September 9th-September 15th // Estimated Reading Time: 7 minutes

Google rolls out pay-per-value AI licensing program to publishers🔒 - Google is quietly expanding a pilot that pays publishers when their content "significantly" shapes AI responses in Gemini, AI Overviews, and AI mode. Publishers in the program see a monthly earnings figure inside Search Console, but Google shares little about how that number gets calculated. Dozens of publishers have reportedly been approached, with small and mid-sized publishers in particular seeming most receptive. But some have called the payouts "peanuts" compared to their ad revenue. Critics see it less as fair compensation and more as a hedge against future legal and regulatory pressure over AI content use.

Opinion: On one hand, this represents Google doing much more for small- and mid-sized publishers than OpenAI or Anthropic. On the other hand, this represents Google not doing nearly enough to compensate publishers for killing their site traffic and consequently their ad revenue. Google should at least be transparent about what goes into the compensation model, so publishers can plan and optimize accordingly.

Chrome has a new way to measure ad overload on the web - Chrome added four new ad metrics to track the ad experience on webpages: ad count, ad density, and ad weight measured by both network and CPU usage. The metrics use the same underlying data that powers Core Web Vitals, which measure page performance. Google isn't setting benchmarks yet and considers the metrics experimental, collecting feedback from the ecosystem before deciding next steps. Publishers likely won't know what "good" looks like until they know if and how buyers are factoring the data into media-buying decisions. 

Opinion: As the AdExchanger story notes, it's ironic that Google is pushing publishers towards a cleaner ad experience while AI Overviews sap the traffic their ad business depends on. It’ll be interesting to see how these new metrics correlate with ad performance, if at all. If they do, measurement providers and marketers will be quick to use them as another signal.

Ad spend forecasts revised upward as more ad dollars are handled by AI tools🔒 - The IAB now expects US ad spend to grow 12.3% this year, up from its earlier 9.5% estimate, thanks to a strong first half driven by the Winter Olympics and World Cup. Madison & Wall estimates AI-powered campaigns like Google's Performance Max and Meta's Advantage+ now make up 12% of US ad spend, up from just 2% in 2023, and that share could hit 27% by 2030. Agency execs told Digiday some clients already route 30% to 50% of their budgets through these automated tools. One practitioner put it bluntly: Automation is heading toward 90% and won't stop at 20% or 30%.

Opinion: The automated black-box tools have an interesting conundrum to solve: no one wants to say anything nice about them publicly, yet marketers are investing more and more into them. The walled gardens have clearly built something valuable that gets marketers to pony up, even if they don’t like it. Automation is definitely the future, but the current black-box solutions probably get disrupted before ever getting to 90% of spend.   

TV manufacturer Telly touts programmatic home screen ads - Telly, the startup known for giving away free smart TVs in exchange for viewer data and ad exposure, is opening up its home screen ads to programmatic buying. The ads run on a second screen below the main display, delivered through Magnite's SpringServe with ad demand from The Trade Desk, Teads, and others. Telly is also collaborating with the IAB Tech Lab to standardize specs for home screen ads in the bid stream, aiming to bring the format the standardization it needs to attract more spend.

Opinion: Smart TV home screen ad standardization is interesting. Smart TV home screen ads are arguably the most valuable piece of non-tentpole ad real estate available today. But they are difficult to buy because they are sold by each individual OEM, with unique terms for each buy. Standardization would make it easier to buy and scale, but it would hurt the direct sales efforts of the biggest OEMs (e.g. Samsung, LG, Roku) in the space. 

As Google's ad tech case nears its end, the antitrust fight moves to AI🔒 - Google lost both its search and ad tech antitrust cases but avoided forced breakups, with judges opting for behavioral remedies instead. Attention is shifting to how Google's AI products, including AI Max and its AI Commerce platform, could give Google an unfair advantage and draw fresh regulatory scrutiny. Critics say Google is building a monopoly across the AI commerce chain and want the EU to regulate it under the Digital Markets Act before it becomes entrenched.

Opinion: It took years of litigation to land on behavioral remedies for Google's search and ad tech monopolies, and even those still aren’t final. Any real check on Google’s anticompetitive AI advantage is years behind already.

OpenAI's next ChatGPT ad format: click to chat, not to site🔒 - OpenAI is testing a new AI-native ad format, called Sponsored Agent, that turns an ad click into a conversation with a brand's AI agent instead of sending users to a website. Wayfair is among the early testers, describing it as a way to explore conversational commerce while testing guardrails around accuracy and transparency.

Opinion: If the conversation happens inside ChatGPT, on OpenAI's platform, does OpenAI take over the customer relationship? To a certain extent, yes. Brands, beware.

Publishers are spending hundreds of millions a month on paid search🔒 - Similarweb data shows the top 100 media outlets spent an estimated $113M on paid search in July, up 41% year over year and 274% over the past three years. Forbes alone spent $72.2M that month, more than eight times what it spent three years ago. Organic traffic across major publishers like Forbes, CNN, and USA Today fell 24% to 29% over the same span. Buying traffic isn't new, but the scale of the spend shows how much publishers now depend on paid traffic to replace what zero-click search has taken away. 

Opinion:

Thanks for reading!