How The Triple-P Model Measures Visibility That Drives Revenue

More and more organizations are seeing their organic traffic generation tank in recent months. Some are claiming that SEO is dead. Others are touting that AI is stealing traffic. There is some truth, as I’ve watched companies celebrate page-one rankings while traffic fell off a cliff or their pipeline sat flat.

Often both.

The rankings they enjoyed were real. But the revenue was not. The gap between those two things is the reason I built a model that allowed me to assess and improve not only an organization’s visibility but its productivity.

In other words, rankings and impressions answer one question, which is were you seen. But they say nothing about whether the right people saw you, whether it registered, or whether any of it turned into money.

Three different questions, three different layers. I call them the Triple-P Model.

Why rankings and impressions mislead

During my career as a leader within marketing agencies, our teams were tasked to give what our clients were asking for. They wanted to appear in more searches (impressions) and be as high as possible in those impressions (rankings).

A ranking is a position. An impression is an appearance. But neither one tells you if the visibility was worth anything.

You can rank number one for a keyword no buyer ever types. You can pile up impressions on content that sends people right back to Google. Both look great on a dashboard. But they don’t always move the needle in a business.

That’s why the Triple-P Model refuses to stop at the first layer. It makes every visibility number answer for itself, with three questions asked in order.

The three layers

For years, I followed this mental model when assessing and improving my client’s visibility. Each layer had a classic way of measuring it. Each has a new one, now that AI answers sit between you and your buyer.

The questions haven’t changed. The units have.

Presence: the claims you own

Presence asks whether you showed up. For years, that meant counting things. How much content you had, how many keywords it covered, how many impressions it earned.

When conducting a visibility audit, I would dive into my client’s corpus of published work, look into how many of those pages were indexed and showing up in search results, and for how many queries they would show up under.

Today, the unit is the claim.

A claim is a specific, defensible statement your content makes that a machine can lift and attribute. When an AI tool builds an answer, it expands your question into related sub-queries (what the industry calls “fan-out queries”) and assembles claims from different sources, happily pulling from four different sites in a single response.

Which means a thousand pages that make no liftable claims is a library with no inventory. The question is no longer “do we have content on this topic.” It’s “which claims on this topic do we actually own.”

Prominence: coverage

Prominence asks whether the showing up counted. The classic read I used to make was: what ranked, for which queries, and how high on the page. Meaning, after getting an understanding of how much presence my client had, I then looked into whether that presence was winning in search results and being high enough to get noticed.

Today, the unit is coverage.

Claims won over claims made. Pull the AI answer for a question that matters to your business. Break it into its individual claims. Count how many cite you. Five out of twenty is a Prominence score, and it’s countable in a way “average position” never honestly was.

Brand mentions inside AI answers live in this layer too. Being named in the answer is today’s version of ranking in the search results (i.e., blue links).

And here’s where established sites get blindsided. Expertise (or experience), authority, and trust (what the SEO industry calls E-E-A-T) still matter, but they now decide whether you’re eligible to be cited at all. They don’t decide which claims you win. Eligibility isn’t ownership.

Performance: demand

Finally, Performance asks whether the visibility paid. This means, I assessed whether my client’s presence was not only prominent enough but also translating into visitors and sales. The classic read was clicks, traffic, and the pipeline traced back to organic.

Clicks can’t carry this layer anymore, because most searches never produce one. SparkToro pegs zero-click at 68% of U.S. Google searches this year. The AI answer absorbs the visit.

But your buyer didn’t vanish with the click. They show up later as a brand search, a direct visit, or a “we saw you mentioned somewhere” on a sales call.

Today, the unit is demand.

So Performance now reads brand search volume, direct traffic, and pipeline, alongside whatever organic sessions remain. Grade it on sessions alone and strong AI-era visibility looks like failure. Grade it on demand and the same work shows up as growth.

Read the layers as a chain

The Triple-P Model is a diagnostic, not a scorecard. The diagnosis comes from reading the three layers in order. One layer creates the basis from which I measure the subsequent one.

