Dec 19 2025
Core Web Vitals Champ: Open Source Versus Proprietary Platforms via @sejournal, @martinibuster

The Core Web Vitals Technology Report by the open source HTTPArchive community ranks content management systems by how well they perform on Google’s Core Web Vitals (CWV). The November 2025 data shows a significant gap between platforms with the highest ranked CMS scoring 84.87% of sites passing CWV, while the lowest ranked CMS scored 46.28%.

What’s of interest this month is that the top three Core Web Vitals champs are all closed source proprietary platforms while the open source systems were at the bottom of the pack.

Importance Of Core Web Vitals

Core Web Vitals (CWV) are metrics created by Google to measure how fast, stable, and responsive a website feels to users. Websites that load quickly and respond smoothly keep visitors engaged and tend to perform better in terms of sales, reads, and add impressions, while sites that fall short frustrate users, increase bounce rates, and perform less well for business goals. CWV scores reflect the quality of the user experience and how a site performs under real-world conditions.

How the Data Is Collected

The CWV Technology Report combines two public datasets.

The Chrome UX Report (CrUX) uses data from Chrome users who opt in to share performance statistics as they browse. This reflects how real users experience websites.
The HTTP Archive runs lab-based tests that analyze how sites are built and whether they follow performance best practices.

Together, the report I generated provides a snapshot of how each content management system performs on Core Web Vitals.

Ranking By November 2025 CWV Score

Duda Is The Number One Ranked Core Web Vitals Champ

Duda ranked first in November 2025, with 84.87% of sites built on the platform delivering a passing Core Web Vitals score. It was the only platform in this comparison where more than four out of five sites achieved a good CWV score. Duda has consistently ranked #1 for Core Web Vitals for several years now.

Wix Ranked #2

Wix ranked second, with 74.86% of sites passing CWV. While it trailed Duda by ten percentage points, Wix was just about four percentage points ahead of the third place CMS in this comparison.

Squarespace Ranked #3

Squarespace ranked third, at 70.39%. Its CWV pass rate placed it closer to Wix than to Drupal, maintaining a clear position in the top three ranked publishing platforms.

Drupal Ranked #4

Drupal ranked fourth, with 63.27% of sites passing CWV. That score put Drupal in the middle of the comparison, below the three private label site builders. This is a curious situation because the bottom three CMS’s in this comparison are all open source platforms.

Joomla Ranked #5

Joomla ranked fifth, at 56.92%. While more than half of Joomla sites passed CWV, the platform remained well behind the top performers.

WordPress Ranked Last at position #6

WordPress ranked last, with 46.28% of sites passing Core Web Vitals. Fewer than half of WordPress sites met the CWV thresholds in this snapshot. What’s notable about WordPress’s poor ranking is that it lags behind the fifth place Joomla by about ten percentage points. So not only is WordPress ranked last in this comparison, it’s decisively last.

Why the Numbers Matter

Core Web Vitals scores translate into measurable differences in how users experience websites. Platforms at the top of the ranking deliver faster and more stable experiences across a larger share of sites, while platforms at the bottom expose a greater number of users to slower and less responsive pages. The gap between Duda and WordPress in the November 2025 comparison was nearly 40 percentage points, 38.59 percentage points.

While an argument can be made that the WordPress ecosystem of plugins and themes may be to blame for the low CWV scores, the fact remains that WordPress is dead last in this comparison. Perhaps WordPress needs to become more proactive about how themes and plugins perform, such as come up with standards that they have to meet in order to gain a performance certification. That might cause plugin and theme makers to prioritize performance.

Do Content Management Systems Matter For Ranking?

I have mentioned this before and will repeat it this month. There have been discussions and debates about whether the choice of content management system affects search rankings. Some argue that plugins and flexibility make WordPress easier to rank in Google. But the fact is that private platforms like Duda, Wix, and Squarespace have all focused on providing competitive SEO functionalities that automate a wide range of technical SEO tasks.

