from this pillar guide. By the end, you will know how to plan a digital PR campaign, build a statistics page that attracts citations for years, use the source-request platforms that replaced the old HARO workflow, qualify prospects by relevance instead of vanity metrics, and measure results in referring domains and assisted rankings rather than raw link counts. You will also get a few opinions that cut against common advice:

• Most failed link campaigns in 2026 are not penalized. They are silently devalued, which is worse because you never get a signal to stop spending.
• Relevance beats authority scores. A niche trade blog often passes more useful signals than a generic high-score site.
• The best link building no longer looks like link building. It looks like PR, product work, and research.
• You need fewer links than you think. You need better pages more than you need more links.

This is the strategy pillar. If you want software recommendations, we keep a separate, regularly updated roundup of the best link building tools for 2026, so I will stay out of tool reviews here.

Why did link building change after Google's spam updates?

Google shifted from punishing link schemes to ignoring them. The December 2022 link spam update extended SpamBrain, Google's AI-based spam system, to detect sites that buy links and sites that exist to sell them. Since then, most manipulative links are neutralized rather than penalized. Your rankings do not tank. Your investment just evaporates.

That distinction matters more than any other fact in this guide. In the Penguin era, roughly 2012 to 2016, bad links could drag a whole site down, so people feared them. Fear kept some discipline in the market. Devaluation removed the fear but also removed the payoff. Google described this shift openly in its December 2022 link spam update announcement, where it said SpamBrain can detect both sites buying links and sites used for the purpose of passing outgoing links.

Layer the March 2024 core update on top, which folded helpful-content signals into the core system and hit low-value content networks hard, and the old playbook collapsed. Guest-post farms lost their audiences and their equity. Niche-edit marketplaces sold placements on pages Google had already stopped trusting. I watched an affiliate site spend around $2,000 a month on niche edits through 2024 with a referring-domain chart going up and a traffic chart staying flat. The links existed. The value did not.

So the question for 2026 is not how to acquire links. It is how to deserve them, then make deserving them efficient. That is what link building means now, and the rest of this guide treats it like the strategy discipline it is.

What counts as link spam in Google's eyes now?

Anything that exchanges value for a ranking-passing link without disclosure is link spam. Google's spam policies documentation is blunt about it. Buying or selling links that pass ranking credit, excessive link exchanges, automated link-building programs, and requiring links in terms of service all qualify.

A few cases trip people up in practice:

• Paid placements are fine if disclosed. A sponsored article is legitimate marketing when the link carries a rel attribute of sponsored or nofollow. It becomes spam when it passes ranking credit. In the US, the FTC's endorsement guides also require disclosure of paid placements, and the UK's ASA enforces similar rules, so undisclosed paid links carry legal risk on top of SEO risk for US and EU brands.
• Guest posting is not dead, but guest posting for links is. Writing for a publication your buyers read, with an author bio and editorial standards, builds the experience and authority signals we cover in our guide to E-E-A-T and helpful content in 2026. Mass-producing posts for sites that publish anyone is what SpamBrain was built to catch.
• Reciprocal links happen naturally. Two relevant sites citing each other once is normal. Spreadsheet-managed "you link me, I link you" programs at scale are the "excessive exchanges" Google names explicitly.

My rule of thumb is simple. If a link would still make sense with rankings turned off, it is safe. If its only purpose is rankings, assume Google will eventually treat it as noise.

Do backlinks still matter in the age of AI search?

Yes, and they now do double duty. Links remain a meaningful ranking input in classic search, and the sites AI systems cite most are overwhelmingly sites with strong earned-link profiles. Ahrefs' well-known search traffic study found that the vast majority of pages get no organic traffic from Google at all, and a lack of referring domains is one of the main reasons the authors point to.

AI search changes the shape of the game rather than the rules. ChatGPT, Perplexity, and Google's AI Overviews synthesize answers from sources they trust, and trust correlates with the same things earned links signal, namely that real publications vouch for you. That is why digital PR has become the rare tactic that pays off in both systems at once. A data story covered by Axios or The Guardian earns a followed link for classic rankings and a brand citation that generative engines learn from.

