In 2023, I published a case study showing the impact of syndicated content on rankings across Google surfaces (including the 10-blue links, Top Stories, the News tab in Search, Google News, and Google Discover). The results underscored the importance of maintaining control over indexing when syndicating content and how that could impact which version of the article ranked in the SERPs.
For example, both versions of the content were often indexed and Google would often rank the syndicated version over the original. In addition, both versions sometimes ranked in the same SERP and I even saw site logos get mixed up in certain situations. You can check out my post to read the full details of the case study, but based on my analysis, I explained that noindexing the syndicated content was the path forward. Rel canonical was just a hint, it’s not foolproof, and some syndication partners push for having self-referencing canonicals on the syndicated content (which could obviously cause issues from an indexing and ranking perspective.)
Well, here we are in 2025 with AI Search on the scene, and growing. And now there’s publisher confusion about how syndicated content ranks across AI Search tools like ChatGPT, Perplexity, Gemini, and Claude. So, I decided to dig in and analyze a number of publishers syndicating content to third-party sites like Yahoo and MSN. My goal was to see how those articles ranked across AI Search, if the third-party publishers outranked the original content, and how the various AI search tools handled the situation.
Strap yourself in. The ride is about to get interesting.
Methodology:
I tested nine different publishers that were syndicating content to either Yahoo or MSN. Then I checked indexing in Google, visibility across Google surfaces, and then visibility across AI Search tools like ChatGPT, Perplexity, Gemini, and Claude based on prompts that made the most sense for the given articles. I did not simply use a keyword-driven approach that most would use in traditional Search.
Quick summary of findings: Like organic search, it’s still all over the place…
Syndication is still messy from a visibility standpoint across Google and now AI Search tools. Sometimes the original url ranked in AI search, sometimes the syndicated content on third-party sites like Yahoo or MSN ranked, and sometimes they both ranked. I did see canonicalization help for certain publishers, but that was not always foolproof.
In my opinion, noindexing the syndicated content is still the safest bet for ensuring your content ranks across AI Search tools and Google surfaces. I know that’s a big ask, but it’s clear that syndicating content without noindexing the urls on third-party sites can cause all sorts of ranking issues like I explained above. I’ll cover four examples below including screenshots of what I found. And I’m sorry for the blurring of results, but my point isn’t to focus on the sites… it’s about the impact of syndicating content on visibility across Google surfaces and AI tools.
Influence of traditional search rankings on AI Search for syndicated content:
We know that there’s a ton of overlap with strong Google rankings and what ranks in AI Search so I made sure to check rankings across Google surfaces for all urls I tested. There definitely were examples where AI Search followed Google rankings, but that wasn’t always the case. I’ll cover more about that in the findings and screenshots below, but I wanted to point that out.
So for those situations, the AI Search tools decided to either rank the canonical url or provide the syndicated content on third-party sites like Yahoo or MSN. Note, I saw much more of Yahoo showing up across AI Search tools than MSN. And in some cases, both urls ranked in the AI answer, with the syndicated urls sometimes ranking above the original content. Again, just another reason to noindex syndicated content on third-party sites if you can swing that in your contracts. I know that’s not easy to do, but that would alleviate the problem.
Examples from my testing of syndicated content across AI search tools (blinded):
Below, I’ll cover some quick examples of my findings based on testing nine different sites that are syndicating content to third-party sites. I think you’ll get the picture pretty quickly. Then I’ll end this post with some recommendations for site owners that are syndicating content or thinking about syndicating content in the future.
Example 1: Mixed results.
Google surfaces:
10-blue links: The syndication partner ranks (Yahoo in this case).
AIO: Neither url ranked in the AI Overview.
News tab: The original url ranked and the syndication partner did not rank.
Google News: The original url also ranked in Google News.
AI Search:
ChatGPT: The original url was cited prominently in the answer and in the sources list.
Perplexity: Third-party syndication partner cited prominently.
Here is the Perplexity sources view with Yahoo ranking #1:
Claude: The original url was cited, but Claude also says it failed to fetch the url for some reason. See my post about JavaScript rendering for other issues with AI search tools and rendering content. The original url was also in the answer. So the original url was not in the sources list, but did show up in the answer.
Gemini: Neither url was cited, which was interesting since the content did rank well across Google surfaces.
Example 2: A ranking mess.
Google surfaces:
AIO: The original content seemed to rank in the AIO, but it’s actually another site that picked up the syndicated content. So it’s not even the third-party site I was monitoring for this example. It’s just another reason to watch how you syndicate content. It can end up on many sites beyond your initial syndication partner.
10-blue links: The original url ranked in the 10-blue links at #4.
News tab: Neither url ranked, but the news tab was not returning much for the queries I was entering.
