Expert quotes are one of the strongest GEO moves in the research, but HARO is gone. Here is the 2026 stack for earning them and making them extractable.
The single strongest content change in the foundational GEO research was not a schema tag or a word count. It was adding a named expert quote. In the KDD 2024 GEO paper, Quotation Addition was the top-performing optimization method across a 10,000-query benchmark, improving position-adjusted word count by 41% and subjective impression by 28% over baseline. An answer engine reaches for a sentence it can lift cleanly, and a quote with a name, a title, and a claim attached is the cleanest unit on the page. The problem in 2026 is that the platform most operators learned to earn those quotes on no longer exists.
HARO shut down on December 9, 2024, and its replacement stack looks nothing like the single daily email operators relied on for a decade. Signals runs an aged Reddit account marketplace plus an editorial network for AI brand mentions across Reddit, Quora, Product Hunt, and Threads, so we watch which source surfaces AI engines actually reuse. Before the tactics, one honest caveat: the GEO paper's lift is a document-use result, measured after a source is already in the model's context. It proves quotes get reused once your page is retrieved. It does not prove a quote alone earns discovery.
Expert quotes work because they hand the model a discrete, attributable fact it can quote without inventing context. The GEO paper (Aggarwal et al., KDD 2024) tested nine content changes on a 10,000-query benchmark. Quotation Addition ranked first, lifting position-adjusted word count by 41% and subjective impression by 28% against the unoptimized baseline. Cite Sources and Statistics Addition followed closely. The pattern is consistent: concrete, sourced, named content gets reused; vague assertion does not.
Read the number carefully. The GEO benchmark measured how much a source was used and credited once it was already retrieved into the answer context. It is a citation-quality signal, not a crawling or ranking signal. That distinction is the difference between a tactic that works and a slogan that overpromises. A quote from a named expert makes your page more likely to be quoted when it is in the running. It does not, by itself, put the page in the running. Discovery still depends on authority, retrieval placement, and freshness.
HARO is gone. Peter Shankman launched Help A Reporter Out in 2008; Vocus acquired it in 2010, Cision absorbed it, rebranded it Connectively, and then shut Connectively down entirely on December 9, 2024. The single thrice-daily email that trained a decade of digital PR no longer sends. If your quote-earning playbook still starts with "reply to the HARO email," it has been dead for over a year.
Two things filled the vacuum. Shankman relaunched the original format as Source of Sources, a free, email-based service that serves 20,000-plus journalists and deliberately copies HARO's early rhythm. Separately, Featured acquired the HARO brand name from Cision and revived "HARO" in April 2025 as a free daily newsletter attached to its own curated platform. So "HARO" now points to a Featured product, not the original service. For an operator, the takeaway is simple: the brand you remember is a marketing label now, and the real work happens across several tools.
There is no single replacement. The four platforms that produce the most editorial quotes in 2026 each cover a different gap, and the right answer for most teams is to run two or three in parallel rather than betting on one. Below is how we weigh them for GEO-oriented placement, where the goal is a named, attributable quote on an authoritative domain.
Source of Sources. Peter Shankman's HARO successor, free and email-based, closest to the original three-queries rhythm. Best for operators who want zero-cost volume and do not mind mixed journalist quality. The fastest way to rebuild a HARO-style habit without a subscription.
$0Qwoted. A modern reactive platform with real-time alerts, structured pitch management, and search. Consistently reported as the highest-conversion paid option. Best when you want to work fewer, better-matched queries instead of firing off replies to everything.
FreemiumFeatured. Owns the revived HARO brand. Uses a curated question model that filters spam and favors longer expert answers, which suits thought-leadership placements. Best for depth over speed, with a free entry tier and paid plans above it.
Curated#JournoRequest on X. Not a platform, a live hashtag. Fastest response window of any option because you reply directly to the reporter. Best as a monitoring layer, not a primary channel; noise is high and you filter it yourself.
Real-timeTreat these as a stack. Source of Sources rebuilds volume, Qwoted raises conversion, Featured lands the longer expert pieces, and the hashtag catches time-sensitive requests the newsletters miss.
