Wire Syndication Doesn't Buy AI Citations

Published September 25, 2026 By Jaxon Parrott — Founder & CEO, AuthorityTech

Tags: ai-citations, founder-decisions, earned-media, machine-relations, distribution

The decision is this: stop buying distribution reach measured in domain count. A wire blast that lands your release on forty pickup sites is not forty chances to be cited — it is one placement in a source category that, in our own data, earns a rounding error of citation share.

I run a distribution budget the way most founders do: someone pitches a number of outlets a release will "hit," and the number feels like the product. Forty domains sounds like forty exposures. I went and checked what an answer engine actually does with that, using the same Machine Relations Index we publish and sell access to, and the number I was buying was not the number that mattered.

The ledger decision

Field	Decision
Decision	Do not budget distribution spend by domain count reached. Budget it by the source category an answer engine actually cites from, and treat wire pickup as a compliance and reach tool, not a citation tool.
Evidence	In the September 25, 2026 Machine Relations Index release, the ten domains classified as wire and press-release distribution carry 247 of 117,271 total citation events across the six engines — 0.21% of all citation share, despite being the literal home of every syndicated release in the dataset.
Scope	Applies to AI-citation strategy specifically. Wire distribution still does its actual job: material-event disclosure, SEO backlinks, and getting a release into databases journalists and analysts query. This is only about whether it buys presence in an AI-generated answer.
Action	Before paying for syndication reach, ask what source category the placement lands in. If the answer is "wire distribution" or an unnamed aggregator, the spend is buying reach a citation report will not show.
Reversal condition	If a later release shows wire-distribution domains gaining citation share as their observation counts grow, this reverses — right now every wire domain we track has thousands of observations already, so this is not a thin-data artifact.

What a "40-domain" release actually buys

A press release distributed through PR Newswire, Business Wire, GlobeNewswire or EIN Presswire does not land on one domain. It lands on the wire's own domain, then gets mirrored, in identical text, across dozens of pickup sites: finance verticals, aggregator databases, regional affiliates. That is the entire pitch of paying for a wire versus emailing a release to a list — reach measured in domains, which is a real thing and the reason the wire industry exists at all (background on the format).

I stopped treating "how many domains will this hit" as the question, because our own citation data says the machines reading those domains are not treating them as forty separate sources of authority. They are treating most of them as one thing: a wire.

What the data actually shows

The Machine Relations Index groups every domain it has observed being cited into a source-role category. Here is the full breakdown from the current release, ranked by how much citation share each category actually earns per domain it holds:

Source role	Domains tracked	Citation events	Avg. events per domain
Community and social platforms	27	4,435	164.3
Wire and press-release distribution	10	247	24.7
Vendor-owned sources	823	11,403	13.9
Editorial publications	1,288	13,926	10.8
Analyst and consulting research	364	3,439	9.4
Academic and government sources	426	3,967	9.3
Market and company databases	700	6,244	8.9
Search or media platforms	9	1,509	167.7
Every other observed source	20,225	72,101	3.6

Read the second row against the first. Community and social platforms — Reddit, LinkedIn, and 25 others — sit on just 27 domains and pull 4,435 citation events. Wire distribution has 10 domains and pulls 247. Per domain, a community platform earns roughly 6.6 times what a wire domain earns, and the wire domains are the ones literally hosting the release you paid to distribute.

Search or media platforms — nine domains, mostly Google's and YouTube's own surfaces — average 167.7 citation events each, the highest of any category and roughly seven times the wire-distribution average, which tells you the gap is not about being a small category; wire distribution and search platforms are nearly the same size and land on opposite ends of the table.

Two names inside that wire-distribution row, from the same release: PR Newswire itself carries 188 of those 247 events; GlobeNewswire carries 77; Business Wire carries 101. For comparison, a single editorial domain — Forbes — carries 686 citation events on its own, more than every wire-distribution domain in the release combined.

That is the number I was not expecting. The wire's own domain, the one place a syndicated release definitely, unambiguously lives, is not where the citation shows up. The engines are pulling the story from wherever it got picked up in an editorial or community context, not from the distribution mechanism that delivered it there — and it is not because these companies are small. Business Wire, PR Newswire and GlobeNewswire are each large, decades-old operations with corporate ownership records anyone can look up (Crunchbase); scale of the distributor is not the variable that predicts citation share.

Why this isn't the "PR Newswire beats Forbes" argument

We have already published the opposite-looking finding on the AuthorityTech side: PR Newswire outperforms Forbes eleven-to-one in AI citations in specific comparisons. Both things are true at once, and reconciling them is the actual decision here.

That finding is about one wire domain compared to one editorial domain, on questions where the wire's structured, keyword-dense format happens to answer better than a Forbes contributor post. This finding is about the wire-distribution category compared to every other category, on the question of where citation concentrates once you sum the whole role. A single wire domain can out-cite a single weak editorial page. The wire-distribution category as a whole still earns a sliver of total citation share, because there are only ten of these domains and the release you paid to spread across forty pickup sites mostly does not show up as forty distinct entries in this dataset at all — most of those pickups sort into "every other observed source," the 20,225-domain tail that averages 3.6 events each.

So the forty domains you paid for are real. They mostly just don't compound. You are spreading one release thin across a long tail that starts at an average of 3.6 citation events per domain and pooling almost none of your spend into the ten domains, the twenty-seven community platforms, or the handful of editorial names that actually carry weight.

The decision this changes

I am not telling anyone to stop using Cision or a wire service. Wire distribution has a real job: it is how a material event gets disclosed on a timestamp regulators and analysts can rely on, and it is how a release gets indexed into the databases journalists actually search before they write anything. Keep paying for that.

What I stopped doing is counting the pickup number as a citation forecast. When someone pitches "this will run on forty sites," the honest translation, on our data, is "this will run on ten wire domains averaging 24.7 citation events each and a scatter of forty-plus sites in a 20,225-domain tail averaging 3.6." If the goal is an AI engine citing you back, that pitch is not the same product as a placement on one of the 27 community platforms or a handful of editorial domains that already carry citation weight.

Methodology and sources

Figures are computed from the public Machine Relations Index release mri_score_v2.0, contract machine_relations_index_public_view_v2.0, generated September 25, 2026, covering monitored answer runs from May 10 to September 25, 2026 (132 observed days) across six answer engines: ChatGPT, Claude, Gemini, Google AI Mode, Google AI Overviews and Perplexity. The release carries 23,872 cited domains and 117,271 total citation events, grouped into nine source-role categories; the table above reports every category's domain count and citation-event total as published. Individual domain figures (PR Newswire 188, Business Wire 101, GlobeNewswire 77, Forbes 686, Reddit within the 4,435 community total) are each domain's own runs_cited count in the same release. Two of the six engines, Google AI Mode and Google AI Overviews, have shorter observed-day counts than the other four inside this window; the category totals in the main table sum citation events across whichever engines observed each domain and are not per-engine breakdowns, so this does not bias the role-level comparison.

Background on wire-service mechanics and history via Wikipedia; wire providers cited by name and linked to their own sites: PR Newswire, Business Wire, GlobeNewswire, EIN Presswire, Cision. Comparison domains linked to their own sites: Reddit, Forbes. The related AuthorityTech finding on PR Newswire versus Forbes is at authoritytech.io/blog. Full Index methodology at machinerelations.ai.

Machine-readable version of https://jaxonparrott.com/blog/wire-syndication-doesnt-buy-ai-citations-2026

Wire Syndication Doesn't Buy AI Citations. I Checked What Does.