TL;DR

D S Simon surveyed 168 PR executives and published the results as “Earned Media in the AI Era: Disrupting the PESO Model®.” Forty-four percent picked earned as the most important media type for AI discoverability, and only 5% picked shared, which isn’t a disruption of the PESO Model; it’s the model working as designed: owned and earned are the visibility engine, and shared and paid amplify.

Then a second survey of 200 marketers asked the same question and paid won. PR pros vote for earned, marketers vote for paid, and everyone votes for the thing they already buy—which is why “most important” is the wrong question. The finding that actually matters is buried on page 10: 87% of agencies say they’re succeeding at this, only 59% of brands agree, and the gap is measurement.

Key Insights

Earned at 44% and shared at 5% isn’t a disruption of the PESO Model®. It’s Visibility Engineering™, as described back to us via survey.
Two surveys asked the same question a few weeks apart and got two different winners. PR executives said earned (44%). Marketers said paid (35%).
Owned media finished third in one survey and dead last in the other, and that’s the finding that should worry you.
87% of agencies say they’re succeeding in optimizing earned media for AI discovery. Fifty-nine percent of brands agree. Brands rank lack of measurement as their number one barrier; agencies rank it fifth and put the client relationship first. That’s one problem described from both sides of the table, and measurement is the cure.

AI Isn’t Disrupting the PESO Model®. It’s Proving It.

A report landed in my inbox a couple of weeks ago with a headline that was clearly designed to get my attention. And yours, too, of course. But it really caught my attention, especially because I didn’t know anything about the report before that moment.

D S Simon surveyed 168 brand and agency PR executives about AI discoverability and published the results as “Earned Media in the AI Era: Disrupting the PESO Model®.”

Before we go any further, earned media in the AI era is not disrupting the PESO Model. AI isn’t, either.

Nobody checked with us before publishing, which is fine. No one needs our permission to write about the PESO Model. But had they, they probably would have at least written a different headline, if not looked at their data a little bit differently.

I read every page because when someone tells you your framework is being disrupted, you want to know how.

The good news is that it’s not being disrupted. Not even the tiniest bit.

And the report itself is one of the cleaner pieces of evidence I’ve seen for what we’ve been saying for two years about how PESO works in an AI-driven discovery environment.

What the Report Actually Found

Let’s start with the data because it’s good, and I don’t want the headline to get in the way. (Can you tell I’m bothered by it? LOL!)

The main question was, “What is the most important media type in the PESO Model to improve AI discoverability?”

Forty-four percent chose earned. Paid was second. Owned was third. Only 5% chose shared.

The email announcing the report says earned is “outpacing paid by 42%.”

That’s a relative figure, not a point gap. Earned was 42% more likely to be chosen than paid, which puts paid somewhere around 31%. So the actual spread between earned and paid is closer to 13 points, and the spread between earned and shared is 39.

But I digress into statistics, and you’re here for marketing or comms. The point is, the way the data is presented is not quite right.

That said, the report then had a run of near-unanimous findings. Ninety-eight percent say optimizing earned media for AI discoverability is important. Ninety-nine percent say it’s important to increase the number of earned media campaigns specifically for AI. Fifty-two percent already spend more than half of their AI discoverability budget on earned media.

And then the part I really care about.

Asked whether they’ve been successful in optimizing earned media for AI discovery, 87% of agency respondents said yes. Only 59% of brand respondents said the same. When asked about barriers, agencies put the agency/client relationship at the top (50%). Brands put a lack of clear measurement at the top (44%). Agencies ranked measurement fifth.

Hold onto that stat. We’re coming back to it. I’ll repeat it so you remember: brands put the lack of clear measurement at the top (44%), and agencies ranked measurement fifth.

Why “Earned Tops PESO” Isn’t a Disruption

The PESO Model has never said the four media types are equally weighted. It says they’re integrated.

Those are different claims, and the difference is the whole ballgame.

Our position, which we’ve discussed ad nauseum is that, in AI-driven discovery, owned and earned media are the primary visibility engine. Owned creates depth, authority, proof, and clarity. Earned creates validation, credibility transfer, and discoverability. Shared distributes and reinforces. Paid amplifies what’s already working.

We call it Visibility Engineering™. Owned and earned create the signals. Shared and paid amplify and extend the signals. Measurement tells you what changed.

Now go back to the survey. Earned on top. Shared at 5%.

That’s not a disruption of the model. That is the model, described by 168 people who may never have read or listened to a word I’ve said about it.

When practitioners independently place “earned” at the top and “shared” at the bottom for AI discoverability, you don’t get disrupted. You get validated.

