Earned Media Playbook

Earned media is coverage you don’t pay for: you give a journalist, host or analyst something their audience needs, and they give you their audience’s attention. A startup can earn it without an agency, with one clear purpose, a story only it can tell, a short list of the right people and a pitch they can judge in thirty seconds.

Many early-stage founders think press is something you buy. You hire an agency, it sends a release to a list, and coverage appears. For a seed to Series B startup that rarely works: the retainer often costs more than the coverage is worth, and the release goes to people who never asked for it.

Earned media is borrowed trust: an article, a quote, a podcast interview, an analyst’s mention, a recommendation in a newsletter or a community thread. A reader believes an editor who chose to write about you more than an ad you paid to place, and the people you sell to check you before they call you. They search your name, your founder’s name and your category, and what someone else wrote carries more weight than anything on your own site. Its job for a startup is narrow: make you credible to the few people who decide whether to buy, hire or invest.

Pitching still works when the story fits the person. In Muck Rack’s 2026 survey of 897 journalists, 86% said at least some of their stories start with a PR pitch^[1]. Coverage also compounds: a good piece keeps getting found, linked and quoted, now by AI answer engines as well as by people (Step 8).

Pick one job: credibility with buyers, visibility for hiring, proof for investors, or authority in search and AI answers. Each points at different outlets (Figure 1). A trade newsletter your buyers read beats a national business title they skim.

Earned media is wider than the press. Besides reporters and editors, it includes newsletter writers, podcast hosts, industry analysts, award juries and the communities where buyers ask each other for recommendations. Choose only the ones your audience actually reads, hears or asks.

Write the job in one line: “[Audience] should see [company] in [outlets] so that [outcome].” Every pitch should serve that line.

“We exist” is not a story. “We launched a product” rarely is either. Journalists need something new, true and useful to their readers. Four kinds of story work for startups:

• Data you own. Something you measured that nobody else has, with a method you can explain.
• A position you can defend. A view your category gets wrong, backed by your own work.
• A customer result. Real numbers, with the customer’s written permission to share them.
• A clear reason for now. A new regulation, a shift in the market or a funding round, tied to what it changes for the reader.

Data is worth the effort: 40% of journalists in Muck Rack’s 2026 survey named original data or research among the most valuable things a pitch can offer^[2]. Test each idea with one question: would a reader forward this to a colleague? If not, keep working on it. Figure 2 adds the other three tests.

Twenty names beat two thousand. Relevance is where most pitches fail: in Cision’s 2026 survey of 1,899 journalists in 19 markets, 72% said fewer than a quarter of the pitches they receive are relevant to them^[3], and 88% of the journalists in Muck Rack’s survey immediately disregard a pitch that misses their beat^[1].

Look for the reporters, newsletter writers, podcast hosts, analysts and community leads who cover your buyers’ problem, and read or listen to their last five pieces before adding them. For a podcast, check that recent guests are people like you and that the listeners are your buyers. Analysts matter once buyers in your category consult them, and you don’t need to be a client to brief one: Gartner, for example, takes briefing requests from vendors that are not its clients, typically 45 minutes with one or two analysts^[4].

For each name, note what they cover, what they wrote recently, and why your story fits. If you can’t fill in that last column, they don’t belong on the list yet.

The easiest coverage comes from people who already know you, and reporters are looking: 42% of journalists in Cision’s 2026 survey said they were looking for verified subject-matter experts^[5]. Answer journalists’ requests for expert comment in your field; services such as Qwoted, Featured and Source of Sources collect them in one place. Reply usefully to reporters’ posts. Share your data openly before you ask anyone to write about it.

Do the same in the communities your buyers use, as a person rather than a brand: answer the question that was asked, and mention your product only when it is the answer. A recommendation there is earned media too, and nobody can buy it. The Community & Dark Social Playbook covers how.

Publish the story on your own site first. A journalist is more likely to cover data that already has a home, a method and a chart.

