CMOs need to measure earned media beyond coverage

Earned media belongs in the CMO’s data program

A chief marketing officer (CMO) can have a complete coverage report and still miss how outside information shapes a buyer’s view of the company. Coverage counts, estimated impressions and comparisons with competitors measure awareness. To understand how that awareness develops, the marketing team also needs to examine what coverage reveals about its messages and how published information shapes the answers buyers receive from AI tools.

For years, earned media, coverage published independently of a company’s paid advertising, has been managed mainly through those awareness measures. Teams can also learn from coverage by examining which messages attract interest, monitoring whether the brand appears in AI answers and using those observations to improve decisions. This responsibility belongs with the CMO even when public relations (PR) reports elsewhere, because the findings affect positioning, content and sales support across marketing. Incomplete attribution, or the ability to assign credit for a buyer’s actions, makes this harder, but teams can still manage what they learn from coverage and how it remains useful through later citations and reuse.

Buyer research makes external information strategically important

The case for broader measurement starts with buyers, who can evaluate a company before reaching its website or speaking with sales. Forrester’s 2026 business-buying research identifies generative AI, which produces answers from a user’s instructions or questions, as a starting point for business-to-business (B2B) research. It also finds that buyers regard AI and conversational search, search through questions and follow-up dialogue, as more meaningful information sources than vendor websites or sales teams. A vendor therefore needs to be visible while buyers are forming their understanding of the category.

At that stage, buyers can compare suppliers and begin forming a shortlist before the vendor knows an opportunity exists. Earned coverage can help a company enter consideration because it supplies information outside the company’s own communications. The practical question is which outside information buyers encounter, including the material AI tools cite when answering their questions.

Citation research offers one way to examine that information. Muck Rack’s May 2026 study, “What Is AI Reading?”, analyzed more than 25 million cited links and attributed 84% of citations to earned sources. Muck Rack sells PR software, so evidence that outside coverage matters to AI visibility supports its commercial interest in measuring and managing that coverage. Its earned-source category includes journalism, academic research, government websites and other third-party content across ChatGPT, Claude and Gemini.

Because that category is broad, marketing teams need to separate media coverage from other independent material when interpreting the result. Academic research or a government website can inform an answer without any connection to a company’s media-relations work. Reviewing the sources cited in answers about a particular category helps a brand identify both the information shaping those answers and the publications its PR team could approach.

The broader pattern has persisted across editions of Muck Rack’s research: earned-source shares ranged from 82% to 89% across three editions. In the May findings, paid and advertorial material, advertising presented in an editorial format, together accounted for 0.3% of citations. These figures describe the mix of cited information, which makes them useful for planning AI visibility. Paid media’s performance in other activities requires measures suited to those activities.

Alongside these findings, a practitioner who works directly with CMOs reports seeing a similar external-source pattern in their own work and describes the opportunity as developing over the past year. The practitioner explains it as models learning from human writing and acquiring a preference for independent voices over vendor sales messages. In that account, media coverage supplies information and credibility cues that can shape what buyers and AI tools say about a company.

To turn that explanation into a measurement plan, teams need to distinguish how AI systems use information. A model learning from text, an AI product retrieving information and an answer citing a page are different activities. The citation studies concern the last of these, while the practitioner’s explanation concerns how models learn. Teams can examine which external material appears in answers to prospects’ questions while keeping those mechanisms separate.

OpenAI’s content-licensing arrangements offer a concrete example of reporting becoming available within an AI search experience. OpenAI has paid licensing relationships with dozens of media organizations, including Axios, Financial Times, Wall Street Journal and Business Insider. ChatGPT Search can present summaries and excerpts of licensed reporting alongside citations and direct links. Licensing makes reporting available within the product; selection for a particular answer is a separate step.

These uses of external information give marketing leaders a reason to examine how the company is represented beyond its website. The company controls its own domain and content, while outside publications contribute independent descriptions, judgments and evidence that buyers and AI tools can encounter. Forrester’s research establishes the importance of early AI-assisted research, Muck Rack measures cited sources, and OpenAI’s arrangements illustrate access to reporting. Each informs a different part of the decision about how to build visibility.

Use coverage to improve positioning and extend each placement’s value

Once outside information becomes part of the visibility strategy, each placement becomes an opportunity to learn. A published story shows that a category issue or angle interested a journalist enough to pursue it. The response to placements can also reveal which messages attract attention and which receive little response. PR practitioners can use those observations to repeat productive approaches and reconsider weaker ones.

Journalist interest can then inform a hypothesis about positioning: how the company explains its relevance and difference to prospective buyers. An angle may suggest what prospects are thinking or feeling, while buyer demand and progress toward a sale require their own evidence. The team can compare the angle with what appears in answers to relevant buyer questions. This connects editorial response with a second observable signal about how the company is represented.

Beyond that learning value, coverage can lead to further exposure. A story adds externally published information and credibility signals about the company, which may increase the chance of a later AI citation. A citation can expose the brand to more people, and that visibility may help support another media opportunity. Further coverage adds more information and another external credential, with each step creating an opportunity for the next.

The opportunity can continue after the story’s initial attention has passed. The practitioner argues that citation benefits can begin at publication and continue for months or longer as AI tools return to the story. A paid campaign ends when its funding stops, while an earned story can remain accessible without continued campaign spending. Teams can monitor and reuse it over time.

To make use of that availability, the practitioner proposes treating publication as the start of a reuse plan. In their generic example, one strong story supplies material for a blog post, a sales asset, several social posts, an email and supporting evidence for a later pitch, while remaining available for AI citation. Each use gives another audience a reason to encounter the coverage or its implications. Planning those uses makes the work after publication part of the placement’s value.