Some people insist that Core Web Vitals make a significant contribution to their rankings and I believe them. But in general, the fact is that CWV performance is a minor ranking factor.

Nevertheless, performance still matters for outcomes that are immediate and measurable, such as user experience and conversions, which means that the November 2025 HTTPArchive Technology Report should not be ignored.

The HTTPArchive report is available here but it will be going away and replaced very soon. I’ve tried the new report and, unless I missed something, it lacks a way to constrain the report by date.

Featured Image by Shutterstock/Red Fox studio

Ecommerce MGMT 0 Comments
Dec 19 2025
Google Says Ranking Systems Reward Content Made For Humans via @sejournal, @martinibuster

Google’s Danny Sullivan discussed SEO and AI where they observed that their ranking systems are tuned for one thing, regardless if it’s classic search or AI search. What he talked about was optimizing for people, which is something I suspect the search marketing industry will increasingly be talking about.

Nothing New You Need To Be Doing For AI Search

The first thing Danny Sullivan discussed was that despite there being new search experiences powered by AI there isn’t anything new that they need to be doing.

John Mueller asked:

“So everything kind of around AI, or is this really a new thing? It feels like these fads come and go. Is AI in fad? How do you think?”

Danny Sullivan responded:

“Oh gosh, my favorite thing is that we should be calling it LMNOPEO because there’s just so many acronyms for it. It’s GEO for generative engine optimization or AEO for answer engine optimization and AIEO. I don’t know. There’s so many different names for it.

I used to write about SEO and search. I did that for like 20 years. And part of me is just so relieved. I don’t have to do that aspect of it anymore to try to keep up with everything that people are wondering about.

And on the other hand, you still have to kind of keep up on it because we still try to explain to people what’s going on. And I think the good news is like, There’s not a lot you actually really need to be worrying about.

It’s understandable. I think people keep having these questions, right? I mean, you see search formats changing, you see all sorts of things happening and you wonder, well, is there something new I should be doing? Totally get that.

And remember, we, John and I and others, we all came together because we had this blog post we did in May, which we’ll drop a link to or we’ll point you to somehow to it, but it was… we were getting asked again and again, well, what should we be doing? What should we be thinking about?

And we all put our heads together and we talked with the engineers and everything else. So we came up with nothing really that different.”

Google’s Systems Are Tuned To Rank Human Optimized Content

Danny Sullivan next turned to discussing what Google’s systems are designed to rank, which is content that satisfies humans. Robbie Stein, currently Vice President of Product for Google Search, recently discussed the signals Google uses to identify helpful content, discussing how human feedback contributes to helping ranking systems understand what helpful content looks like.

While Danny didn’t get into exact details about the helpfulness signals the way Stein did, Danny’s comments confirmed the underlying point that Robbie Stein was making about how their systems are tuned to identify content that satisfies humans.

Danny continued explaining what SEOs and creators should know about Google’s ranking systems. He began by acknowledging that it’s reasonable that people see a different search experience and conclude that they must be doing something different.

He explained:

“…I think people really see stuff and they think they want to be doing something different. …It is the natural reaction you have, but we talk about sort of this North Star or the point that you should be heading to.”

Next he explained how all of Google’s ranking systems are engineered to rank content that was made for humans and specifically calls out content that is created for search engines as examples of what not to do.

Danny continued his answer:

“And when it comes to all of our ranking systems, it’s about how are we trying to reward content that we think is great for people, that it was written for human beings in mind, not written for search algorithms, not written for LLMs, not written for LMNO, PEO, whatever you want to call it.

It’s that everything we do and all the things that we tailor and all the things that we try to improve, it’s all about how do we reward content that human beings find satisfying and say, that was what I was looking for, that’s what I needed. So if all of our systems are lining up with that, it’s that thing about you’re going to be ahead of it if you’re already doing that.

To whereas the more you’re trying to… Optimize or GEO or whatever you think it is for a specific kind of system, the more you’re potentially going to get away from the main goal, especially if those systems improve and get better, then you’re kind of having to shift and play a lot of catch up.