There is a related shift worth understanding on its own. Mentions without links now carry visibility value in AI answers, which we unpack in our post on why brand mentions are becoming the new backlinks in AI search. And if you want the full picture of optimizing for generative engines, our guide to ranking in AI search covers that discipline end to end. The strategic takeaway for this pillar is that every link building campaign you run should be judged on links earned and citations earned, together.

What is digital PR, and why does it out-earn classic outreach?

Digital PR means creating stories journalists want to cover, then pitching them like a publicist rather than an SEO. Instead of asking a blogger to insert your link, you give a reporter at a real publication a reason to write about you. The links arrive as citations inside coverage, which is exactly the pattern Google's systems treat as natural, because it is natural.

The economics surprise people. A solid digital PR campaign in the US or UK typically costs somewhere between $5,000 and $20,000 through an agency, based on rate cards I have seen from mid-market firms in both countries. That sounds expensive next to $200 guest posts until you compare outcomes. One campaign that lands 20 to 40 referring domains from news sites your audience has heard of will usually outperform a year of bought placements that Google ignores. In my own client work, a single survey-based story in 2025 earned 31 referring domains in six weeks, more than the previous 14 months of manual outreach combined.

Why does it work so well now? Three reasons:

• Journalists need material. Newsrooms are smaller than they were a decade ago, and reporters under deadline genuinely want credible data and expert comment.
• The links are unfakeable. Competitors cannot buy their way into equivalent coverage, so the advantage compounds.
• It builds the authority layer. Press coverage feeds the trust signals we describe in our piece on building authority that Google and AI systems trust.

My honest caveat is that digital PR fails often. Industry practitioners commonly report that a large share of campaigns land few or no links. The next section is about stacking the odds.

How do you run a data-driven digital PR campaign?

Start with the journalist's job, not your keyword list. A reporter needs a story that is new, surprising, timely, and relevant to their beat. Your campaign is the raw material. Here is the process I use, with realistic time estimates:

1. Find the angle (1 to 2 weeks). Read the publications you want coverage in. Note what their journalists covered this quarter. Your angle should extend a conversation they are already having, not start a new one.
2. Generate the data (2 to 4 weeks). Options include commissioning a consumer survey through a panel provider, analyzing your own product data, or aggregating public datasets like Eurostat or US Census figures into something nobody has compiled before. Internal product data is the most underrated source because only you have it.
3. Package it (1 week). A landing page with the headline findings, clear methodology, and quotable numbers. Journalists cite pages that make their job easy.
4. Build the media list (3 to 5 days). Tools like Muck Rack or Prowly help you find reporters by beat. Fifty well-matched journalists beat five hundred random ones every time.
5. Pitch and follow up (2 to 3 weeks). Short email, finding first, offer of exclusive data or expert comment, one polite follow-up.

The classic proof that this model works is OkCupid's old data blog, which turned dating statistics into national news coverage for years. Zillow does it with housing data, Glassdoor with salary reports, NerdWallet with personal finance studies. None of them think of it as link building. All of them earn more links than almost anyone who does. For the outreach and monitoring software that supports this workflow, from BuzzStream to Pitchbox, see our link building tools roundup rather than this guide.

Why do original research and statistics pages earn so many links?

Because every writer on the internet needs a citation, and Google rewards writers who cite sources. When someone writes "email marketing statistics" into a search box, they are usually about to link whatever credible page they find. Owning that page means earning links passively, every week, without outreach.

A statistics page works when it meets three bars. The numbers must be current, the sourcing must be transparent, and the page must be scannable. I have seen stats pages earn hundreds of referring domains over a few years, and I have seen copies of those pages earn nothing, because aggregating other people's aggregations adds no value. The strongest version includes at least some numbers that are yours alone, from a survey, your product, or your community. If you run forums or reviews, the audience data angle pairs well with our user-generated content strategy guide, since community-sourced data is both content and citation bait.

