Barry Adams recently argued that Google creates an ecosystem of public loyalty. His take on the mechanics is that preferred sources, search profiles, and subscription links provide publishers with new tools to stay visible to trusted readers.
His article explains to companies with loyal audiences how to build their loyalty. The harder question is what happens to those who aren’t on anyone’s favorite list yet?
Google’s loyalty features allow users to choose which sources they want to appear more often. This creates a new discovery issue for sites who must gain notoriety before they can gain preference.
What Favorite Sources Do
Favorite Sources allows users to select publishers they want to see more of in search results.
Google launched the feature in the United States and India for Top Stories. He expanded globally in all supported languages in April. Then in May, Google introduces favorite sources to AI previews and AI mode.
In Top Stories, chosen sources appear more often, or in a dedicated “From your sources” section. In AI previews and AI mode, links from these sources are labeled with a badge so people can spot them. Google says searchers are twice as likely to access a preferred source, and more than 345,000 unique sources have been selected so far.
Barry’s article covers complete mechanics in depth.
When preference becomes distribution
The features support the same goal.
Search profileslaunched in June in the United States, offers its many subscribers a dedicated search page. A Follow button can surface more content from that source in Discover. And Subscription Binding allows paid readers to connect publisher subscriptions to Google accounts, so that content from paid subscriptions can be highlighted in Search, Discover, and other Google products.
Each feature rewards publishers people already know. This is a reasonable design choice, but it means the discovery layer becomes thinner for publishers who haven’t yet built that audience.
This is not the same as classic algorithmic filter bubbles. Preferred sources are different because people deliberately select websites from which they want to benefit more.
This changes the ethics of the debate. You can’t blame an algorithm for decisions people made on purpose. But the structural effect is similar. This is the filter bubble problem in a user-directed form.
The benefit extends to all features. Search profiles need 100,000+ followers on YouTube, Instagram or X, or 300,000 on TikTok. Subscription linking requires an existing subscriber base. Each feature is easier to activate with an established audience.
What custom queries add
Queries add another layer of personalization on top of your chosen sources.
Google’s Robbie Stein gave an example of how people search in AI mode. Instead of “Nashville restaurants,” people enter queries like “restaurants in Nashville but a friend has an allergy, we have dogs and want to sit outside.” This simple query gives Google more context about the user than a traditional search ever did.
Add to that source preferences and Google’s Personal Intelligence feature, which connects Gmail and Photos data to AI mode, and the picture becomes more individual.
An iPullRank experience published in May found a 46 percentage point increase in brand mentions for accounts connected to Personal Intelligence. Referenced brands increased from 23.9% to 66.8% of relevant responses in AI mode, with Gmail showing the largest effect. The test covered three accounts over 17 days with the Personal Intelligence option only.
The combined effect is a research experience that two people asking the same question cannot share. The query, sources and basic data for registered users are personalized, giving Google more individual context than traditional keyword search.
How content creators are trying to break through
The problem is achieving awareness before the preference exists. For publishers not part of a user’s chosen set, visibility must come from places that Google’s preferences layer does not fully control.
One option is to become the source cited by Preferred Sources. If the sources a user trusts reference your work, your content can still reach them. That means building your presence in podcasts, industry publications, original research, ChatGPT, social platforms, and peer recommendations—places where people encounter new sources and where AI systems can retrieve, cite, or mirror your work.
Another is to use the tools provided by Google. Search Central Documentation includes a deep link format and downloadable buttons that you can add to your site along with other calls to action. Deep linking takes users directly to the source preferences tool with the publisher URL pre-populated. Google says the buttons are intended to accompany social media follow prompts and newsletter signups.
Writing for custom queries is a third option to watch out for. People using AI mode give Google detailed context about themselves. Content with direct experience and depth beyond AI summaries can perform better in conversational search.
We have covered the growth of AI-driven citation models on these surfaces, and this pattern points in the same direction. Publishers cited in AI responses tend to have strong brand recognition across all channelsnot just the traditional top rankings.
None of these paths are guaranteed. Google has not disclosed the weight of preferred sources compared to other signals, and adoption numbers are still early. But these are the options that correspond to how the feature works.
What we don’t know
Google reported 345,000 unique curated sources, but did not report how many people enabled preferred sources.
If adoption is low, the structural effect on discovery is limited. If adoption grows alongside AI mode, what Sundar Picahi said in May this year has already exceeded 1 billion monthly active usersthe effect could be much greater.
Digiday reported in February that publishers cannot yet measure the effect of preferred sources on their traffic. There is no Search Console filter, so you can’t see how many people have added your site as a favorite source.
Google says Preferred Sources see a 2x click-through rate, but there’s no way to verify this number on your own site. In AI Previews and AI Mode, Google currently labels preferred sources with a badge rather than improving their ranking. How much this changes and when is an open question.
Looking to the future
Whether this creates significant barriers to discovery depends on adoption and how Google evaluates these signals against content quality and relevance. For businesses and research professionals, these features are already important. The question is how to become the source people choose before preference-based distribution becomes a bigger part of how search works.
More resources:
Featured Image: Konstantin Faraktinov/Shutterstock