So, you know, we’re going to talk about some of that stuff here with the big caveat, we’re only talking about Google, right? That’s who we work for. So we don’t say what, anybody else’s AI search, chat search, whatever you want to kind of deal with and kind of go with it from there. But we’ll talk about how we look at things and how it works.”

What Danny is clearly saying is that Google is tuned to rank content that’s written for humans and that optimizing for specific LLMs sets up a situation where it could backfire.

Why Optimizing For LLMs Is Misguided

Although Danny didn’t mention it, this is the right moment to point out that OpenAI, Perplexity, and Claude together have a total traffic referral volume of less than 1%. So it’s clearly a mistake to optimize content for LLMs at the risk of losing significant traffic from search engines.

Content that is genuinely satisfying to people remains aligned with what Google’s systems are built to reward.

Why SEOs Don’t Believe Google

Google’s insistence that their algorithms are tuned toward user satisfaction is not new. They have been saying it for over two decades, and over the years it has been a given that Google was overstating their technology. That is no longer the case.

Arguably, since at least 2018’s Medic broad core update, Google has been making genuine strides toward actually delivering search results that are influenced by user behavior signals that guide Google’s machines toward understanding what kind of content people like, plus AI and neural networks that are better able to match content to a search query.

If there is any doubt about this, check out the interview with Robbie Stein, where he explains exactly how human feedback, in aggregate, influences the search results.

Is Human Optimized Content The New SEO?

So now we are at a point where links no longer are the top ranking criteria. Google’s systems have the ability to understand queries and content and match one to the other. User behavior data, which has been a part of Google’s algorithms since at least 2004, plays a strong role in helping Google understand what kinds of content satisfy users.

It may be well past time for SEOs and creators to let go of the old SEO playbooks and start focusing on optimizing their websites for humans.

Featured Image by Shutterstock/Bas Nastassia

Ecommerce MGMT 0 Comments
Dec 18 2025
Coursera Acquiring Udemy via @sejournal, @martinibuster

Coursera has agreed to acquire Udemy in a stock-for-stock transaction that will combine two large online learning platforms with consumer and enterprise businesses.

Under the terms of the deal, each Udemy share will be exchanged for 0.800 Coursera shares. Following the transaction, Coursera shareholders will own approximately 59% of the combined company, while Udemy shareholders will own about 41%. The merged company will continue operating as Coursera, Inc., headquartered in Mountain View, California. Greg Hart will remain CEO, and Coursera’s Andrew Ng will serve as chairman. The companies expect the transaction to close in the second half of 2026, subject to shareholder approval and regulatory clearances.

Coursera’s platform is built around partnerships with universities, institutions, and industry organizations, with a focus on credentialed learning programs. Udemy operates an open marketplace of instructors and provides training programs used by enterprise customers. The combined company is expected to offer academic courses, professional skills training, and enterprise learning programs through a single platform.

The companies report a combined total of more than 270 million registered learners and nearly 19,000 enterprise customers. Coursera contributes institutional partnerships and credential-focused offerings, while Udemy contributes a large instructor marketplace and a broad enterprise customer base. Udemy generates a majority of its revenue outside North America, while Coursera generates a larger share of revenue in the United States.

If completed, the transaction will bring together institutional learning programs and an open instructor marketplace within a single company.

General reaction online was surprise, with Udemy instructor unsure about where he stood, writing on X:

“I can’t tell what this acquisition by Coursera means for my future as a Udemy instructor. Time will tell.
I will definitely keep on teaching – on one platform or another.

But learning that a brand that was THE main part of my professional life for the last 10 years will go away is really very, very sad.”

Read more at the Coursera website:

Coursera to combine with Udemy

Featured Image by Shutterstock/ShutterStockies

Ecommerce MGMT 0 Comments
Dec 18 2025
Google’s Robby Stein Names 5 SEO Factors For AI Mode via @sejournal, @martinibuster

Robby Stein, Vice President of Product for Google Search, recently sat down for an interview where he answered questions about how Google’s AI Mode handles quality, how Google evaluates helpfulness, and how it leverages its experience with search to identify which content is helpful, including metrics like clicks. He also outlined five quality SEO-related factors used for AI Mode.