Two operational notes from experience:

• Refresh on a schedule. A 2024-dated stats page bleeds citations to fresher rivals. We cover the mechanics of keeping pages current in our guide to content refreshing, pruning, and decay. Put stats pages on a quarterly review cycle.
• Cite primary sources with dates. Pew Research, Statista, government data. Writers check whether your numbers hold up before linking them. Broken sourcing kills the compounding effect.

Expect the first links within one to three months of ranking, then a long tail. This is the slowest tactic in this guide and, in my opinion, the highest total-return one.

How do expert commentary and source requests work after HARO?

The HARO model survived its own corporate death, and it still earns strong editorial links. Quick history, because the landscape confused everyone. Cision rebranded HARO into a platform called Connectively, then shut Connectively down in late 2024. Peter Shankman, who founded the original HARO, had already launched a free successor called Source of Sources. The HARO name itself returned in April 2025, when Cision sold the platform to Featured.com, which relaunched the service. Meanwhile, Qwoted and Featured built their own journalist-and-expert marketplaces, and Help a B2B Writer serves business niches specifically.

The workflow is the same across all of them. Journalists post requests for expert comment. You reply with a short, quotable answer and credentials. When quoted, you usually get a link or at least a named mention, and both matter now.

What separates people who win placements from people who waste hours:

• Speed. Reporters often pick from the first useful replies. Responding within a couple of hours beats responding tomorrow.
• Specificity. One concrete number or example from your real work beats three paragraphs of generic advice. Give the quote they can paste.
• Genuine credentials. A "marketing enthusiast" gets skipped. A person with a title, a company, and demonstrable experience gets quoted. This is E-E-A-T applied to your people, not just your pages.

My own hit rate across these platforms has hovered around one placement per ten to fifteen quality pitches, and I would treat anyone promising much better than that with suspicion. Budget 30 to 45 minutes a day and treat it as a steady drip, not a campaign.

What makes a linkable asset worth building?

A linkable asset is a page that solves a problem so cleanly that other sites reference it instead of explaining it themselves. Free tools, calculators, templates, glossaries, and original datasets all qualify. This site is itself an example of the model. Utility pages like a free domain authority checker or a domain age checker earn references from tutorials and roundups because writers would rather link a working tool than describe a manual process.

Before you build, apply the test I use with clients. Search what you plan to create. If ten good versions exist, yours needs a tenfold improvement or a genuinely new angle to earn anything. If nothing good exists and people clearly need it, even a modest version can collect links for years. A niche mortgage-overpayment calculator aimed at UK borrowers, or a salary-conversion tool for cross-border EU hires, will out-earn a generic word counter every time because it serves a specific audience nobody else serves well.

Three placement details decide whether an asset earns its keep:

• Do not bury it. Assets need internal prominence to get crawled, ranked, and found. Our guide to site architecture and crawl depth explains why depth kills discoverability.
• Surround it with a cluster. Supporting articles that answer related questions make the asset the hub of a topic, a structure we map out in our content clusters and pillar pages guide.
• Scale carefully. Spinning one calculator into five hundred thin variants is how assets become liabilities. Our post on programmatic SEO without penalties covers where that line sits.

How do you reclaim unlinked brand mentions?

Someone already wrote about you and forgot the link. Asking for it is the highest-conversion outreach that exists. The writer already knows you, already found you worth naming, and adding a link costs them thirty seconds. In my experience, reclamation emails convert at several times the rate of cold pitches, and the only reason more teams skip it is that nobody owns the monitoring.

The workflow is light. Set up alerts for your brand name, product names, and executives' names. When a mention appears without a link, check whether a link would genuinely help that page's readers. If yes, send a two-line note thanking the writer and pointing to the most useful destination, which is often a specific resource rather than your homepage. If the mention is on a high-authority page, prioritize it. If it is on a scraper site, ignore it.