Win the placement by writing the quote the reporter can paste without editing. Journalists working a HARO-style query are on deadline and reading dozens of pitches. The response that gets used is short, on-topic, and already in publishable form: two or three sentences, a specific number or example, and a clean attribution line with your name, title, and company. Lead with the answer, not with your bio.
Three habits separate placements from ignored pitches. Respond inside the first hour, because early, complete answers get read before the reporter fills the slot. Answer the exact question asked, not the question you wish they asked. And give a concrete claim a generalist could not, since specificity is what makes a quote quotable. "Founders underprice by anchoring to competitors instead of value" beats "pricing is very important." The same trait that wins the journalist, a discrete and attributable claim, is what an AI engine later extracts from the published article.
Host your own expert quotes with the same discipline you use to earn third-party ones. Earned placements build authority off-site; on-site expert quotes give the engine something to lift when your page is already retrieved. Both matter, and the on-site version is fully under your control. The rule is that a quote only counts if it is attributed in text near the claim, not buried in an image or a video caption an LLM cannot parse.
Format the quote as an extractable unit. Put the speaker's name, title, and organization in the same sentence or the line directly adjacent, so the model does not have to guess who said it. Keep the quoted claim self-contained: it should still make sense if lifted out of the page. Pair it with a short answer capsule at the top of the section and clean 120 to 180 word chunks so the surrounding structure does not fight the extraction. A named quote inside a wall of text is far less useful to an engine than the same quote given room to breathe.
Budget mostly time, not money. The tools themselves are cheap or free: Source of Sources costs nothing, the X hashtag costs nothing, and paid tiers on Qwoted and Featured generally run in the low tens of dollars per month. The real cost is the operator hour spent reading queries and writing tight, specific pitches daily. Expect a low hit rate at first; conversion climbs as you learn which queries match your expertise and how fast you need to reply.
This program fits founders, marketers, and subject-matter experts who can speak credibly on a narrow topic and can commit to a daily habit for a few weeks. It is a poor fit for teams that want a one-time push, that have no genuine expert to quote, or that need results this week. If discovery is the bottleneck rather than citation quality, editorial placements that put your brand on authoritative domains move the needle faster than reactive pitching. Reach for reactive tools when you have the expert and the patience; reach for placement infrastructure when you need the authority to exist first.
In the KDD 2024 GEO paper, adding expert quotes was the top-performing content change, lifting position-adjusted word count by 41% and subjective impression by 28%. Treat that as a post-retrieval reuse result: quotes make a page more citable once it is already in the model's context, not a guarantee that the page gets discovered.
Yes. Cision shut down Connectively, the platform formerly known as HARO, on December 9, 2024. Featured later acquired the HARO name and revived it in April 2025 as a free newsletter, but the original service no longer operates.
Source of Sources, launched by HARO's original founder Peter Shankman, is the closest free replacement. It is email-based, serves more than 20,000 journalists, and copies HARO's original rhythm. Featured's revived HARO newsletter and #JournoRequest on X are also free.
Run two or three in parallel. Source of Sources rebuilds volume, Qwoted improves conversion on better-matched queries, Featured suits longer expert answers, and #JournoRequest on X catches time-sensitive requests. No single platform replaces the reach HARO once had.
Not on its own. A quote improves how a page is used once it is retrieved. Getting retrieved still depends on domain authority, source placement, and freshness. Combine earned quotes on authoritative sites with well-structured on-page quotes so engines have a reason to find the page and a reason to cite it.
Expert quotes are one of the strongest GEO moves in the research, but HARO is gone. Here is the 2026 stack for earning them and making them extractable.
The single strongest content change in the foundational GEO research was not a schema tag or a word count. It was adding a named expert quote. In the KDD 2024 GEO paper, Quotation Addition was the top-performing optimization method across a 10,000-query benchmark, improving position-adjusted word count by 41% and subjective impression by 28% over baseline. An answer engine reaches for a sentence it can lift cleanly, and a quote with a name, a title, and a claim attached is the cleanest unit on the page. The problem in 2026 is that the platform most operators learned to earn those quotes on no longer exists.
