📊 What's Actually Happening

Two Studies Just Made This Commercially Urgent

Measurement standards matter more when there is something worth measuring. In June, two separate pieces of research put numbers on that.

Similarweb, published June 23, tracked user behavior across finance, travel, and beauty. Users who received an AI recommendation were 2.5 times more likely to visit that brand's site within seven days. Those visitors viewed nearly twice as many pages and stayed roughly twice as long. And more than 56 percent of that AI-influenced traffic arrived through branded search, compared with about 40 percent for standard visits.

Read that last number again. The influence happened in the AI conversation. The visit was recorded as branded search.

A preprint from Scrunch AI, posted to arXiv on June 9, joined opt-in clickstream data to users' ChatGPT, Claude, and Gemini conversations. Among users with no recent observed engagement with the brand, the effects were:

The gap between those two columns is the finding. A recommendation moved behavior roughly two to three times more than an incidental name-drop.

Which means counting AI mentions as if they are interchangeable may be the most expensive measurement mistake this industry makes next. Mentioned and recommended are not the same commercial event, and treating them as one number hides the difference

🎯 This Week's Strategic Lens

The Four P's Measure a Moment. Nothing Measures Whether It Holds.

Go back through the IAB pillars and notice what each one describes.

Presence: did you appear? Prominence: where did you appear? Portrayal: how were you described? Persuasion: did that appearance move someone?

Every one of those is a measurement of an event. A response happened. Something was true about that response.

None of them asks the question a CMO asks second.

Does it still happen next week?

Change the prompt. Add a constraint. Change the company size in the scenario. Ask a follow-up. Run it on a different platform. Run it after the next model update. Are you still there?

That is not a fifth P and I am not going to give it a clever name. It is a question, and it is the one I would add to any AI visibility report before I let it inform a budget decision.

The reason it matters is structural. Presence, prominence, portrayal, and persuasion can all be true once and false the following Tuesday. A brand can look excellent in a Monday scan and vanish from the same query on Friday because the model updated, or because the evidence behind the recommendation was thin enough that one new comparison article displaced it.

The distinction I keep coming back to:

One recommendation is an event. Repeated recommendation is a position.

You cannot build a strategy on an event. You can build one on a position.

And this is exactly where the two halves of this week's news meet. The IAB says single-response measurement is not measurement. The correct response to that is not just more queries. It is repeated queries over time, so you can see whether what you measured was a fluke or a foundation.

A brand that appears in eight out of ten runs of the same buying question has something. A brand that appeared once, screenshotted it, and put it in a board deck has a screenshot.

🛠 THE 10-MINUTE CMO MEASUREMENT CHECK

Pull up whatever AI visibility report your team last handed you. Run it against five questions.

1. Presence. Does this tell me whether we appear in the buying conversations that actually matter, not just the broad category term? "Best CRM platforms" and "best CRM for a 200-person B2B SaaS company replacing HubSpot" are different conversations. Your report should reflect the second one.

2. Prominence. When we appear, where are we? First, third, or a footnote after the recommendation has already been made?

3. Portrayal. How were we described? Which category were we placed in? Was anything about us wrong? Inaccurate portrayal at high prominence is worse than modest prominence with accurate portrayal.

4. Persuasion. Was the language confident or hedged? "A strong option for teams that need X" and "you could also consider X" are not the same recommendation, and the June research suggests they do not produce the same downstream behavior.

5. Persistence. Does the report tell me whether any of the above held across prompts, sessions, platforms, and weeks? Or is it a snapshot?

Then ask one question about the report itself, which is the IAB's real contribution:

Is this directional or decision-grade? How many queries, across how many intent types, run how often, with what variation range? If your vendor cannot answer that, you have a trend indicator. Useful. Just not something to move budget against.

If your report answers one or two of the five, you have visibility measurement. If it answers four, you are approaching Recommendation Intelligence. If it answers all five and tells you what changed and why, AI stops being a marketing metric and starts being acquisition intelligence.

🔔 CLOSING SIGNAL

AI visibility became measurable this month in a way it was not before. That is real progress and the IAB deserves credit for it.

Now comes the harder part, which is making it meaningful.

Presence tells you that you showed up. Prominence tells you where. Portrayal tells you how you were described. Persuasion tells you whether it moved anyone.

Persistence tells you whether any of it is a position or an accident.

And underneath all five sits the question this month quietly put on the table: can the evidence itself withstand inspection? Yours, your competitor's, and your vendor's.

The brands that win the next phase will not be the ones with the most measurements. They will be the ones who can explain which changes mattered, why they happened, and what to do next.