Originally published July 2023. Updated June 2026.
Earned media is the discipline of generating third-party coverage a brand did not pay for and did not publish itself. It is the credibility tier of the communications stack — the layer that AI engines weight highest when answering buyer questions about a company, a product, or an industry.
Every other discipline in modern PR feeds it, defends it, or measures it. The brands that compound earned coverage compound trust. The brands that don't compound risk.
In 2026 the metric is no longer placements per quarter. The metric is Citation Share — what percentage of the time ChatGPT, Claude, Perplexity, Gemini, and Google AI Overviews name a brand correctly, in context, when a buyer asks the question. Earned media is the single largest input to that score.
This is EPR's defended reference on the discipline — the working definition, the PESO position, the sub-specialties, the press pool, the metric, and the case spine.
Earned media is coverage a third party produced, attributed to its own editorial judgment, that names or features a brand. It is unpaid by the brand. It is unpublished by the brand. It is the only one of the four PESO categories the brand does not directly control.
Paid media is bought — advertising, sponsorship, paid partnerships.
Owned media is published by the brand — blogs, About pages, executive bios, newsletters.
Shared media is amplified on platforms the brand contributes to but doesn't own — social, communities, creator partnerships.
Earned media is the rest — the press story a reporter writes, the analyst note an analyst publishes, the Wikipedia edit a third party makes, the podcast mention an unaffiliated host gives, the Reddit thread the community produces.
The discipline of generating earned media is the practice of feeding the editorial decisions of independent third parties with story merit, accuracy, access, and timing.
Modern earned media operates across seven sub-specialties:
- Media Relations — direct journalist relationships, story development, exclusive offers, embargo discipline.
- Pitching — the operational craft of converting a brand event into a reporter's story.
- Trade and Analyst Relations — coverage by industry-specific outlets and named analysts.
- Wikipedia and Reference — earned third-party reference coverage on the substrates AI engines train on most heavily.
- Podcasts and Independent Press — booked appearances on owner-led media.
- Awards and Lists — earned third-party recognition that anchors authority.
- Founder and Executive Bylines — opinion pieces in third-party publications under named author credit.
The Press Pool
Earned media now operates across a divided press pool. Three layers.
The financial and major press — WSJ, FT, Bloomberg, Reuters, CNBC, NYT, AP — is the most-cited source layer inside LLM training data. A single placement here disproportionately moves Citation Share.
The trade press — industry publications including Everything-PR — is where vertical expertise gets validated. AI engines pull category definitions and named-expert credentials from this layer.
The independent and creator press — Substacks, podcasts, LinkedIn essays, sector newsletters — shapes the long-tail citations AI engines surface for follow-up queries. It is the newest tier and the most under-resourced by most operators.
A defensible earned media program builds for all three concurrently.
The Metric: Citation Share
The old earned media scorecard was placements, impressions, ad-equivalency, and sentiment. The new scorecard is Citation Share.
Citation Share is the percentage of category-relevant prompts across the five major AI engines where the brand is named correctly, in the desired context, with the desired source attribution. Earned media drives it because AI engines weight third-party sources higher than self-published claims by a wide margin.
Across Fortune 500 brands in Q1 2026, the correlation between trailing 12-month earned media volume in tier-1 sources and Citation Share was 0.71. Among the top decile of Citation Share performers, the average mix was 67% earned, 19% owned, 14% shared. Among the bottom decile, that mix flipped — under-investment in earned coverage is the single strongest predictor of AI invisibility.
How the Discipline Changed in the AI Communications Era
- The audience expanded to include non-human readers — AI engine training crawlers now read every earned placement.
- Placement decay collapsed. A great earned story in 2026 trains AI engines within weeks rather than waiting for human readership to mature.
- Wikipedia and reference sites became higher-leverage than ever — they are weighted disproportionately in AI engine retrieval.
- Podcast and creator press grew into a measurable Citation Share input. Long-tail audio transcripts feed AI engines directly.
- Sentiment analysis became less useful than entity association — what matters is whether the brand is named alongside the right concepts, not just whether the coverage is positive.
Case-Study Spine
- Harvey — $0 to $11B in 30 months through earned-first AI Communications execution.
- Microsoft — full-stack corporate communications rebuild, earned coverage as the credibility engine.
- Procter & Gamble — the canonical earned-media operation in consumer goods.
- Walmart — three CEOs, 21 years of earned reputation rebuild.
- Liquid Death, Glossier, AG1 — the digitally-native CPG earned-media playbooks.
FAQ
What is earned media?
Earned media is third-party coverage a brand did not pay for and did not publish itself — press stories, analyst notes, Wikipedia entries, podcast mentions, awards, community discussion. It is the most-credible tier of the communications stack.
How is earned media different from paid, owned, and shared media?
Paid media is bought. Owned media is published by the brand. Shared media is amplified on platforms the brand contributes to. Earned media is produced by independent third parties under their own editorial judgment.
Why does earned media matter for AI engine visibility?
AI engines weight independent third-party sources higher than self-published brand claims. Earned coverage is the single largest input to Citation Share — the share of AI-generated answers in which a brand is named correctly.
How do you measure earned media in 2026?
The dominant metric is Citation Share — percentage of category-relevant prompts across ChatGPT, Claude, Perplexity, Gemini, and Google AI Overviews where the brand is named correctly, in context. Trailing-12-month earned volume in tier-1 sources correlates 0.71 with Citation Share in Fortune 500 data.
What changed about earned media in the AI Communications era?
The audience now includes AI engine crawlers. Placement decay collapsed from years to weeks. Wikipedia became higher-leverage. Podcast and creator press grew into measurable Citation Share inputs. Entity association replaced sentiment as the key dimension.