HARO shut down on December 9, 2024, and its replacement stack looks nothing like the single daily email operators relied on for a decade. Signals runs an aged Reddit account marketplace plus an editorial network for AI brand mentions across Reddit, Quora, Product Hunt, and Threads, so we watch which source surfaces AI engines actually reuse. Before the tactics, one honest caveat: the GEO paper's lift is a document-use result, measured after a source is already in the model's context. It proves quotes get reused once your page is retrieved. It does not prove a quote alone earns discovery.
Key takeaways
Expert quotes were the highest-performing method in the GEO paper, but the lift is a post-retrieval "how much gets reused" number, not a discovery guarantee.
HARO (rebranded Connectively) was shut down by Cision on December 9, 2024. The brand was later revived by Featured, but the original platform is gone.
The 2026 stack is a set of tools, not one replacement: Source of Sources, Qwoted, Featured, and #JournoRequest on X each cover a different gap.
A quote only helps AI citation if it is attributed and structured. A floating sentence with no name is noise; pair it with an answer capsule so the engine can extract it.
Earning third-party quotes and hosting your own expert quotes are two different plays. Run both, because original data and structure decide whether the quote survives summarization.
Expert quotes work because they hand the model a discrete, attributable fact it can quote without inventing context. The GEO paper (Aggarwal et al., KDD 2024) tested nine content changes on a 10,000-query benchmark. Quotation Addition ranked first, lifting position-adjusted word count by 41% and subjective impression by 28% against the unoptimized baseline. Cite Sources and Statistics Addition followed closely. The pattern is consistent: concrete, sourced, named content gets reused; vague assertion does not.
Read the number carefully. The GEO benchmark measured how much a source was used and credited once it was already retrieved into the answer context. It is a citation-quality signal, not a crawling or ranking signal. That distinction is the difference between a tactic that works and a slogan that overpromises. A quote from a named expert makes your page more likely to be quoted when it is in the running. It does not, by itself, put the page in the running. Discovery still depends on authority, retrieval placement, and freshness.
HARO is gone. Peter Shankman launched Help A Reporter Out in 2008; Vocus acquired it in 2010, Cision absorbed it, rebranded it Connectively, and then shut Connectively down entirely on December 9, 2024. The single thrice-daily email that trained a decade of digital PR no longer sends. If your quote-earning playbook still starts with "reply to the HARO email," it has been dead for over a year.
Two things filled the vacuum. Shankman relaunched the original format as Source of Sources, a free, email-based service that serves 20,000-plus journalists and deliberately copies HARO's early rhythm. Separately, Featured acquired the HARO brand name from Cision and revived "HARO" in April 2025 as a free daily newsletter attached to its own curated platform. So "HARO" now points to a Featured product, not the original service. For an operator, the takeaway is simple: the brand you remember is a marketing label now, and the real work happens across several tools.
There is no single replacement. The four platforms that produce the most editorial quotes in 2026 each cover a different gap, and the right answer for most teams is to run two or three in parallel rather than betting on one. Below is how we weigh them for GEO-oriented placement, where the goal is a named, attributable quote on an authoritative domain.
Source of Sources. Peter Shankman's HARO successor, free and email-based, closest to the original three-queries rhythm. Best for operators who want zero-cost volume and do not mind mixed journalist quality. The fastest way to rebuild a HARO-style habit without a subscription.
$0Qwoted. A modern reactive platform with real-time alerts, structured pitch management, and search. Consistently reported as the highest-conversion paid option. Best when you want to work fewer, better-matched queries instead of firing off replies to everything.
FreemiumFeatured. Owns the revived HARO brand. Uses a curated question model that filters spam and favors longer expert answers, which suits thought-leadership placements. Best for depth over speed, with a free entry tier and paid plans above it.
