TL;DR
- USA Today's Google referral traffic share fell from over 70% to about 40%, pushing it toward Threads, TikTok, Reddit and AI-licensing deals with Meta and Microsoft.
- Ziff Davis ties about 17.5% of its revenue specifically to search, one of the few public search-exposure figures shared at the September 2026 summit.
- Future Publishing's GEO product has signed more than 30 clients with successful renewals, per Digiday's recap of the September 2026 Publishing Summit.
- Reuters' machine-learning-driven dynamic paywall now generates more ad revenue than it did before the paywall existed, Digiday reported on September 28, 2026.
Publishers at Digiday's September 2026 Publishing Summit shared concrete numbers on what replaces collapsing Google search traffic, and the standout data point is that Future Publishing's generative engine optimization (GEO) product has already signed more than 30 paying clients with renewals among them. That single number is the clearest evidence yet that GEO has moved from a theoretical service to one buyers actually renew.
What publishers reported at the summit
USA Today: referral share cut nearly in half
USA Today's Google referral traffic share fell from over 70% to about 40%, according to Digiday's recap of the September 2026 summit. The publisher is responding by pushing audience and distribution toward Threads, TikTok and Reddit, alongside AI-licensing deals with Meta and Microsoft. Digiday did not report the dollar value of those licensing deals or the specific timeframe over which the referral share declined.
The Verge: Google traffic is now "incremental"
The Verge now treats Google traffic as merely incremental to its business rather than central to it, per the summit recap. Digiday did not report what The Verge considers central to its business instead, or what share of traffic Google referrals still represent.
Ziff Davis: putting a number on search exposure
Ziff Davis ties about 17.5% of its revenue specifically to search, giving one of the few publicly shared figures quantifying how exposed a large publisher is to search-driven revenue. Digiday did not report whether that figure has changed over time or how it compares to Ziff Davis's exposure to other channels.
Future Publishing: GEO as a renewing client service
Future Publishing's GEO product, which optimizes content and visibility for generative AI engines on behalf of clients, has signed more than 30 clients, and Digiday's recap specifically notes successful renewals among them. Future Publishing is a media company selling GEO as a service to its own clients, not a marketing agency, and Digiday's report does not name Future Publishing as a competitor to any marketing agency; it is a different category of business selling a comparable service.
Reuters: a paywall that earns more than no paywall did
Reuters' machine-learning-driven dynamic paywall now generates more ad revenue than the business did before the paywall existed. Digiday did not report the specific mechanism by which the dynamic paywall increases ad revenue, or what percentage increase it represents.
Why Future Publishing's number matters more than it looks
Most of what gets written about generative engine optimization is forward-looking: predictions about how answer engines might change discovery, or frameworks for what GEO work could look like. Future Publishing's 30-plus signed clients, with renewals, is a different kind of evidence. Clients do not renew a service that is not producing a result they are willing to keep paying for. That makes this data point, as reported by Digiday in its recap of the September 2026 Publishing Summit, one of the more concrete public signals that paid GEO work has commercial traction right now, not just as a concept.
It is also worth noting what the report does not say. Digiday's recap does not disclose how much Future Publishing charges for its GEO product, how long a typical contract runs, which industries its 30-plus clients come from, or what specific renewal rate it is seeing beyond the fact that renewals happened. Those gaps matter: a renewal count alone does not tell a buyer whether the product works for a business its own size, in its own category, at a price it can justify. What the number does establish, reliably, is that the underlying premise of paid GEO work, that a business will pay a specialist to optimize its visibility inside AI answer engines, has cleared the first commercial test of being worth paying for twice.
It is worth being precise about what Future Publishing is. It is a media company, the publisher behind a large portfolio of specialist magazines and websites, that built a GEO product and sells it to outside clients as a service line. That is a different category of business from a marketing agency, even though the service it sells, GEO, is the same category of work that marketing agencies also offer. Digiday's report does not describe Future Publishing as a competitor to marketing agencies, and nothing in the source supports treating the two as interchangeable.
