HARO Link Building: The Queue Didn't Die, It Started Banning People
HARO outlived its own platform twice. What people were buying was never the product, it was a queue of pre-qualified intent, and that queue is now actively defended.
HARO bans an account every 23 minutes. That is not a metaphor, it is the title of a post its operator published on 10 September 2026. Run the rate out over a year and it is roughly 22,800 accounts removed from a service most of the SEO industry still describes as dead.
Dead things do not need moderation teams.
HARO link building is the only tactic I know that outlived its own product twice. It got bought, buried, sold and resurrected, and people are still arguing about whether it works. That is a strange thing for a channel to do, because channels do not normally survive the death of the thing that carries them.
It survived because almost nobody was buying the platform. They were buying one structural property, and that property is portable.
What people called HARO link building was never a platform. It was a queue of pre-qualified intent.
A journalist writing "I need a source on X by Thursday" is doing something that almost never happens in outreach: declaring a need, in writing, in advance, with a deadline attached. That is the rarest input in link building. Everything else is guesswork about whether someone wants what you have.
So the death of HARO was a distribution problem, not a tactic problem. The queue did not evaporate. It fragmented, came back, and then started defending itself.
What actually happened to HARO: the verified timeline, not the folklore#
Most posts on this keyword get the chronology wrong, so let us fix it with primary sources.
HARO was founded as a Facebook group in 2008 by Peter Shankman, acquired by Vocus in June 2010, and passed to Cision when Vocus merged with it in 2014. That chronology is the one part of this post resting on a secondary source, Wikipedia, and you should weight it accordingly. Everything after 2024 is quoted from the companies themselves.
Cision rebranded HARO to Connectively, then killed it. Cision's own notice is unambiguous: "As of December 9, 2024, the Connectively (HARO) platform has been permanently discontinued".
That is where most SEO content stops. It is out of date by more than a year and a half.
On 16 April 2025, a PR Newswire release datelined "SCOTTSDALE, Ariz., April 16, 2025" announced that "Featured.com, a platform that connects subject-matter experts with journalists, announced today that it has acquired Help A Reporter Out (HARO) from Cision." The same release sets out the model: HARO "will remain free for journalists and sources, supported by newsletter ads," and will "return to its original format of delivering daily opportunities via email." Cision confirmed the sale from its own side the same week.
HARO is not dead. It died in December 2024, was sold in April 2025, and runs today as a free email.
The official HARO blog confirms the current shape: it is "the free, three-times-a-day newsletter you've counted on," with Connectively repositioned as "HARO's sister brand," the platform tier for people who want filtering, response tracking and a way to manage opportunities in one place.
How to do HARO link building in 2026, in five steps#
If you came here for the method rather than the argument, here it is in full. Nothing is held back for later.
- Subscribe to more than one queue. Start with HARO's free newsletter, three emails a day, then add Qwoted, SourceBottle and MentionMatch so no single platform owns your pipeline. The alternatives section below covers what each one is actually good for, checked this month.
- Pitch as a named expert, not as a brand. Reporters quote people with a title and a track record. Decide who that person is before you subscribe, and keep a standing note of the four or five things they can speak to credibly.
- Triage hard on arrival and answer very few. Kill any request where the outlet does not link, where the brief is so vague it will draw four hundred replies, or where you could not survive one follow-up question. Three or four real answers a week beats forty templated ones, and the qualification filters are below.
- Answer the question first, credentials second, in your own words. Lead with the actual answer, put the proof under it, keep it to a few short paragraphs. HARO's operator reports that 21% of pitches are now entirely AI-generated and that 35% of journalists block those outright, so anything that reads as generated is competing in the one category the buyer is filtering out.
- Work the follow-up inside the deadline. Send from a mailbox with real sending history rather than a cold domain, follow up once before the deadline closes and never after it, and answer the reporter's clarifying question in minutes rather than days. This is where most landed placements quietly die.
That is the method. The rest of this post is why each of those steps is what it is, and the evidence that the queue changed underneath the advice everywhere else on this keyword.
Why the declared need, not the platform, is the asset in HARO link building#
Here is the structural claim, and it is the whole post.
Ordinary outreach is a guess. You find a page, decide it should link to you, and email someone who has expressed no interest in hearing from you. Your reply rate is a function of how well you guessed.
A journalist request inverts that. The other party published their criteria before you arrived. You are not persuading them a need exists, only proving you fit one they already wrote down.
Published acceptance criteria are the highest-leverage input in link acquisition, and journalist requests are one format of them.
If that sounds familiar, it should. It is the same property that makes resource page link building work: a page saying "we link to tools that do X, submit here" is publishing its acceptance criteria in advance. Broken link building is a weaker version of the same idea, where the need is implied by a 404 rather than declared in a sentence.