How Google Controls Hallucinations

Stein answered a question about hallucinations, where an AI lies in its answers. He said that the quality systems within AI Mode are based on everything Google has learned about quality from 25 years of experience with classic search. The systems that determine what links to show and whether content is good are encoded within the model and are based on Google’s experience with classic search.

The interviewer asked:

“These models are non-deterministic and they hallucinate occasionally… how do you protect against that? How do you make sure the core experience of searching on Google remains consistent and high quality?”

Robby Stein answered:

“Yeah, I mean, the good news is this is not new. While AI and generative AI in this way is frontier, thinking about quality systems for information is something that’s been happening for 20, 25 years.

And so all of these AI systems are built on top of those. There’s an incredibly rigorous approach to understanding, for a given question, is this good information? Are these the right links? Are these the right things that a user would value?

What’s all the signals and information that are available to know what the best things are to show someone. That’s all encoded in the model and how the model’s reasoning and using Google search as a tool to find you information.

So it’s building on that history. It’s not starting from scratch because it’s able to say, oh, okay, Robbie wants to go on this trip and is looking up cool restaurants in some neighborhood.

What are the things that people who are doing that have been relying on on Google for all these years? We kind of know what those resources are we can show you right there. And so I think that helps a lot.

And then obviously the models, now that you release the constraint on layout, obviously the models over time have also become just better at instruction following as well. And so you can actually just define, hey, here are my primitives, here are my design guidelines. Don’t do this, do this.

And of course it makes mistakes at times, but I think just the quality of the model has gotten so strong that those are much less likely to happen now.”

Stein’s explanation makes clear that AI Mode is encoded with everything learned from Google’s classic search systems rather than a rebuild from scratch or a break from them. The risk of hallucinations is managed by grounding AI answers in the same relevance, trust, and usefulness signals that have underpin classic search for decades. Those signals continue to determine which sources are considered reliable and which information users have historically found valuable. Accuracy in AI search follows from that continuity, with model reasoning guided by longstanding search quality signals rather than operating independently of them.

How Google Evaluates Helpfulness In AI Mode

The next question is about the quality signals that Google uses within AI Mode. Robby Stein’s answer explains that the way AI Mode determines quality is very much the same as with classic search.

The interviewer asked:

“And Robbie, as search is evolving, as the results are changing and really, again, becoming dynamic, what signals are you looking at to know that the user is not only getting what they want, but that is the best experience possible for their search?”

Stein answered:

“Yeah, there’s a whole battery of things. I mean, we look at, like we really study helpfulness and if people find information helpful.

And you do that through evaluating the content kind of offline with real people. You do that online by looking at the actual responses themselves.

And are people giving us thumbs up and thumbs downs?

Are they appreciating the information that’s coming?

And then you kind of like, you know, are they using it more? Are they coming back? Are they voting with their feet because it’s valuable to you.

And so I think you kind of triangulate, any one of those things can lead you astray.

There’s lots of ways that, interestingly, in many products, if the product’s not working, you may also cause you to use it more.

In search, it’s an interesting thing.

We have a very specific metric that manages people trying to use it again and again for the same thing.

We know that’s a bad thing because it means that they can’t find it.

You got to be really careful.

I think that’s how we’re building on what we’ve learned in search, that we really feel good that the things that we’re shipping are being found useful by people.”

Stein’s answer shows that AI Mode evaluates success using the same core signals used for search quality, even as the interface becomes more dynamic. Usefulness is not inferred from a single engagement signal but from a combination of human evaluation, explicit feedback, and behavioral patterns over time.

Importantly, Stein notes that just because people use it a lot, presumably in a single session, that the increased usage alone is not treated as success, since repeated attempts to answer the same query indicate failure rather than satisfaction. The takeaway is that AI Mode’s success is judged by whether users are satisfied, and that it uses quality signals designed to detect friction and confusion as much as positive engagement. This carries over continuity from classic search rather than redefining what usefulness means.

Five Quality Signals For AI Search

Lastly, Stein answers a question about the ranking of AI generated content and if SEO best practices still help for ranking in AI. Stein’s answer includes five factors that are used for determining if a website meets their quality and helpfulness standards.

Stein answered:

“The core mechanic is the model takes your question and reasons about it, tries to understand what you’re trying to get out of this.

It then generates a fan-out of potentially dozens of queries that are being Googled under the hood. That’s approximating what information people have found helpful for those questions.

There’s a very strong association to the quality work we’ve done over 25 years.

Is this piece of content about this topic?

Has someone found it helpful for the given question?

That allows us to surface a broader diversity of content than traditional Search, because it’s doing research for you under the hood.

The short of it is the same things apply.

  1. Is your content directly answering the user’s question?
  2. Is it high quality?
  3. Does it load quickly?
  4. Is it original?
  5. Does it cite sources?

If people click on it, value it, and come back to it, that content will rank for a given question and it will rank in the AI world as well.”

Watch the interview starting about the one hour and twenty three minute mark:

Ecommerce MGMT 0 Comments
Dec 18 2025
Let’s Be Honest About The Ranking Power Of Links via @sejournal, @martinibuster

What link building should be trying to accomplish, in my opinion, is proving that a site is trustworthy and making sure the machine understands what topic your web pages fit into. The way to communicate trustworthiness is to be careful about what sites you obtain links from and to be super careful about what sites your site links out to.

Context Of Links Matter

Maybe it doesn’t have to be said but I’ll say it: It’s important now more than ever that the page your link is on has relevant content on it and that the context for your link is an exact match for the page that’s being linked to.

Outgoing Links Can Signal A Site Is Poisoned

Also make sure that the outgoing links are to legitimate sites, not to sites that are low quality or in problematic neighborhoods. If those kinds of links are anywhere on the site it’s best to consider the entire site poisoned and ignore it.

The reason I say to consider the site poisoned is the link distance ranking algorithm concept where inbound links tell a story about how trustworthy a site is. Low quality outbound links are a signal that something’s wrong with the site. It’s possible that a site like that will have its ability to pass PageRank removed.

Reduced Link Graph

This is how the Reduced Link Graph works, where the spammy sites are kicked out of the link graph and only the legit sites are kept for ranking purposes and link propagation. The link graph can be thought of as a map of the internet with websites connected to each other by links. When you kick out the spammy sites that’s called the reduced link graph.

Search engines are at a point where they can rank websites based on the content alone. Links still matter but the content itself is now the highest level ranking factor. I suspect that in general the link signal isn’t very healthy right now. Less people are blogging across all topics. Some topics have a healthy blogging ecosystem but in general there aren’t professors blogging about technology in the classroom and there aren’t HR executives sharing workplace insights and so on like there used to be ten or fifteen years ago.

Links for Inclusion

I’m of the opinion that links increasingly are useful for determining if a site is legit, high quality, and trustworthy, deeming it worthy for consideration in the search results. In order to stay in the SERPs it’s important to think about the outbound links on your site and the sites you obtain links from. Think in terms of reduced link graphs, with spammy sites stuck on the outside within their own spammy cliques and the non-spam on the inside within the trusted Reduced Link Graph.

In my opinion, you must be in the trusted Reduced Link Graph in order to stay in play.

Is Link Building Over?

Link building is definitely not over. There’s still important. What needs to change is how links are acquired. The age of blasting out emails at scale are over. There aren’t enough legitimate websites to make that worthwhile. It’s better to be selective and targeted about which sites you get a (free) link from.

Something else that’s becoming increasingly important is citations, other sites talking about your site. An interesting thing right now is that sponsored articles, sometimes known as native advertising, will get cited in AI search engines, including Google AI Overviews and AI Mode. This is a great way to get a citation in a way that will not hurt your rankings as long as the sponsored article is clearly labeled as sponsored and the outbound links are nofollowed.

Takeaways

  • Links As Trust And Context Signals, Not Drivers Of Ranking
    Links increasingly function to confirm that a site is legitimate and topically aligned, rather than to directly push rankings through volume or anchor text manipulation as in the old days.
  • The Reduced Link Graph Matters
    Search engines filter out spammy or low-quality sites, leaving a smaller trusted network where links and associations still count. Being outside this trusted graph puts sites at risk of exclusion.
  • Content Matters, Links Qualify
    Search engines can rank many pages based on content alone, but links can still act as a gatekeeper for credibility and inclusion, especially for competitive topics.
  • Outbound Links Are A Risk Signal
    Linking out to low-quality or problematic sites can damage a site’s perceived trustworthiness and its ability to pass value.
  • Traditional Link Building Is Obsolete
    Scaled outreach, anchor text strategies, and chasing volume are ineffective in an AI-driven search environment.
  • Citations Are Rising In Importance
    Mentions and discussions of a website can cause a site to rank better in AI search engines
  • Sponsored Articles
    Sponsored articles that are properly labeled as sponsored content and containing nofollowed links are increasingly surfaced in AI search features and contribute to visibility.

Link building is still relevant, but not in the way it used to be. Its function now is likely more about establishing whether a site is legitimate and clearly associated with a real topic area, not to push rankings through volume, anchors, or scale. Focusing on clean outbound links, selective relationships with trusted sites, and credible citations keeps a site inside the trusted reduced link graph, which is the condition that allows strong content to compete and appear in both traditional search results and AI-driven search surfaces.

Featured Image by Shutterstock/AYO Production

Ecommerce MGMT 0 Comments
Dec 18 2025
Google Says What To Tell Clients Who Want SEO For AI via @sejournal, @martinibuster

Google’s Danny Sullivan offered advice to SEOs who have clients asking for updates on what they’re going to do for AI SEO. He acknowledged it’s easier to give the advice than it is to have to actually tell clients, but he also said that advancements in content management systems drive technical SEO into the background, enabling SEOs and publishers to focus on the content.

What To Tell Clients

Danny Sullivan acknowledged that SEOs are in a tough spot with clients. He didn’t suggest specifics for how to rank better in AI search (although later in the podcast he did offer suggestions for what to do to rank better in AI search).

But he did offer suggestions for what to tell clients.

Danny explained:

“And the other thing is, and I’ve seen a number of people remark on this, is this concern that, well, I’ve been doing SEO, but now I’m getting clients or people saying to me, but I need the new stuff. I need the new stuff. And I can’t just tell them it’s the same old stuff.

So I don’t know if you feel like you need to dress it up a bit more, but I think the way you dress it up is to say, These are continuing to be the things that are going to make you successful in the long-term. I get you want the fancy new type of thing, but the history is that the fancy new type of thing doesn’t always stick around if we go off and do these particular types of things…

I’m keeping an eye on it, but right now, the best advice I can tell you when it comes to how we’re going to be successful with our AEO is that we continue on doing the stuff that we’ve been doing because that is what it’s built on.

Which is easy for me to say ’cause I don’t got someone banging on the door to say, Well, actually we do. And so we are doing that.

So that’s why, as part of the podcast, it’s just to kind of reassure that, look, just because the formats are changing didn’t mean you have to change everything that you had to do and that everything you had to shift around.”

Downside Of Prioritizing AEO/GEO For AI Search Visibility

There are many in the SEO community who are suggesting fairly spammy things to do to rank better in AI chatbots like ChatGPT, like creating listicles that recommend themselves as best whatever. Others are doing things like tweaking on keyword phrases, the kind of thing SEOs stopped doing by 2005 or 2006.

The problem with making dramatic changes to content in order to rank better in chatbots is that ChatGPT, Perplexity, and Anthropic Claude’s search traffic share is a fraction of a percent for each of them, with Claude close zero and ChatGPT estimated to be 0.2% – 0.5%.