Strong Presence, weak Prominence? You’re visible to the wrong people, or you’re answering a question related to the one being asked instead of the exact one. Strong Prominence, flat Performance? You’re earning attention that doesn’t convert, and that’s usually a positioning problem, not a search problem.

The layer that breaks tells you where the real constraint sits. Most teams never get to the diagnosis because they stop reading at Presence and call it a win. It’s the same diagnostic-first logic I bring to everything else: find the constraint before prescribing the fix, because organic visibility only compounds into revenue when all three layers connect.

Why the model survives the AI shift

I built the Triple-P before AI answers existed, and I’ve been using it in audits and teaching it for over 10 years now. And it carried over without modification. That’s because rankings were, more often than not, the middle layer. They were never the goal.

A model built on rankings as the end goal collapsed the moment the results page stopped sending clicks. A model built on three questions (did you show up, did it count, did it pay) keeps working, because AI search changed how each question gets answered without changing whether it needs asking.

Claims are the new keywords. Coverage is the new rankings. Demand is the new clicks.

That’s the whole shift in one line. The coverage logic you’ve known for two decades still wins the game. Whoever covers the most questions with the most citable answers gets assembled into the most responses. Only the unit of measurement changed.

What the numbers look like in practice

Here’s a recent engagement where all three layers moved at once. My client has real authority. Over the years, they built hundreds of pages, owned a healthy backlink profile, earned top-three positions on their most important terms, and successfully drove qualified, revenue-producing traffic to their door.

However, their traffic started to flatline. And I was brought in to investigate.

Across their three best terms, Google’s AI Overview made 20 distinct claims and cited them on 5. On one term they rank third organically and got cited zero times. Even the sites winning those claims had less authority, fewer backlinks, and thinner content. They just answered the exact sub-question the machine asked, while my client answered a related one.

That’s a Prominence problem, and no rankings report on earth will show it to you.

The Performance layer told the other half of the story. One page I checked was the most-cited source in its answer, quoted four times, and it earned two clicks. On the same site, brand searches are up 28% and direct traffic is up 13% while organic clicks fall. Sessions say decline. Demand says the visibility is working, and the buyer is arriving through a different door.

How I run it in an engagement

The Triple-P is one of many tools in my toolkit. But this model is one of the simplest to run. If you want to test it yourself, give it an afternoon. That’s all the first pass takes, and I’d run it before touching a single page.

Pick five money questions, plus five questions only your real buyer would ask. That second five matters, because the full question gives the asker away. The old SEO playbook used to be about keywords. Then, with the help of advances in semantic search, it became about topicality and intentionality.

But AI takes it one step further.

For example, one client, a training firm for consultants, ranked for years on “consulting fees.” But that string is the same whether a beginner types it or a seven-figure firm owner does. There’s a difference between “how do I set consulting fees” and “how do I raise my fees with existing consulting clients,” and the question tells the machine exactly who’s asking.

In turn, your page has to match that reader or lose to one that does. Knowing where your buyer sits on the awareness spectrum is what makes those buyer-only questions worth choosing carefully.

Pull the AI answer for each question in a private browser session. Break every answer into its individual claims. Count two numbers, claims made and claims won. That ratio is your new scoreboard, and you rerun it monthly, because AI answers reshuffle constantly even when their substance holds.

Then look at the claims you lost. In my experience, most aren’t lost to better expertise. They’re lost to better formatting (i.e., structure). The claims you win sit in liftable modules with clear labels, many with a number in them. The ones you lose are buried in prose. Fixing that is structural work, and it’s exactly what my SOME framework exists for: Summarize, Organize, Modularize, Emphasize.

Google’s own tooling has caught up, too. Search Console added a generative AI performance report in late August, so the measurement excuse is gone.

One more thing. The Triple-P read is also where visibility work hands off to revenue architecture. Presence and Prominence problems get fixed in the content and its structure. Performance problems almost always trace upstream, to positioning, offer, or audience, and no amount of additional content fixes those.

The model made its public debut as a field note on the Fortin File, where I walked through the study that forced the unit change. This page is its permanent address.

If your rankings look healthy while your traffic and pipeline drift apart, that’s the Triple-P read I run in fractional engagements. Let’s have a chat.

---

Frequently Asked Questions