Seen as a category rather than a product, the HARO obituaries stop mattering. The question is never "is HARO alive." It is "where are people currently writing down what they need, and can I read it faster than the competition."
This is also why it belongs under digital PR link building rather than under link schemes. The link is a byproduct of being cited as a source.
Worth being precise there, because the accusation comes up constantly. Google's spam policies define link spam as "the practice of creating links to or from a site primarily for the purpose of manipulating search rankings," and the named examples include "Exchanging money for links, or posts that contain links," "Excessive link exchanges," and "Using automated programs or services to create links to your site."
Answering a reporter's published question with genuine expertise is not on that list. Mass-generating answers to questions you are not qualified to answer arguably is, under "Using automated programs or services to create links to your site." The line is real and it is not where most people put it.
The HARO queue is not dead, it is being actively defended#
Now the part nobody in SEO is writing about, because it was published three weeks ago.
The original queue had three properties: free, public, and unqualified entry. Any one alone is fine. All three together is a tragedy of the commons with a deadline attached. Free plus public plus unqualified means the cost of a pitch trends toward zero while the value of a journalist's attention stays fixed. Once templated and then automated pitching arrived, volume per query rose until reading the queue stopped being worth a reporter's time.
Cision's Connectively answer was friction via price, moving to a paid credit model. It did not save the product. Featured's answer was the opposite: stay free, and spend the money on enforcement instead.
We now have numbers on what that enforcement looks like, from the operator, covering a year of traffic. Featured's Brett Farmiloe published an analysis of "more than 2 million pitches and 100 million emails sent through HARO" on 31 August 2026, and a follow-up on moderation, bluntly titled after the ban rate, on 10 September 2026.
Four figures matter.
21% of all pitches on HARO are 100% AI-generated. Against the two million pitches in the same analysis, that is on the order of 420,000 fully automated pitches in a single year, into a queue whose entire value proposition is human expertise.
35% of journalists opt out of receiving AI pitches entirely. Over a third of the demand side has already pulled the lever the platform gave them.
32% of journalist requests are rejected during the approval process, after review of the journalist, the publication and the request. The filtering runs in both directions, which is the part that distinguishes today's HARO from the 2019 free-for-all.
HARO bans an account every 23 minutes. Roughly 22,800 accounts a year, on my arithmetic rather than theirs, and the operator is explicit about the profile: accounts that have "built entire operations around sending large volumes of responses to HARO opportunities."
And the outcome figure, the one that should reset how you budget this channel: human-written pitches are "roughly three times more likely to get featured than AI-generated ones."
Note what Featured is betting on. Not that pitching gets cheaper, but that verified human expertise gets scarcer and therefore more valuable. Farmiloe's line at the acquisition was "As AI floods the internet with generic content, journalists need credible, human sources more than ever," and the enforcement numbers are that sentence turned into a budget.
Align your process with the bet. The tactic has not got harder, it has got narrower, and the narrowing is being done by someone with an incentive to protect the value of the queue you want access to.
Where journalist requests live now: HARO alternatives I actually checked#
I loaded each of these before naming it. Where a site is behind a bot challenge, I say so.
HARO. Live, owned by Featured.com, free for journalists and sources, three emails a day per the official blog. The root domain sits behind a bot checkpoint, but the blog subdomain is publishing actively.
Featured.com. Live. Positions itself as "Your co-pilot for PR" and invites you to "Log in to search, save, and respond to journalist requests, podcast invites, speaking gigs, and more." Broader than media queries.
Connectively. Now Featured's platform tier rather than a Cision property, described on the HARO blog as the option for people who want filtering, response tracking, a clean interface, or a way to manage opportunities in one place. It recently passed 100,000 users, with HARO's newsletter list materially larger than that.
Qwoted. Live. Headline reads "Stop chasing sources. Start choosing them," with a free join tier. Skews business, finance and tech, with stricter journalist verification than open-email models.
SourceBottle. Live and free, and useful precisely because it is not crowded by the US SEO industry. It describes itself as "Free publicity for expert sources and PR professionals," with visible weighting toward Australian outlets.
MentionMatch, formerly Help a B2B Writer. This one moved and most listicles have not noticed. helpab2bwriter.com now redirects to mentionmatch.com, which describes itself as "Connecting B2B writers with expert sources." If a HARO alternatives post still points at the old domain, it has not been checked since the rebrand.
Muck Rack. Live, but a different animal, and it blocks automated requests so I could not read it directly. It positions around a media database and outbound outreach rather than an inbound query feed, which puts you back to guessing who wants your story.
#JournoRequest on X and Bluesky. Both feeds resolve. Least structured, and for that reason the least contested, because it is not aggregated into the tools everyone else is reading. I have no reliable figure for volume or hit rate here and I am not going to invent one.
How to qualify journalist requests before you spend a pitch#
The queue moved, got smaller and got policed, so qualification now matters more than volume. Three filters, in order.
Is the outlet real, and does it link. Plenty of queries come from content mills that quote you and strip the link, or from sites with a blanket nofollow policy on contributor mentions. Check an existing article before you answer, not after.
Is the request specific enough to be genuine. "Looking for marketing experts" is a fishing expedition that will draw four hundred replies. "Looking for a CFO at a Series B SaaS company who has moved off annual budgeting" is a person with a half-written article and a hole in it. The narrower the request, the better your odds, which is the inverse of how most people sort their queue.
Can you actually answer it. Not "can you produce text about it." If your answer would not survive a follow-up question from the reporter, do not send it. That is the filter automated pitchers skip, and per HARO's own moderation numbers it is now the filter that gets accounts removed.
Answer fewer requests, from people who could plausibly quote you, with things only you could have said.
The operational pain is not finding requests. It is that a good pitch takes fifteen minutes of real thought, deadlines are short, and the follow-up loop is where the value leaks. Replies land in a shared inbox, nobody answers the reporter's clarifying question about your title or your headshot, and a landed placement quietly becomes nothing. That leak is the entire reason GenLink exists: it keeps an opportunity alive from the pitch to the published link, so a deadline you were going to make never dies in an inbox nobody opened.
That is worth being concrete about, because it is the half of this job that does not scale with effort. Finding the request is free. Writing the substance is irreducibly yours. Everything in between — identifying the right contact at the outlet, sending from a mailbox with real sending history rather than a cold domain that lands in spam, following up once inside the deadline and not after it, and triaging the reply so the reporter's question gets answered in minutes instead of days — is mechanical, and it is where most placements die.
The fifteen minutes of real thought is yours. Everything after it should not be. GenLink replaces the shared pitch inbox nobody owns, the follow-up reminder you set and then miss, and the cold sending domain that lands you in spam before a word gets read. You write the answer. It keeps the opportunity alive until the link is.
The case against HARO link building, argued in the open#
I would rather put the counter-evidence in the post than let you find it later.
The strongest argument against the whole exercise is that the prize is shrinking. The Reuters Institute's Journalism, Media and Technology Trends and Predictions 2026, based on a survey where "280 people completed a closed survey between 18 November and 20 December 2025" across 51 countries and territories, reports that publishers expect traffic from search engines to almost halve, a 43% fall, over the next three years.
Their Chartbeat data is measurement rather than forecast: Google traffic from organic search to over 2,500 sites was down by a third globally between November 2024 and November 2025, and by 38% in the United States, though the report is careful that it is not clear how much of this is down to AI overviews.
The second argument comes from Google. Gary Illyes said at a search conference in Bulgaria in 2024, as reported by Search Engine Journal, "We need very few links to rank pages… Over the years we've made links less important." His follow-up was "I shouldn't have said that… I definitely shouldn't have said that," which is a confirmation, not a retraction.
Take both seriously. If search referral volume is falling and links are a weaker signal than they were, the case for spending an expert's afternoon on media queries genuinely weakens.
Here is my counter, and it comes from the same independent report. Publishers plan to invest more in "original investigations and on the ground reporting" at a +91 percentage point difference between "more" and "less," "contextual analysis and explanation" at +82, "human stories" at +72 and fact-checking and verification at +63, while scaling back general news at -38, evergreen content at -32 and service journalism at -42, which many expect to become commoditised by AI chatbots.
The categories being cut are the ones that never needed a source. The categories being funded cannot be written without one.
That is a queue getting smaller and much better qualified at the same time, which is the same conclusion HARO's moderation data reaches from the other direction. It is also why the durable payoff is being the named, quoted expert inside original reporting rather than a link in a roundup. That survives whether the reader arrives via a blue link, an AI answer, or memory.
What this evidence does and does not support
Be clear about the limits, because the strongest numbers in this post come from an interested party.
Featured.com owns HARO. Its figures on AI pitch share, ban rate, request rejection and the three-times feature advantage for human pitches are first-party, unaudited, and describe only its own platform. They are also commercially convenient: "human expertise is scarce and we police the queue" is precisely the story that makes its free newsletter worth advertising against. Treat them as the operator's account of its own service, not as an industry measurement. I use them because there is no independent census of journalist-request queues and because the direction they point is corroborated by a source with no stake in HARO at all, the Reuters Institute, whose survey and Chartbeat data are the strongest evidence in this post that cuts against the tactic as well as for it.