One nuance for 2026. Not every unlinked mention needs converting anymore. Mentions themselves now feed AI-search visibility, so a healthy mention footprint has standalone value even before reclamation. We cover that dynamic in depth in our post on brand mentions as the new backlinks, so I will not re-cover it here. The strategic point for this pillar is sequencing. Reclamation only scales with the size of your mention footprint, which means digital PR and reclamation form a flywheel. PR creates mentions, reclamation converts them, and both feed authority.

Do broken-link building and resource pages still work in 2026?

They work at a smaller scale than the blog posts of 2015 promised, and only when you do the curation for the site owner. Broken-link building means finding dead outbound links on relevant pages and offering your live resource as a replacement. Resource-page outreach means asking curated "useful links" pages to include you. Both survive because they align incentives. The site owner gets a better page, and you get a link.

Here is the honest math. Reply rates on this kind of outreach are low, and conversion is lower. I consider a campaign successful if 3 to 5 percent of contacted pages add the link, and I have run campaigns that did worse. What separates the campaigns that clear that bar:

• Your replacement must actually be better. If the dead resource was a detailed guide and yours is a 600-word summary, you are asking someone to downgrade their page. They will not.
• Point to the exact spot. Name the page, the section, and the dead destination. Site owners will not hunt for the problem you found.
• Only pitch true topical matches. A resource page for university writing centers does not want your marketing tool, no matter how good it is.

I treat these as opportunistic tactics, not programs. When crawling turns up a relevant dead link, capture it. Building a whole quarter around it is a 2015 strategy running on 2026 reply rates.

Where do local and niche citations fit in?

For any business with a location or a defined industry, citations are the cheap, safe base layer of the link profile. Chambers of commerce, industry associations, accreditation bodies, local news, event sponsorships, and legitimate directories all produce links that Google expects a real business to have. None of them will vault you up competitive rankings on their own. Their job is corroboration. They confirm you are a real entity in a real place serving a real market.

For local businesses in the US and Europe, these citations work alongside your Google Business Profile, and consistency of name, address, and phone details across them still matters. We cover that whole stack in our local SEO and Google Business Profile guide. For B2B and niche sites, the equivalent is trade-body memberships, industry award listings, and speaker pages at real conferences. One genuine association link beats twenty generic directory submissions.

A note for international operators. If you serve multiple countries, earn citations in each market's own ecosystem, in its own language where possible. A German .de industry portal does more for your German visibility than another US directory. This also applies to high-value markets like South Korea, where local platforms and Naver shape discovery, and where your international setup needs the technical foundation we describe in our hreflang and international SEO guide. Links are geographic signals too, and most teams forget that.

How should you qualify link prospects in 2026?

Relevance first, real traffic second, authority scores a distant third. I think sorting prospect lists by Domain Rating or Domain Authority is the most overrated habit in SEO. Those are third-party estimates from Ahrefs and Moz, not Google metrics, and link sellers have learned to inflate them. If you are fuzzy on what these scores measure and what they do not, our explainer on what Domain Authority actually is sets the record straight.

My qualification checklist, in order:

1. Topical relevance. Would this site's audience plausibly care about yours? A DR 35 trade blog read by your actual buyers beats a DR 80 general "business" site every time, for rankings and for referral value.
2. Organic traffic. Does the site rank for anything? A site with authority scores but no rankings is usually a link farm wearing makeup.
3. Editorial standards. Real authors, real editing, outbound links that make sense. A "write for us" page promising fast publication is a red flag, not an opportunity.
4. Outbound-link hygiene. If the site links to casinos, essay mills, and payday loans from unrelated articles, it sells links, and Google likely knows.
5. Then, and only then, authority metrics as a tiebreaker. A quick pass through our domain authority checker plus a look at site age with the domain age checker helps you rank an already-qualified list.

Ten minutes of qualification saves weeks of outreach to prospects that were never going to help you.

How do you write outreach that actually gets replies?

Short, specific, and obviously written by a human who read the page. Every journalist and site owner in the US and Europe now receives a steady stream of templated, AI-generated pitches, and their spam radar has never been sharper. The bar for standing out has paradoxically dropped to "sound like a person."

