A few weeks ago, a Digiday reporter working on a story about new Ozone benchmarking data reached out to me. The numbers she shared were quite interesting. In Q2 2026, publisher ad request volumes had fallen 32% to 37% year over year in the U.S. and 39% to 41% in the U.K.

But most interestingly, at least in the U.K., rising eCPMs made up most of the difference. In June, average eCPMs were up about 30% year over year in the U.K. and 7% in the U.S.

It’s only one set of data in two markets, but the conclusion is that, at least in the U.K., we’re starting to see signs that a second-order effect of the post-traffic era is the supply-demand curve for publisher inventory is inverting. Buyers have fewer impressions to buy so, at least in some cases, they’re paying more for what’s left. 

Is this a sign of things to come? Maybe. 

I think what’s happening is that, as supply goes down, we’re starting to see a bifurcation in the market between publishers who can demonstrate the quality of their inventory to buyers and those who can’t. The market is willing to pay more for the former (especially when there’s less of it to go around) but not for the latter. 

As I told Jessica Davies for her story: “What you’re seeing is publishers realising that buying into the 15-year programmatic strategy of ‘infinite supply’ has effectively just steadily driven CPMs down, and AI-driven traffic declines have now killed the idea that there actually is infinite supply. The opportunity sits with publishers that can surface meaningful signals through first-party, contextual, attention, or other data to show that their impressions meet the quality thresholds the buy side is now actively seeking.”

In addition, as I told Digiday, some publishers are “deliberately shaping supply, cutting ad load and low-value bid requests to protect or increase attention and price.”

The evidence this works keeps arriving: A new Lumen Research study with Mail Metro Media found that cutting a simulated page from 15 ads to five lifted the share of readers who actually viewed an ad from 53% to 78%, with 4.2x higher spontaneous recall and an 8% lift in purchase intent.

If you read one deeper piece on this, make it Mimmo Palmieri’s essay in What’s New in Publishing. Palmieri, who spent more than a decade leading yield, data, and measurement at Condé Nast and the Financial Times, saw the crunch coming a year ago in Condé Nast’s own numbers. 

His argument is the same as mine: Programmatic was built on the assumption of abundance, from RTB plumbing to DSP bidding algorithms, and scarcity changes the math on both sides. His warning is aimed at yield teams tempted to answer falling ad requests with volume through more ad units, bought traffic, or lower-quality inventory. 

That treats a macro shift as temporary and dilutes the very quality signals now commanding a premium.

The better response, he argues, is to make fewer, better impressions worth more to protect premium supply for direct and guaranteed deals (already more than three-quarters of U.S. programmatic spend, per eMarketer), prove quality through first-party, contextual, and attention data, and move the headline metrics from pageviews and RPM to retention and lifetime value.

The cultural and organisational change, he notes, may be harder than the revenue hit itself.

This is not just a U.S. and U.K. story either. As I talked about in my report from South Asia, Times Internet COO Puneet Gupt is forecasting the same dynamics in India, expecting display eCPMs there to rise at least 30% over the next few years, and HT Digital says it has returned to double-digit CPM growth after deliberately decoupling revenue from traffic.

And as for the difference between the U.S. and the U.K., my thesis is the dynamics and trajectory of the U.S. market explain the gap. The U.S. still carries a long tail of commodity supply that dilutes yields, which is why repricing is showing up more slowly there. 

My prediction, as I said to Digiday, is that it accelerates. The thing to watch over the rest of the year is how this dynamic evolves and which publishers report falling traffic alongside rising ad revenue. That will give us more insight and depth on the variables of supply and demand — and the value of publishers signals to change the math — during this transformation.

If you’d like to subscribe to my bi-weekly newsletter, INMA members can do so here.

Banner art: Adobe stock image by bestforbest.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .