Personalization can help small publishers


Google vice president and head of search Liz Reid said personalized search and preferred sources can help small publishers gain visibility, addressing concerns that personalization makes them less accessible.

Reid shared his views on AI Inside Podcastin the same interview where she told editors that the key to AI visibility lies in create content that resonates with people. When the hosts expressed concerns that personalization could make some editors “more invisible,” she took the opposite position.

Personalization as a path of discovery

Reid suggests that generic, universal search results tend to cause everyone to see the same results. She mentioned that when there are more detailed signals about what a user is looking for, it opens up opportunities for niche publishers to be seen.

Reid said:

“If the only thing you enter is a few keywords and it’s not personalized, then it all looks the same.”

She gave the example of someone searching for “eco-friendly” brands who never specifically uses that term. In this situation, personalization could surface small merchants or specialized reviewers matching the user’s preferences.

Reid added that personalization favors creators and journalists who focus on specific topics and are difficult to match to queries. She described it as “pushing more in the tail”.

Favorite sources and subscription question

In arguing that personalization is beneficial for websites, Reid mentioned Preferred Sources, a search feature that allows users to tell Google which publishers they prefer.

When someone likes a particular website and lists it as a favorite source, she said, that signal can help the publisher’s content appear more prominently than the same information elsewhere.

She added:

“If you have the same information as someone else, yours should appear stronger.”

Reid wasn’t as positive about paywalls. She says surfacing gated content doesn’t do much when most people can’t read it, and publishers who add a paywall and then watch traffic drop see the predictable result.

“Yes, that’s what will happen if you charge,” she said. The solution she proposes is for Google to route subscribers to the publishers they already pay for.

The complaint is delivered without data

Reid didn’t provide any data in the interview to show that personalization helps small publishers or that preferred source status makes their content more visible. His argument is similar to his recurring “bounce clicks“AI-related traffic loss explained.

A iPullRank Experience on Google’s Personal Intelligence feature found that personal signals increased the frequency with which predefined brands appeared in AI mode. It showed that personalization adds to the foundation of the web rather than replacing it. Although the test covered three accounts over 17 days, these were only activated accounts.

Why it matters

Reid’s position is that insider sources can help smaller sites. However, this is still a claim and Google hasn’t provided a way to determine whether personalization or preferred source status actually impacts their viewability.

There’s a problem that Reid doesn’t address. Preferred Source status rewards publishers that a reader already trusts, which doesn’t help a site the reader hasn’t yet heard of.

I have argued that Google’s loyalty tools, like Preferred Sources, create a problem for publishers who are not already on someone’s list. Reid takes the opposite view, saying that preferred sources always surface the best organic results alongside those chosen by the user.

Looking to the future

Reid notes that Google will continue to expand its preferred sources and subscription features. Whether publishers see improvement depends on metrics that Google didn’t provide. In the meantime, the case for personalization helping publishers is worth testing against your own analytics rather than relying on them.


Featured image: screenshot from youtube.com/@aiinsideshow, June 2026.



Source link

Leave a Reply

Your email address will not be published. Required fields are marked *