If you have the same information as somebody else, yours should show up stronger.
As artificial intelligence reshapes how people discover information, Google's head of search offered a reassuring vision this week: that personalization, rather than erasing smaller voices, might actually amplify them. Liz Reid argued that knowing more about a reader's interests allows search to surface niche publishers a generic result would never find — a hopeful premise, but one she advanced without data, measurement tools, or independent verification. In an era when a single company mediates most of humanity's access to knowledge, the distance between a plausible argument and a proven one carries real weight for the publishers whose livelihoods depend on the answer.
- Small publishers face an existential fear that AI-personalized search will bury them beneath better-known outlets, making discovery by new readers nearly impossible.
- Google's VP of Search offered a counterintuitive reassurance — that personalization creates space for niche voices — but presented no supporting data, metrics, or measurement tools to substantiate the claim.
- A preferred sources feature lets readers designate trusted publishers for higher prominence, yet critics note it structurally favors outlets readers already know, offering little help to publishers trying to reach new audiences.
- An independent 17-day experiment found personalization did increase brand visibility in AI Mode, but the narrow scope leaves the broader question unanswered.
- Publishers are left in an uncomfortable position: Google controls the discovery layer and is asking for trust it has not yet earned with evidence.
Google's head of search made a case this week that personalization could actually help smaller publishers get discovered in the AI era — but offered no evidence to support it.
Liz Reid, speaking on the AI Inside podcast, pushed back against publisher fears that tailored search results would favor large, well-known outlets. Her argument: when Google understands a user's specific interests, it gains room to surface niche publishers and specialist reviewers that a generic, one-size-fits-all result would miss entirely. She called this pushing content "more into the tail" — toward the long tail of specialized publishers that broad searches overlook.
Reid also highlighted a preferred sources feature, which lets readers designate publishers they trust. When a source is marked as preferred, she said, its content should appear more prominently than the same information from a competitor. She was cooler on paywalled content, arguing that surfacing gated articles helps no one — and suggested Google should instead route subscribers directly to publications they already pay for.
The problem is that Reid offered none of this as demonstrated fact. No data showed personalization lifting small publisher traffic. No metrics confirmed that preferred-source status moves the needle. A small independent experiment did find that personalization increased brand visibility in AI Mode, but it ran for only 17 days across three accounts — far too narrow to draw conclusions.
There's also a structural tension Reid didn't fully address: preferred sources reward publishers a reader already knows. For an unknown outlet trying to reach new audiences, that feature offers nothing. Reid noted that top organic results still appear alongside preferred sources, but that's a different claim than saying personalization helps undiscovered publishers break through.
Google says it will keep expanding these features. Whether publishers see any real traffic benefit remains unproven. The company that controls how most people find information is asking publishers to trust that personalization will serve them — without providing the tools to verify whether it actually does.
Google's head of search made a case this week that might sound reassuring to publishers worried about disappearing in the age of AI: personalization, she argued, could actually help smaller outlets get discovered. The problem is she offered no evidence to back it up.
Liz Reid, Google's VP and Head of Search, laid out her thinking during an appearance on the AI Inside podcast. Publishers have been anxious that as Google's search results become more personalized—tailored to individual users rather than showing everyone the same ten blue links—smaller, less-known outlets would get buried. Reid's counterargument was straightforward: personalization does the opposite. When Google knows more about what you're actually looking for, she suggested, it has room to surface niche publishers and specialist reviewers that a generic search couldn't find.
Her logic hinged on a simple observation: a one-size-fits-all search result tends to look the same for everyone. But when a search engine has detailed signals about a user's interests—say, a history of reading about sustainable products—it can surface small merchants or specialized reviewers that match those preferences, even if the user never typed "eco-friendly" into the search box. Reid called this pushing "more into the tail," meaning the long tail of smaller, more specialized publishers that generic searches would miss.
Reid also pointed to a feature called preferred sources, which lets readers tell Google which publishers they trust. When someone marks a publication as a preferred source, Reid said, that signal should make that outlet's content appear more prominently than the same information from a competitor. "If you have the same information as somebody else, yours should show up stronger," she said. She was less enthusiastic about paywalls, noting that surfacing gated content does little good when most readers can't access it. Her solution: Google should route subscribers directly to the publications they already pay for.
But here's the catch: Reid presented none of this as fact. She offered no data showing that personalization actually helps small publishers gain visibility, no metrics showing that preferred-source status moves the needle on traffic, no measurement of any kind. An independent experiment by iPullRank on Google's Personal Intelligence feature did find that personalization increased how often seeded brands appeared in AI Mode, but that test ran for only 17 days across three opted-in accounts—hardly a comprehensive picture.
The absence of data matters because preferred sources, as Reid describes them, have a built-in limitation she didn't address: they reward publishers a reader already knows and trusts. For a small outlet trying to break through to new readers, preferred-source status is useless. Reid counters that preferred sources still surface top organic results alongside a user's chosen ones, but that's a different claim than saying personalization helps unknown publishers get discovered.
Google says it will continue expanding preferred sources and subscription features. Whether publishers actually see a traffic lift from any of this remains an open question. Reid's argument is worth testing against your own analytics, but taking it on faith would be premature. The company that controls how most people find information online is asking publishers to trust that personalization will help them—without providing the tools or data to verify whether it actually does.
Citações Notáveis
If the only thing you enter is a few keywords and it's unpersonalized, then everything kind of looks the same.— Liz Reid, Google VP and Head of Search
If you have the same information as somebody else, yours should show up stronger.— Liz Reid, on preferred sources
A Conversa do Hearth Outra perspectiva sobre a história
So Reid is saying personalization helps small publishers. What's the actual mechanism she's describing?
She's saying that when Google knows more about what you want—your interests, your reading history—it can surface niche publishers that match those preferences, rather than showing everyone the same generic results. It's the "long tail" idea.
That sounds plausible. Why is everyone skeptical?
Because she offered no data. No numbers showing it actually happens. And there's a logical problem: preferred sources reward publishers you already know. They don't help you discover new ones.
So it's circular.
Exactly. If you've never heard of a small publisher, marking it as a preferred source doesn't help. And Reid didn't address that gap.
Has anyone tested whether this actually works?
iPullRank ran a small experiment—17 days, three accounts—and found personalization did increase visibility for seeded brands. But that's not the same as proving it helps unknown publishers break through.
What would actually prove it?
Google would need to publish data showing traffic changes for small publishers after personalization and preferred sources rolled out. They haven't done that. They're asking publishers to trust the mechanism without showing the results.