Google News: Similar to the News tab, Google News didn’t return much for the queries I was testing.
AI Search:
ChatGPT: The original url cited and third-party url didn’t rank at all.
Perplexity: Third-party syndication partner cited prominently at top source.
Claude: Both urls were cited and they were side-by-side in the answer from Claude. That’s fitting actually. :) And both are in the sources list with the original ranking higher in the list.
And here is the sources list with both ranking:
Gemini: Neither url was cited but no sources were cited at all. I saw this several times by the way. If I changed the prompt slightly to more of a question, Gemini did cite sources and the original url ranked. The syndicated content did not rank.
Example 3: More confusion.
Google surfaces:
Google Search: The syndicated url ranks #1 in the AIO and #1 in the 10-blue links.
News tab: Yahoo ranks #1 in the News tab.
Google News: Even though the syndicated content ranked across other Google surfaces, the original content ranked in Google News. Again, it’s all over the place.
AI Search:
ChatGPT: The syndicated url is cited as a source but in the “More” section below the sources list. The original url is not cited. And there are two urls from Yahoo listed there, one that’s the syndicated content and another article about a similar topic.
Perplexity: Like ChatGPT, the third-party url is cited as a source. And also like ChatGPT, there was a second article from Yahoo about a similar topic.
Here is the sources view showing both articles from Yahoo:
Claude: Again, both urls (the original and syndicated content) were cited and ranking right next to each other in the answer.
Gemini: Neither url was cited but the sources were super interesting for this prompt. Perplexity finance was cited several times in Gemini and filled up the sources list on the right side. Very interesting and something to watch out for if Perplexity expands verticals beyond finance.
Example 4: A unicorn for AI Search and syndicated content.
Google surfaces:
AIO: The original content ranks #1 in the AIO and #1 in the 10-blue links.
News tab: The original url ranks #1 in the News tab.
Google News: Like the News tab and 10-blue links, the original content ranks well in Google News.
Note, this was the cleanest example I found during my research. It was also interesting to see similar results over and over for the site I was analyzing. It’s like a unicorn for syndicated content and AI Search.
AI Search:
ChatGPT: The original article ranks prominently as a source.
Perplexity: Like ChatGPT, the original url is cited as a source, and #1 in the answer and sources list.
Here is the sources view in Perplexity:
Claude: The original url was cited but ranked below other sources. The syndicated content did not rank at all.
Gemini: Not only was the original url cited prominently in the answer, but so were several other articles from the publisher. Again, a unicorn. The syndicated content did not rank at all.
Syndicating content is still tough visibility-wise:
I’ll stop here after four examples since I think you get the picture… For publishers syndicating content, it’s a tough environment for making sure your original content ranks in Google Search and across AI search tools like ChatGPT, Perplexity, Claude, and Gemini. Syndication partners can often rank above the original content. Below, I’ll cover some key points and recommendations before ending the post.
Recommendations for site owners syndicating content:
- My testing of syndicated content in AI Search underscores the ranking challenges for the original publishers. As you can see above, the syndicated content ranks often, and sometimes along with the original content.
- In my opinion, this is what noindexing the syndicated content is the path forward if you want to ensure your original content gains the most visibility across Google and AI Search tools like ChatGPT, Perplexity, Claude, and Gemini.
- Canonicalization does seem to help in certain situations, but it’s not foolproof. Both urls are often indexed and the visibility across the various platforms is mixed. Sometimes it’s the original content ranking and sometimes it’s the syndicated content. But at a minimum, try to have the syndicated content canonicalized to the original urls (if noindex is not an option).
- Negotiate your syndication contracts wisely. Push for noindexing if you want your own content to rank in traditional search as well as AI search. A second choice would be having the syndicated content canonicalized to the original content.
- Run through this analysis for your own content. You can manually run through examples like I did above or you can leverage AI visibility tracking tools to run some tests. Note, you probably won’t have many prompts per AI visibility tool… so be smart with how you set up the tests. For example, they might have to be short and concise tests so you can reuse those prompts for other examples.
- Document everything in detail so you can track visibility changes over time across the various AI tools like ChatGPT, Perplexity, Claude, Gemini, and others. I believe we will continue to see a lot of change with the ranking systems across AI Search tools, so what’s happening today could change next week, month, or quarter. That’s both exciting and frustrating, but that’s the stage we are at with AI Search.
Summary: For syndicated content, control what you can control.
Syndicating content can be a minefield for organic search and it seems that’s the case for AI Search as well. Moving forward, I would control what you can control. That’s why I recommend noindexing syndicated content or at least canonicalizing that content to the original urls. If both urls get indexed (which they often do), then you cannot guarantee visibility for your original content over the syndicated content — other than the unicorn I covered earlier of course. :)
GG