I’d also gently point out that the question itself is the tell. You can’t ask “which media type in the PESO Model matters most” unless everyone answering already agrees PESO is the map. The report doesn’t challenge the framework. It runs on it.

Ask Marketers and You Get a Completely Different Answer

But here is where this data gets even more interesting. D S Simon wasn’t the only firm to run this question this summer.

Sword and the Script Media surveyed 200 U.S. marketing and communications professionals in July—76% of them in-house, and skewing marketing rather than PR—and asked which of the four PESO media types has the biggest effect on AI visibility.

Paid won. Thirty-five percent. Shared was second at 28%. Earned was third at 24%. Owned was last at 14%.

So, in the span of a few weeks, one survey of PR executives found that earned media is the most important media type for AI discoverability, and another survey of marketers found that paid media is the most important.

Earned was first in one and third in the other. Paid was second in one and first in the other.

That’s not two contradictory findings. That’s the same finding.

The same finding isn’t which media type is most important in AI visibility; ,it’s that people vote for the one they already buy.

PR pros buy earned. Marketers buy paid. Everybody is answering “what do I spend on?” and calling it “what matters?”

(And in the Sword and the Script survey, only 3% of respondents said PR or communications owns AI visibility at their organization. Marketing operations was the top answer at 17%. If you’re a comms pro reading this and wondering why the marketing team is running your discoverability strategy with a paid-first playbook—that’s why.)

To their credit, Sword and the Script drew the right conclusion from their own data. They said, “For most organizations, the priorities ought to be owned, earned, shared, and paid—and in that order.”

And they added that they have “yet to see a single study that suggests paid media, including advertising, affects AI visibility,” while there are plenty showing owned and earned do.

Owned, earned, shared, paid. In that order. I didn’t write that, but I 100% agree. And I completely agree that there isn’t a single shred of research that suggests paid media affects AI visibility.

The Finding That Should Actually Worry You

Now, look at where “owned” landed in both surveys. Third in one. Dead last in the other. Fourteen percent of marketers think the media type they control most completely matters the least for AI visibility.

I’d love to tell you that’s a survey artifact. And partly it is—”most important” invites people to name the thing they’re actively spending on, and owned media doesn’t feel like spending. But I think it also reflects something real, which is that a lot of teams still treat owned media as the website that already exists rather than as the thing every other media type points to.

Here’s why that’s a problem for the 99% who are about to increase their earned media so they show up in AI.

Large language models place heavy weight on third-party, high-authority sources when deciding what’s true and who’s credible. A mention in a trusted publication is a signal that’s hard to fake.

That’s why earned wins the survey, and the survey is right about that.

But the mention has to point somewhere. It has to land on an owned property that goes deep enough to be the source the model cites—the case study, the methodology page, the executive’s byline, the FAQ that answers the question the model is trying to answer. If it lands on a thin About page, its credibility has nowhere to transfer.

Earned without owned is a citation with nothing behind it. Owned without earned is a library nobody has heard of. Either one alone is a media type. Together, they’re a system.

So when I see owned ranked third by the people about to buy more earned media, and last by the people about to buy more paid, what I see is a lot of coverage and a lot of ad spend about to point at a lot of websites that aren’t ready to receive them.

No, the P Doesn’t Mean Paid Comes First

While we’re on the subject of things the model doesn’t say, let me address a misconception I’ve heard more than once over the years. That is the idea that because paid is the first letter in PESO, the model puts paid first.

Nope.

Never has. Never will.

PESO is an acronym, not a ranking. The letters are in that order for exactly one reason—it spells a word you can say out loud.

If the order reflected how the model actually works, it would be closer to OESP: owned and earned building the foundation, and shared and paid amplifying it.

But OESP isn’t a word, nobody would remember it, and a framework nobody can remember doesn’t become the industry’s shared language. It might also be the number one rule in branding. Create something memorable.

The sequence of the work has never been paid first. For most programs, it runs from owned to earned to shared to paid.

You build the owned content that goes deep enough to be worth citing. You earn third-party validation that lends it credibility. You use shared to distribute and reinforce. And you put paid behind it only once you know what’s working, because amplifying something unproven just means being wrong at a larger scale.

So if you’ve been reading the P as a priority, allow me to reframe it in your mind: the acronym tells you which ingredients to use. The strategy tells you the order. And in an AI-driven discovery environment, that order starts with owned and earned.

“Most Important” Is the Wrong Question for an Operating System

Here’s where I do have a real disagreement with the survey, and it’s not with the data. It’s with the question.