Keep it short and specific. The subject line is the story, not your company name. The first two lines say what’s new and why their readers care. Then say what you can offer: the data, an interview, a customer who will talk.

• Address one person by name, about their beat.
• No attachments. Link to the data instead.
• One pitch per person. No mass sends with the name swapped in.

That matches what journalists say they want. In Muck Rack’s 2026 survey, 69% preferred pitches under 200 words, 62% a one-to-one email and 78% pitches sent before noon; 71% immediately disregard a pitch that is overly promotional, and 50% one that looks like a mass email^[2]. For a podcast, pitch an episode rather than yourself: the topic, why the host’s listeners care and two or three questions you can answer well. How to pitch a journalist when nobody has heard of your startup works through a cold pitch line by line.

If you can’t explain the story in two sentences, it isn’t ready to pitch.

When a journalist says yes, they need things fast. Prepare them before the first pitch goes out:

• A press page with a short company description, the founder’s bio and a contact who answers the same day.
• Current high-resolution headshots and a logo pack.
• The method behind any data you share.
• A customer who has agreed, in writing, to speak.

Images matter more than founders expect: 37% of journalists in Muck Rack’s 2026 survey named high-resolution images among the most valuable things a pitch can offer^[2]. Nothing loses a story faster than a two-day wait for a photo.

Send one follow-up after a few days, with something new if you have it. Then stop. Silence is an answer, and a third email closes a door you may want later: 46% of journalists in Muck Rack’s 2026 survey disregard pitches sent repeatedly without a response^[2].

Track every pitch in one sheet: who, when, the angle, the follow-up date and the result. Over a quarter it shows you which stories land and which people respond.

Coverage fades fast if you only share it once. Link to it from your site and your press page. Quote it in your sales deck. Post about it on LinkedIn with what you learned, not just “we were featured”. Add it to your proposals.

It also works for you in AI answers. Muck Rack’s May 2026 analysis of more than 25 million links cited by ChatGPT, Claude and Gemini found that earned sources made up about 84% of citations and journalism alone 27%, while paid and advertorial content made up 0.3% (Figure 7); more than half of the journalism cited had been published in the previous 12 months^[6]. Muck Rack counts academic, government, encyclopedic and third-party company pages as earned, so the 84% is broad. A 2025 study of AI search engines answering consumer product questions found the same pattern: “a systematic and overwhelming bias towards Earned media” over brands’ own sites and social platforms^[7]. The GEO & AEO Playbook covers the rest of what engines cite.

Then thank the journalist and send them your next useful idea when you have one. The second story is easier than the first.

Use your own numbers, not an industry benchmark. Track four things from the first quarter:

• Reply rate on pitches. Low replies mean the story or the list is wrong, not the channel.
• Pieces published in the outlets from your Step 1 line.
• Where buyers heard of you. Ask “How did you hear about us?” on every intake form and first call, and count the mentions of coverage. The Self-Reported Attribution Coder codes the answers consistently.
• Whether AI answers name you. Each quarter, ask the assistants your buyers use the same set of questions your buyers ask, and count how often you are named. The AI Visibility Significance Test tells you whether a change is more than noise.

If replies rise but nothing publishes, sharpen the materials in Step 6. If pieces publish but no buyer mentions them, revisit the outlets in Steps 1 and 3.

What changes by stage?

What mistakes should I avoid?

• Pitching the company instead of a story. Nobody covers a startup because it exists (Step 2).
• Mass pitching. Two thousand identical emails reach nobody who cares (Steps 3 and 5).
• Promising data you can’t back. One figure that falls apart costs you the source relationship; publish the method first (Step 4).
• Letting the facts drift. When your site, press page, profiles and coverage describe the company differently, readers and AI answers repeat whichever version they find. Keep one description on the press page and use it everywhere.
• Stopping after one piece. Earned media works as a habit, not a campaign, and AI answers lean on recent journalism (Step 8).
• Having no plan for coverage you did not ask for. Press attention is not always attention you asked for. Decide who speaks and what is said before a reporter calls about a problem, using the Crisis Communications Playbook.