The structure that keeps working for me:

1. A subject line that states the value, not the ask. "New data: 62 percent of UK renters delayed moving in 2025" outperforms "Collaboration opportunity" by a mile.
2. One line proving you read their work. Reference the specific article, honestly. Fake flattery reads as fake.
3. The value in two sentences. The finding, the resource, or the fix, plus why it fits their beat or their page.
4. A single low-friction ask. One link, one quote, one question. Multiple asks kill replies.
5. One follow-up, three to five business days later, then stop. Persistence past two touches converts almost nothing and burns your domain's sending reputation.

A practical benchmark from my own campaigns and from what practitioners share publicly. Reply rates of 5 to 10 percent on cold link building outreach are respectable, and 15 percent means your targeting is excellent. If you are below 3 percent, your list is bad, your pitch is generic, or your asset is not worth linking, and it is usually the asset. Fix the asset before you buy more email credits.

What should you never do in link building now?

Anything that leaves a paper trail of manipulation, because SpamBrain is very good at reading paper trails. The definitive list lives in Google's spam policies, but here is my practitioner translation of what to refuse in 2026:

• Private blog networks. PBNs are networks of sites that exist to link out. Footprint detection has gotten relentless, and when a network is caught, every link it ever passed is void. Renting a PBN link is renting a countdown timer.
• Link exchanges at scale. The "ABC swap," where site A links B, B links C, and C links A to hide reciprocity, is a named violation. Occasional natural reciprocity is fine. Managed swap programs are not.
• Paid followed links. Sponsorships and paid placements are legitimate with a sponsored or nofollow attribute on the link. Paying for followed links violates Google's policies and, undisclosed, US and EU advertising rules too.
• Bulk guest posting on general-purpose sites. If a site publishes articles about plumbing, crypto, and skincare from different "contributors" every day, it is a link farm with a content management system.
• Mass automated outreach. Beyond being ineffective, scraped-and-blasted email campaigns create deliverability damage that outlasts the campaign, and in the EU they can raise GDPR questions about how you sourced the addresses.

The pattern behind every item is the same. Shortcuts that used to be risky are now merely worthless, and worthless at real cost. The budget they consume is the budget that should fund the link building strategies above, the ones that actually earn.

How do you measure a link building campaign?

Measure referring domains, link quality, and assisted rankings, not raw link counts. One hundred links from three domains is three votes, not one hundred. Here is the measurement stack I set up for every link building campaign:

• New referring domains per month, from Ahrefs, Semrush, or Majestic, filtered to followed links from pages with real traffic. This is your primary volume metric.
• Quality distribution. What share of new domains are topically relevant sites your buyers might actually read? I aim for at least half. A campaign that produces thirty irrelevant domains failed even if the chart looks good.
• Assisted rankings. Track the target pages' positions in Google Search Console for their priority queries, and compare movement against similar pages that received no links. Links typically take four to twelve weeks to show ranking effects in my experience, so judge campaigns at the quarter, not the week.
• Referral and citation value. Real coverage sends real visitors, and increasingly it earns citations in AI answers. Log both.

Two things multiply the value of every link you earn. First, distribute the equity. A page that attracts links should pass authority to the money pages that need it, which is an internal problem, and our internal linking strategy guide covers exactly how to route it. Second, make sure the linked pages deserve the rankings the links enable, using our on-page SEO checklist. For the analytics and monitoring software behind all of this, our 30 best SEO tools roundup has current picks and pricing.

Where should you start this quarter?

Back to that SaaS founder with $19,000 in dead invoices. We stopped all purchased placements, spent one month building a statistics page from his product's anonymized usage data, and pitched the findings to fifteen trade journalists. It earned 11 referring domains in the first quarter, and his two most important pages moved from page three to the bottom of page one. Fewer links than he bought. Infinitely more effect.

If you take one priority from this pillar, take this sequence. First, set up brand-mention monitoring and reclamation this week, because it is nearly free. Second, commit to one linkable asset or original-data piece this quarter, chosen by searching for the gap first. Third, build a daily source-request habit on the HARO successors. Everything else, including broken-link opportunism and citations, slots in around those three.

My prediction for the next two years is that the gap widens. As AI-generated content floods the low end, editorial links and cited data become scarcer and therefore more decisive, in classic rankings and AI answers alike. The teams that build earning machines now will be very hard to catch. What is the one asset in your niche that everyone would cite if it existed? That is your next campaign.

Frequently Asked Questions

Is link building still worth it in 2026?

Yes, but only the earned kind. Links remain a meaningful ranking signal in Google, and strong link profiles correlate with being cited by AI search systems. What stopped being worth it is purchased and traded links, which SpamBrain now largely devalues. Shift the same budget into digital PR, original data, and linkable assets and the return is typically far higher.

How many backlinks do I need to rank?

There is no universal number. It depends on what the current top ten pages for your query have earned. Check the referring domains of ranking pages and treat the median as a rough bar. In low-competition niches, a handful of relevant links can be enough. In competitive US and EU niches, you may need dozens of quality referring domains plus strong content and on-page work.

Are paid links always against Google's rules?

No. Paying for placement is legitimate advertising when the link is marked with a rel attribute of sponsored or nofollow, so it passes no ranking credit. The violation is buying or selling links that pass ranking signals. Undisclosed paid placements can also breach FTC guidance in the US and advertising rules in the UK and EU, so disclosure protects you twice.

What replaced HARO for expert quotes?

Several platforms serve the same journalist-request workflow. Qwoted and Featured run active marketplaces, Help a B2B Writer covers business niches, and HARO founder Peter Shankman runs the free Source of Sources newsletter. HARO itself relaunched in April 2025 under Featured.com, which bought the platform from Cision after Cision shut down Connectively in late 2024. Pick one or two, answer fast with specific, quotable expertise, and expect roughly one placement per ten to fifteen good pitches.

Do nofollow links have any value?

Yes, indirect but real. Google treats nofollow as a hint rather than an absolute directive, and nofollow coverage from major publications still drives referral traffic, brand awareness, and mentions that AI systems learn from. A natural link profile also contains a healthy mix of followed and nofollow links. Do not chase nofollow links, but do not dismiss coverage because of them.

How long does it take for new backlinks to affect rankings?

Plan on four to twelve weeks in most cases. Google must recrawl the linking page, process the link, and recalculate signals, and effects often build gradually rather than jumping. Links to pages on established sites tend to show impact faster than links to new domains. Judge campaigns quarterly, and track target-page positions in Google Search Console rather than watching daily.

Is guest posting dead?

Guest posting as a link scheme is effectively dead. Google's systems recognize sites that exist to publish paid contributor content, and those links get devalued. Guest posting as genuine thought leadership is alive and useful. Writing for a respected publication your buyers actually read builds authority, referral traffic, and E-E-A-T signals, with any link being a bonus rather than the point.

What is a good reply rate for link outreach?

For cold outreach, 5 to 10 percent replies is respectable and 15 percent is excellent. Conversion to an actual link usually runs lower, and 3 to 5 percent of contacted prospects linking is a solid campaign. Brand-mention reclamation converts far better because the writer already knows you. If replies sit under 3 percent, fix your targeting or your asset before scaling volume.

Should I disavow bad links in 2026?

Usually not. Google says its systems ignore most spammy links automatically, and John Mueller has repeatedly discouraged routine disavowing. The disavow tool remains relevant mainly if you have a manual action for unnatural links or a documented history of large-scale link buying. For everyday sites, random spam links pointing at you are noise, not a threat worth hours of file maintenance.

How much does digital PR cost?

Agency campaigns in the US and UK commonly run from around $5,000 to $20,000 or more per campaign, depending on research scope and media targets. In-house costs less in cash and more in time, since survey panels, data work, and pitching take weeks. Judge cost per quality referring domain rather than sticker price. One strong campaign often beats a year of low-grade placements.