Asking which media type is most important for AI discoverability is like asking which part of a car is most important for getting to work.

The engine, obviously.

Duh.

Except…now try driving to the office with an engine and nothing else.

And when two surveys ask the same question a few weeks apart and get two different winners, the question has told you everything it’s going to tell you. It’s not measuring what works. It’s measuring who you asked.

The wrong conversation is “which media type should we bet on?”

The right conversation is “how do the four work together, in what sequence, and how do we know it’s working?”

Which brings me to the part of the report I think everyone should be talking about.

The Confidence Gap Is the Real Story

Eighty-seven percent of agencies say they’re succeeding at this. Fifty-nine percent of brands say they are.

Twenty-eight points is a lot of daylight between the people doing the work and the people paying for it.

I don’t think it’s a mystery. Look at the barriers.

Brands say their number one problem is a lack of clear measurement. Agencies rank measurement fifth, behind the client relationship, internal alignment, expertise, and earned, with measurement kept separate from their AI work.

The agency’s top barrier is the client. The client’s top barrier is not knowing whether the agency’s work made any difference.

That’s not two different problems. That’s one problem, described from both sides of the table.

If you’re an agency and you can’t show a client what changed in their AI visibility after an earned media push, the relationship is going to be your top barrier, because the client has no reason to believe you.

If you’re a brand and your agency keeps telling you the media tour went great, but you can’t connect it to anything, measurement is going to be your top barrier, because it is.

We saw exactly this in a recent client engagement. Their internal team is full of rock stars who are curious and thoughtful and ask really tough questions. They have strong agency partners, their earned media operation almost runs itself, their content is among the best I’ve ever seen, there are lots of positive discussions about them online, they’re targeted in their paid efforts, and their teams are highly coordinated.

Not integrated, but they are coordinated.

They also didn’t have a shared definition of what success looks like. Each line of business focused on its own KPIs, and in many cases, they didn’t ladder up to what the organization was working to achieve.

And, of course, me being me and having hired people who are like me when it comes to this stuff, we saw this as a challenge and got to work.

We built them a measurement framework built around PESO, and guess what happened?

You guessed it!

They didn’t start getting results overnight. They were already getting them. But now they were visible, and everyone knew exactly what works and what doesn’t—and they make quick decisions based on what the data tells them.

Suddenly, their barrier was not measurement, and their agencies’ barriers were no longer the client.

Back to the survey. Nearly everyone is about to spend more on earned media for AI discoverability. Ninety-nine percent. That’s a stampede.

And most of them are running toward a media type with no agreed-upon way to measure whether they’ve reached it, even as they admit it’s one of their biggest barriers.

What This Means for Your Team

If you’re in marketing and your instinct is paid-first, the Sword and the Script data is your reality check, plus the fact that no research on earth has shown paid moves AI visibility on its own.

Paid amplifies a signal. It doesn’t create one. Put it behind owned and earned, that’s already working, not in front of them.

If you’re a brand, don’t take “increase earned media” as the takeaway. Take “build the measurement framework before you increase anything” as the takeaway.

Define what AI visibility means for your organization. Which prompts matter? Which sources do you want to be cited in? Figure out what your owned properties need to contain so that when earned points at them. Then increase your earned media activities.

If you’re an agency, the 28-point confidence gap is your competitive opening. The firms that can show a client what changed, in the client’s own terms, are going to win the next round of budget from the ones that can’t.

“The relationship is the barrier” is a symptom. Measurement is the cure.

And for everyone, stop treating PESO as a planning checklist where you tick the four boxes and move on. Run it as an operating system. Owned and earned create the signals. Shared and paid amplify them. Measurement closes the loop and tells you what to do next.

The D S Simon report is right that earned media matters enormously for AI discoverability. It’s right that budgets are following. It’s right that measurement is the thing brands are stuck on.

It just has the headline backward.

Earned media isn’t disrupting the PESO Model.

It’s the part of the PESO Model that’s working.

If You’d Rather Not Figure This Out Alone

And listen. If you want help with any of this, it’s what we do all day, every day. I clearly geek out about how a fully integrated PESO Model program will affect your AI visibility, and we test things constantly to ensure we’re advising clients correctly about this stuff.

More than that, though, we’ve found a niche in helping brands build their marketing and comms measurement framework. It is, of course, built around PESO, but it’s a driving passion of mine that I’ve built a business around solving.

If you want to chat about how we might help, 30 minutes of my day is yours! And if you need to figure it out on your own, don’t forget that we have the Spin Sucks Community, where we have conversations like this every day. It’s free and comes with zero strings attached.