The dates, ownership and pricing above are documented on live primary pages and quoted directly. The claim that declared intent outperforms cold prospecting is a structural argument I find persuasive, not a measured result, and I have no controlled study to hand you.
No source cited here measures conversion from a journalist request to a live backlink. Anyone quoting you a precise rate for that is guessing.
The 22,800-bans-a-year figure is mine, not HARO's: it is the published one-per-23-minutes rate extrapolated across a year, and it assumes the rate is steady, which nobody has claimed.
Building a HARO workflow that survives the next shutdown#
Design for the fourth shutdown, because there will be one.
Subscribe to more than one queue and never let a single platform own your pipeline. HARO, Qwoted, SourceBottle, MentionMatch and the journalist hashtags all break differently, and the last two years prove any one of them can be discontinued by a parent company on short notice.
Build a named expert, not a brand account. Reporters quote people. The asset that survives every platform migration is a person with a track record and a title, which is what makes the resulting coverage useful inside a wider link building programme rather than a side quest.
Keep a standing library of what your expert can credibly speak to, with proprietary data where you have it. Speed only helps if the substance is already assembled, and with three emails a day and short deadlines, the assembly is the bottleneck.
Treat this as one lane rather than the whole strategy. The skills transfer directly, because the qualification discipline you learn reading journalist requests is the same discipline that makes cold link building outreach work. In both cases you are proving fit against criteria rather than arguing someone into a new need.
The platform was never the asset. Learn to read declared need and you can rebuild the pipeline wherever it reappears.
Frequently asked questions#
Is HARO still free in 2026?
Yes. Featured.com's acquisition release committed to keeping it free for journalists and sources, funded by newsletter advertising rather than subscriptions, and the official HARO blog reconfirmed the free three-times-daily newsletter. The paid tier that caused most of the backlash belonged to the Cision-era Connectively product, which is a different thing from today's newsletter.
What is the difference between HARO, Connectively and Featured now?
All three sit under Featured.com and they are deliberately layered. HARO is the free email digest, Connectively is the platform experience with filtering and response tracking for people who want more than an inbox, and Featured is the broader opportunity aggregator covering podcasts, speaking slots and awards alongside media queries. Pick based on whether your bottleneck is discovery or management.
Can I use AI to write pitches to journalists?
Use it to draft, structure and tighten. Do not use it to invent the substance. HARO's operator reports that 21% of pitches are entirely AI-generated, that 35% of journalists have opted out of receiving AI pitches, and that human-written pitches are roughly three times more likely to be featured. A pitch whose value is generated rather than sourced is competing in the one category the buyer is actively filtering out, and at scale it is the profile that gets accounts banned.
How do I find journalist requests without a platform at all?
Monitor the hashtags directly on X and Bluesky, subscribe to the newsletters of the specific reporters on your beat, and watch for the "sources wanted" posts editors publish on LinkedIn and in industry Slack groups. It is slower and less complete, which is exactly why it is less contested than the aggregated feeds everyone else reads.
Do journalist request links actually pass value, or are they nofollow?
It varies by outlet and you should check before pitching rather than after. Many large publishers apply nofollow or UGC attributes to contributor mentions as policy, while trade and mid-tier outlets more often link normally. Treat the attribute as a qualification input alongside relevance, and remember that a nofollow citation in a serious publication still does work for brand recall and for how people and models describe you.
How long should I wait before following up on a pitch?
Match the reporter's clock, not yours. If the query carried a deadline, your window closes when it does, and a follow-up afterwards is noise rather than persistence. Inside an open deadline, one short follow-up that adds something new is reasonable. A second usually is not.
Why do my pitches get ignored even when I am qualified?
Usually deliverability first, specificity second. A cold domain sending a burst of pitches lands in spam before anyone judges the content, and a pitch that opens with your credentials rather than the reporter's question reads as a template even when it is not. Lead with the answer, put the proof underneath, and send from a mailbox with a real sending history.
Is HARO link building worth it if links matter less than they used to?
The honest answer is that it depends what you are optimising for. If you want raw link volume, this was never the efficient channel and it is less so now. If you want to be the named expert inside the category of journalism publishers are actively funding, it is one of the few repeatable ways in, and that citation keeps working in AI answers and in reader memory whether or not the link passes ranking credit.
Expertise is the only part of this job that cannot be delegated. Almost everything else can. Hand GenLink the chasing — finding the right person at the outlet, sending from a mailbox with real history instead of a cold domain, the one follow-up that lands inside the deadline rather than after it, the reply triaged the same hour — and keep the fifteen minutes that actually win the citation. The free tier runs one agent, so you can point it at a single campaign before you commit to anything.