Incomplete attribution makes baseline-building necessary

Managing that continuing value requires measurement, and many teams are still learning how to observe AI visibility. In findings released in June, Semrush reported that 45% of marketing leaders could not accurately measure their brand’s presence in AI answers. Only 9% had tools to track that presence across platforms. Semrush sells marketing software, so demand for AI visibility tools is commercially relevant to its business.

The reported gap makes budget decisions harder because methods for measuring AI influence and assigning credit are still developing. Available approaches can be inconsistent and expensive, leaving teams without stable benchmarks for comparison. A marketing leader then struggles to judge whether a change represents progress or variation in the answers the tools return. Building a repeatable record is an early step toward making those judgments more useful.

Even with that record, last-click attribution presents a separate problem. This method gives credit to the final recorded interaction before a conversion, such as an inquiry or purchase, so earlier exposure can remain unrecognized. A buyer may encounter a company in an AI answer and later reach its website through another route. Without a direct referral from the cited story, the team has difficulty assigning credit to information that helped shape the decision.

CMOs are tackling the difficulty of observing AI visibility in several ways. The practitioner reports that their CMO contacts manually check prompts, the questions or instructions entered into AI tools, test more advanced third-party software and build their own agents, software that automates parts of the checking process. Each approach helps teams investigate whether the brand appears and what information supports the answer. These practices offer different ways to begin collecting a record.

A baseline becomes useful as those observations accumulate. Repeated checks establish what the team currently sees, making later changes easier to recognize and investigate. Their immediate value is better observation and decision-making; revenue lift, conversion improvement, acquisition-cost savings and campaign return require commercial measurement. Teams can develop that measurement while using the baseline to guide current work.

Build a repeatable process from buyer questions to pitching and reuse

To build the baseline, start with questions prospects actually ask, because generic brand queries provide only a partial view of discovery. Ask those questions in ChatGPT, Claude and Gemini, and record whether the company appears in the answers. Alongside each appearance, note the sources cited to support it. This connects the brand’s visibility to a buyer’s information need and the material used in the response.

The cited sources give media planning a more specific starting point. Review which outside outlets appear in answers about the category and use that review to prioritize publications for pitch consideration. An outlet’s presence in relevant answers gives the team a reason to assess its editorial fit. A future placement would still need to be selected for citation, so outreach planning should account for relevance to both journalists and buyer questions.

To understand the balance of visibility, compare citations of outside coverage with citations of the company’s own website. The proposed earned-to-owned citation ratio tracks that relationship over time, with “owned” referring to content the company controls. Treat it as a management comparison, with counting rules the team defines and applies consistently. Keep PR coverage separate from research, government information and other independent material so the comparison reflects the activity being managed.

Consistent counting becomes more useful with repeated monitoring. An isolated answer records what one tool returned on one occasion, while a continuing record gives the team a baseline for assessing change. Keep the buyer question attached to each observation so appearances remain understandable in context. The team can then examine where the brand is represented and which sources accompany it.

Those observations should feed back into positioning and outreach decisions. Compare the response to placements with brand appearances in AI answers, then use the combination to refine messages and future pitch priorities. For example, an angle that attracts journalist interest becomes more strategically relevant when the resulting coverage also appears in answers to an important buyer question. The team then has a specific signal to investigate and act on.

Reuse should fit into the same workflow through a checklist attached to each placement. Assign responsibility for the blog, sales, social, email and pitching uses described earlier, so publication triggers planned follow-up work. Connect those assignments with the people responsible for distributing content and supporting sales. This makes the placement’s continued use part of the team’s routine.

The CMO’s scorecard should make room for that work alongside existing awareness measures. Retain coverage totals, estimated impressions and competitor comparisons, then add observations about appearances in answers to buyer questions, cited sources and the balance between external and owned visibility. Budget for the monitoring, learning and reuse needed to produce those observations. The allocation can reflect the company’s needs while making responsibility for the work explicit.

Connect external visibility with the company’s website

Alongside the expanded earned-media program, paid search and company-controlled content continue to serve buyers during research. External reporting can contribute to discovery and credibility, while the website gives visitors a place to examine the company’s own explanation and offering. Prospects who reach the domain need a strong experience there, including when they arrive after encountering coverage or an AI answer.

That arrival creates a practical test for the website: can the prospect find a clear, relevant account of the company that addresses the questions raised during earlier research? The brand has limited control over how outside sources describe it, but it controls what visitors encounter on its own pages. Keeping those pages relevant to the buyer’s questions gives external visibility somewhere useful to lead.

Key takeaways for decision-makers

• Put earned media in the CMO’s data program: CMOs need to connect coverage measurement with positioning, content and sales support, even when PR reports elsewhere. Add AI visibility, message insights and reuse activity to existing awareness measures.
• Track the sources buyers encounter: Marketing teams need to examine citations in AI answers to relevant buyer questions. Separate media coverage from academic research, government information and other independent sources to identify useful PR opportunities.
• Turn placements into learning and reuse: PR teams can use journalist interest to develop positioning hypotheses, then compare those signals with buyer evidence. Assign follow-up uses across sales materials, email, social posts and future pitches when each story publishes.
• Build a baseline for AI visibility: Marketing measurement owners need repeated observations across ChatGPT, Claude and Gemini to assess changes in brand presence and cited sources. Keep visibility findings separate from claims about revenue impact, which require commercial measurement.
• Connect monitoring with outreach decisions: Start with questions prospects actually ask, record brand appearances and citations, and use relevant cited outlets to inform pitch priorities. Apply consistent counting rules when comparing citations of external coverage with company-owned content.
• Give external visibility a useful destination: Website owners need to address the questions prospects bring from coverage and AI-assisted research. Clear, relevant company pages help visitors assess the offering after encountering outside information.