Recently I caught up with Heiko Desruelle, chief product officer at Mediafin, about how his team is tackling one of the most pressing questions facing publishers right now: how to manage the growing reach of Google’s AI Overviews while preserving search traffic. 

In short, they have found a workaround which seemingly every news organisation can learn from and replicate.  

Mediafin has been experimenting with the “no snippet” metadata tag, a small technical switch with big implications. When applied, it prevents Google from generating AI Overviews, summary boxes, and even blue link descriptions, while still allowing the blue link itself to appear in search results. Documentation is here.

In August, they tested the approach on personal finance content. Late September, they rolled it out across their entire site. Below is a screenshot from search showing the links. No additional info, i.e., “snippets” are shown, just the title and blue link.

I was surprised to hear there was no significant decline in traffic so far. This suggests blocking snippets doesn’t necessarily mean losing visibility.

It’s early days, but the experiment highlights a path other publishers are exploring: a technical, non-confrontational way to protect content while maintaining visibility. 

This is huge.

The next step will be to apply the tags to overview and dossier pages, which Google continues to scrape for overviews.

Protecting content in 2025 is proving harder than ever. Archive.today is exploiting paywall exceptions, using Google referrers in incognito mode to unlock free articles. The problem compounds when AI platforms like Perplexity pull from these archives to bypass paywalls entirely. 

For Mediafin, and many others, this is forcing a rethink of “free article” strategies. Login walls and tighter trial periods are likely to be core to this strategy working.

There’s also a new layer of leakage: AI systems stitching together narratives from social media discussions about news articles. The quality is lower, but it still undermines the value of original reporting.

Heiko was clear about the financial stakes. Advertising still accounts for a large part of Mediafin’s revenue, which creates significant vulnerability in an environment where traffic-based models are under strain. Subscription growth is showing more promise for AI-era resilience. As AI overviews reduce the incentive to click through, the logic of customer-first subscription models only strengthens.

This isn’t just about resilience; it’s about control. A shift toward logged-in, paying customers creates relationships that bots and AI overviews can’t easily replicate.

Mediafin’s approach captures the tension many publishers feel: experimenting tactically with tools like “no snippet,” closing loopholes in paywall access, and at the same time rethinking the broader revenue model.

The other conclusion I can’t help thinking about: AI is putting traffic-based models under pressure, and customer-first subscription models are the most sustainable path forward. 

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

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