Curated#JournoRequest on X. Not a platform, a live hashtag. Fastest response window of any option because you reply directly to the reporter. Best as a monitoring layer, not a primary channel; noise is high and you filter it yourself.
Real-timeTreat these as a stack. Source of Sources rebuilds volume, Qwoted raises conversion, Featured lands the longer expert pieces, and the hashtag catches time-sensitive requests the newsletters miss.
Win the placement by writing the quote the reporter can paste without editing. Journalists working a HARO-style query are on deadline and reading dozens of pitches. The response that gets used is short, on-topic, and already in publishable form: two or three sentences, a specific number or example, and a clean attribution line with your name, title, and company. Lead with the answer, not with your bio.
Three habits separate placements from ignored pitches. Respond inside the first hour, because early, complete answers get read before the reporter fills the slot. Answer the exact question asked, not the question you wish they asked. And give a concrete claim a generalist could not, since specificity is what makes a quote quotable. "Founders underprice by anchoring to competitors instead of value" beats "pricing is very important." The same trait that wins the journalist, a discrete and attributable claim, is what an AI engine later extracts from the published article.
Host your own expert quotes with the same discipline you use to earn third-party ones. Earned placements build authority off-site; on-site expert quotes give the engine something to lift when your page is already retrieved. Both matter, and the on-site version is fully under your control. The rule is that a quote only counts if it is attributed in text near the claim, not buried in an image or a video caption an LLM cannot parse.
Format the quote as an extractable unit. Put the speaker's name, title, and organization in the same sentence or the line directly adjacent, so the model does not have to guess who said it. Keep the quoted claim self-contained: it should still make sense if lifted out of the page. Pair it with a short answer capsule at the top of the section and clean 120 to 180 word chunks so the surrounding structure does not fight the extraction. A named quote inside a wall of text is far less useful to an engine than the same quote given room to breathe.
Budget mostly time, not money. The tools themselves are cheap or free: Source of Sources costs nothing, the X hashtag costs nothing, and paid tiers on Qwoted and Featured generally run in the low tens of dollars per month. The real cost is the operator hour spent reading queries and writing tight, specific pitches daily. Expect a low hit rate at first; conversion climbs as you learn which queries match your expertise and how fast you need to reply.
This program fits founders, marketers, and subject-matter experts who can speak credibly on a narrow topic and can commit to a daily habit for a few weeks. It is a poor fit for teams that want a one-time push, that have no genuine expert to quote, or that need results this week. If discovery is the bottleneck rather than citation quality, editorial placements that put your brand on authoritative domains move the needle faster than reactive pitching. Reach for reactive tools when you have the expert and the patience; reach for placement infrastructure when you need the authority to exist first.
In the KDD 2024 GEO paper, adding expert quotes was the top-performing content change, lifting position-adjusted word count by 41% and subjective impression by 28%. Treat that as a post-retrieval reuse result: quotes make a page more citable once it is already in the model's context, not a guarantee that the page gets discovered.
Yes. Cision shut down Connectively, the platform formerly known as HARO, on December 9, 2024. Featured later acquired the HARO name and revived it in April 2025 as a free newsletter, but the original service no longer operates.
Source of Sources, launched by HARO's original founder Peter Shankman, is the closest free replacement. It is email-based, serves more than 20,000 journalists, and copies HARO's original rhythm. Featured's revived HARO newsletter and #JournoRequest on X are also free.
Run two or three in parallel. Source of Sources rebuilds volume, Qwoted improves conversion on better-matched queries, Featured suits longer expert answers, and #JournoRequest on X catches time-sensitive requests. No single platform replaces the reach HARO once had.
Not on its own. A quote improves how a page is used once it is retrieved. Getting retrieved still depends on domain authority, source placement, and freshness. Combine earned quotes on authoritative sites with well-structured on-page quotes so engines have a reason to find the page and a reason to cite it.
Expert quotes get reused once your page is retrieved. Editorial placements are what get it retrieved in the first place. Signals places brand mentions across a 20,000-plus site editorial network, so named quotes about your brand appear on the authoritative domains AI engines already read.
Sources