Summit numbers compared
The table below lists only the figures and facts Digiday's recap attributes to each publisher.
| Publisher | Reported figure or detail | Context |
|---|---|---|
| USA Today | Google referral share fell from over 70% to about 40% | Shifting toward Threads, TikTok, Reddit, plus AI-licensing deals with Meta and Microsoft |
| The Verge | Google traffic now "incremental" | No longer treated as central to the business |
| Ziff Davis | About 17.5% of revenue tied to search | One of the few publicly shared search-exposure figures |
| Future Publishing | 30+ clients signed for its GEO product | Includes successful renewals |
| Reuters | Dynamic paywall now generates more ad revenue than before it existed | Paywall is machine-learning-driven |
What this means for Thai marketers
Digiday's recap does not mention Thailand or any Southeast Asian market, so the numbers above describe US and UK publishers, not the Thai market. What transfers is the underlying argument, not the figures themselves.
The case for treating generative engine optimization as a current budget line, rather than something to revisit next year, rests on exactly the kind of evidence this summit produced: a business with an actual renewing client base for the service, not just a pitch deck about it. A business deciding whether to fund GEO work now or wait typically weighs two things: whether answer engines are already sending a measurable share of its audience, and whether a credible provider can show the work holds up past a first contract. Future Publishing's renewal data, as limited as the public detail is, speaks to the second question. It does not answer the first question for any individual Thai business, which still needs its own data on how much of its audience already arrives through AI answer engines before deciding how much to invest.
The other numbers from the summit are useful mainly as a checklist of questions to ask about one's own traffic mix before treating GEO as optional. USA Today's drop from over 70% to about 40% of referral traffic shows how fast a single channel's share can move once an alternative discovery surface, in that case AI-driven platforms and social apps, starts pulling attention away from traditional search results. The Verge's description of Google traffic as merely incremental is a useful prompt for any business to ask what its own primary traffic source actually is today, rather than assuming it is still search by default. Reuters' dynamic paywall result is a reminder that the revenue side of a post-search strategy is not only about replacing lost traffic; it can also come from monetizing the audience a publisher already has more precisely, which is a separate lever from discovery work entirely.
Ziff Davis's 17.5% search-revenue figure is also a useful prompt rather than a benchmark. A Thai business does not need to match that number to find it useful; it is evidence that at least one large publisher has gone to the trouble of quantifying its search exposure precisely, which is itself a reasonable first step before any GEO or AI-SEO budget conversation: knowing what share of traffic and revenue is currently search-dependent before deciding how urgently to diversify away from it.
FAQ: GEO and the post-search publishing summit
Does this apply to Thailand or Thai publishers?
Not stated by the source. Digiday's recap covers USA Today, The Verge, Ziff Davis, Future Publishing and Reuters, all US or UK businesses, and does not mention Thailand or any Southeast Asian publisher or market.
Is Future Publishing a marketing agency and a competitor to Relevant Audience?
No. Future Publishing is a media company that built a GEO product and sells it as a client service; Digiday's report does not describe it as a marketing agency, and it is a different category of business from a marketing agency even though it sells a comparable GEO service.
How many clients does Future Publishing's GEO product have?
More than 30, according to Digiday's recap, which also notes successful renewals among those clients. The report did not break down pricing, contract length, or which industries those clients are in.
What exactly caused USA Today's traffic share to drop?
Digiday's recap reports the size of the drop, from over 70% to about 40% of referral traffic from Google, but does not report the specific causes behind it or the exact timeframe over which it happened.
Does a dynamic paywall like Reuters' always increase ad revenue?
Not stated by the source. Digiday reported that Reuters' specific machine-learning-driven dynamic paywall now generates more ad revenue than the business did before the paywall existed, but did not report whether this result generalizes to other publishers or paywall designs.
In Digiday's recap, "post-search" describes a publisher's shift away from depending on Google referral traffic as its primary source of audience and revenue, toward a mix of social platforms, AI-licensing deals, subscription and paywall revenue, and services like GEO sold to other businesses. The report frames each publisher's response, USA Today's platform shift, The Verge's reclassification of Google traffic, Ziff Davis's search-exposure figure, Future Publishing's GEO product, and Reuters' paywall, as five different answers to the same underlying pressure rather than one shared playbook. A business weighing whether 2026 is the year to fund GEO work, rather than wait another year, can treat Future Publishing's renewal data as one data point among several worth checking before deciding.
