Ask 500 SEO professionals what's working in link building and you get a clear answer: digital PR, by a wide margin. Ask 897 journalists what makes them delete a pitch without reading past the subject line and you get an equally clear answer: almost anything that isn't relevant to their beat. Both surveys are right. The gap between them is where most link building budget currently goes to die.
Methodology
Reporter Outreach fielded its State of Link Building survey to 500 working SEO professionals in 2026, covering which tactics respondents consider best-performing, what they pay per link, how budgets moved year over year, and how AI search factors into current thinking. It is a self-reported perception survey: it measures what practitioners believe is working, not a controlled measurement of actual ranking or citation outcomes. Muck Rack's State of Journalism report, published March 19, 2026, is a different instrument entirely: it surveyed the people receiving pitches, not the people sending them, on what makes a pitch land, what makes it get deleted, which platforms they trust, and how AI has changed their own workflow. Read separately, each is a normal industry survey. Read together, one measures the seller's theory of the market and the other measures the buyer's actual behavior, and the difference between the two is the most useful part of either data set.
Neither survey set out to be read this way. Reporter Outreach is a link building vendor surveying its own market, and its numbers should be read with that context attached, the same way any vendor-run survey deserves a grain of salt about incentive to flatter the category it sells into. Muck Rack has the opposite incentive: it sells media relations software to the same PR and communications teams sending the pitches its own survey reports 88% of journalists delete, which is an unusually candid number for a vendor to publish about its own customers' output. That asymmetry is part of why the two data sets are worth reading together rather than either one alone: the buyer-side survey is optimistic in the direction you'd expect, and the recipient-side survey is blunt in a direction that costs its publisher nothing to admit and, if anything, makes the case for better targeting tools.
What SEO pros say is working
Digital PR is now the dominant link building method by adoption, used by 67.3% of the marketers Reporter Outreach surveyed, and it is not close on performance either. Asked to name their single best-performing tactic, 34% of respondents picked digital PR, nearly twice the 18% who picked guest posts and well ahead of the 21% who picked HARO-style journalist request platforms. That ranking lines up with what Something Inc. has argued about earned media's durability relative to platform-gatekept placements: a link a journalist chooses to include, because the story is genuinely useful to their readers, survives algorithm changes and AI-answer shifts in a way a paid or templated placement does not.
| TACTIC | NAMED BEST-PERFORMING | WHAT THE JOURNALIST-SIDE DATA SAYS ABOUT IT |
|---|---|---|
| Digital PR | 34% | Matches how journalists say they actually select stories: relevance to their beat, not outreach volume |
| HARO / journalist request platforms | 21% | Depends entirely on matching a live request; misses on relevance get deleted at the same 88% rate as any other pitch |
| Guest posts | 18% | Weakest performer in the SEO-side data, and the format journalists in this survey did not single out as a preferred structure |
The pattern holds up against the site's own past coverage of the same underlying dynamic. Digital PR and link building for the AI era has consistently found that AI engines weight third-party corroboration, coverage a brand did not control the wording of, more heavily than owned or paid placements. A journalist choosing to include a link because the underlying claim is genuinely newsworthy is the mechanism that produces exactly the kind of citation-worthy corroboration that both classic SEO and generative-engine visibility now reward. Guest posts and templated HARO submissions do not carry the same signal, because the selection bar on the other end is lower and the resulting content reads that way to both search engines and AI answer systems.
What it costs, and where it's going
Pricing has moved up across the board. 76% of link buyers surveyed now pay $300 or more per placement, and 31% are in the $500 to $1,000 range, a tier that would have been considered premium-only pricing a few years ago. Budgets followed: 58% of respondents increased their link building budget in 2026 compared to 2025, against only 14% who cut back, and 75% expect costs to keep climbing over the next two years. None of this is a market in decline. It is a market where buyers are paying more for a narrower set of tactics they trust, which tracks with digital PR's dominance in the performance data above: fewer, better-targeted placements at a higher price point rather than volume-based link acquisition at scale.
Link building budget and pricing signals, 2026 (Reporter Outreach survey of 500 SEOs)
The rising price is worth reading as a signal about scarcity, not inflation for its own sake. If the best-performing tactic is coverage a journalist chooses to grant rather than a placement a vendor sells outright, the supply of that coverage is capped by how many genuinely newsworthy stories a brand can produce and how many journalists are covering that beat, not by how much budget gets thrown at outreach volume. Our own analysis of link building budget allocation reached a similar conclusion from the AI-citation side: brand-controlled, durable channels are worth paying up for precisely because they cannot be bought at scale the way a directory listing or a paid guest post slot can.
What journalists actually do with a pitch
The Muck Rack side of the data is where the SEO-side optimism runs into a harder floor. 86% of the 897 journalists surveyed say PR pitches inspire at least some of their stories, which is the number that keeps digital PR budgets flowing. But 88% say they delete pitches that miss their beat, regardless of how well-written or well-targeted the pitch otherwise is. Those two numbers describe the same profession from two angles: pitches genuinely do become stories, and the single fastest way to lose a journalist's attention is sending them something that has nothing to do with what they cover.
The AI-adoption number is worth sitting with on its own. A journalist using ChatGPT or Gemini as part of their research process is a journalist who can, and increasingly will, verify a pitch's central claim before deciding whether it is worth their time, which raises the cost of an exaggerated or thinly sourced pitch beyond what it was two years ago. It also means the earned-media coverage that already dominates AI citation share is being produced by reporters who are themselves running the story's claims through the same class of tools that will later decide whether to cite the resulting article. A pitch built on a stretched statistic is now more likely to get caught before it ever becomes a placement, not just penalized after the fact.
Where belief and behavior diverge
The most useful finding in either survey is not in either survey alone, it's in the gap between them. On the SEO side, 74% of respondents believe backlinks affect AI search visibility. But only 24% are actually tracking whether that belief holds up, and just 19% report having changed their tactics in response to it. That is a profession that has updated its opinion faster than its instrumentation, which is a familiar pattern to anyone who has watched a metric get widely believed before anyone built the dashboard to check it.
That last point is the practical payoff of reading these two surveys together. A team that measures only its own send volume and placement count cannot distinguish between a genuinely weak digital PR program and a well-run one that keeps missing journalists' actual beats. The journalist-side data supplies the missing half of that diagnosis: if placements are down, the first question is not whether digital PR still works as a category, since the SEO-side survey and this site's own earned media research both say it does, the first question is whether the pitch list was ever actually built around what each specific journalist covers.
There is a second, quieter gap worth naming: the 74% who believe backlinks affect AI visibility and the 82% of journalists now using AI tools in their own workflow are describing the same shift from two different chairs, without either survey connecting them. A journalist running a claim through ChatGPT before covering it is applying exactly the kind of scrutiny that determines whether the resulting article later gets cited by the same class of tools. The SEO-side belief that backlinks matter for AI visibility is not wrong, it is just measuring the wrong end of the pipeline. The mechanism running underneath both numbers is the same one this site has argued for in earlier link building coverage: AI engines and the journalists whose coverage feeds them are converging on the same standard for what counts as a citable claim, and a pitch that would survive an AI fact-check is now also the pitch most likely to survive the journalist's own AI-assisted vetting before it ever becomes a placement.
The audit to run this quarter
None of this argues for abandoning guest posts or HARO-style platforms outright; both still produced real placements in the survey data, just at a lower reported hit rate than digital PR. It argues for spending the next quarter's link building budget the way the two surveys, read together, actually recommend: fewer, better-targeted pitches built around what a specific journalist covers, priced at what the market has already settled on, measured against tracked outcomes instead of an assumed one. For a B2B team weighing where the next quarter's earned-media budget goes, that is a more defensible allocation than optimizing for send volume and hoping the 88% deletion rate lands on someone else's list.
It is worth restating the limits of what two surveys, however large, can actually establish. Neither data set traces a specific placement to a specific ranking or citation outcome; both describe beliefs and self-reported behavior, not a controlled experiment. The 500-SEO sample and the 897-journalist sample were drawn independently, on different timelines, using different sampling methods, and nothing here claims they describe a matched pair of senders and recipients on the same pitches. What the two surveys support is narrower and still useful: a consistent picture, from both sides of the transaction, that relevance to the recipient's actual beat is the dominant variable in whether a pitch becomes a placement, and that the tactic performing best in the buyer-side data, digital PR, is also the one whose underlying mechanic, a journalist choosing to cover something because it's genuinely useful to their readers, matches what the recipient-side data says actually drives that choice. That alignment, not either survey's headline number in isolation, is the finding worth building next quarter's program around.
See where you are cited today
A free snapshot audit of your rankings and AI citations before we ever talk.
Josh leads work at the intersection of SEO and generative engines at Something Inc., helping B2B brands get ranked and cited across every major AI engine.