So it absolutely makes zero sense to prioritize AEO/GEO over Google and Bing search at this point because the return on the investment is close to zero. It’s a different story when it comes to Google AI Overviews and AI Mode, but the underlying ranking systems for both AI interfaces remain Google’s classic search.

Danny shared that focusing on things that are specific to AI risks complicating what should be simple.

Google’s Danny Sullivan shared:

“And in fact, that the more that you dramatically shift things around, and start doing something completely different, or the more that you start thinking I need to do two different things, the more that you may be making things far more complicated, not necessarily successful in the long term as you think they are.”

Technical SEO Is Needed Less?

John Mueller followed up by mentioning that the advanced state of content management systems today means that SEOs and publishers no longer have to spend as much time on technical SEO issues because most CMS’s have the basics of SEO handled virtually out of the box. Danny Sullivan said that this frees up SEOs and creators to focus on their content, which he insisted will be helpful for ranking in AI search surfaces.

John Mueller commented:

“I think that makes a lot of sense. I think one of the things that perhaps throws SEOs off a little bit is that in the early days, there was a lot of almost like a technical transition where people initially had to do a lot of technical specific things to make their site even kind of accessible in search. And at some point nowadays, I think if you’re using a popular CMS like WordPress or Wix or any of them, basically you don’t have to worry about any of those technical details.

So it’s almost like that technical side of things is a lot less in the foreground now, and you can really focus on the content, and that’s really what users are looking for. So it’s like that, almost like a transition from technical to content side with regards to SEO.”

This echoed a previous statement from earlier in the podcast where Danny remarked on how some people have begun worrying less about SEO and focusing on content.

Danny said:

“But we really just want you to focus on your content and not really worry about this. If your content is on the web and generally accessible as most people’s content is, that’s it.

I’ve actually been heartened that I’ve seen a number of people saying things like: I don’t even want to think about this SEO stuff anymore. I’m just getting back into the joy of writing blogs.

I’m like, yes, great. That’s what we want you to do. That’s where we think you’re going to find your most success.”

Listen to Danny Sullivan’s remarks at about the 8 minute mark:

Featured Image by Shutterstock/Just dance

Ecommerce MGMT 0 Comments
Dec 18 2025
Google Explains How To Rank In AI Search via @sejournal, @martinibuster

Google’s John Mueller and Danny Sullivan discussed their thoughts on AI search and what SEOs and creators should be doing to make sure their content is surfaced. Danny showed some concern for folks who were relying on commodity content that is widely available.

What Creators Should Focus On For AI

John Mueller asked Danny Sullivan what publishers should be focusing on right now that’s specific to AI. Danny answered by explaining what kind of content you should not focus on and what kind of content creators should be focusing on.

He explained that the kind of content that creators should not focus on is commodity content. Commodity content is web content that consists of information that’s widely available and offers no unique value, no perspective, and requires no expertise. It is the kind of content that’s virtually interchangeable with any other site’s content because they are all essentially generic.

While Danny Sullivan did not mention recipe sites, his discussion about commodity content immediately brought recipe sites to mind because those kinds of sites seemingly go out of their way to present themselves as generically as possible, from the way the sites look, the “I’m just a mom of two kids” bio, and the recipes they provide. In my opinion, what Danny Sullivan said should make creators consider what they bring to the web that makes them notable.

To explain what he meant by commodity content, Danny used the example of publishers who used to optimize a web page for the time that the Super Bowl game began. His description of the long preamble they wrote before giving the generic answer of what time the Super Bowl starts reminded me again of recipe sites.

At about the twelve minute mark John Mueller asked Danny:

“So what would you say web creators should focus on nowadays with all of the AI?”

Danny answered:

“A key thing is to really focus on is the original aspect. Not a new thing.

These are not new things beyond search, but if you’re really trying to reframe your mind about what’s important, I think that on one hand, there’s a lot of content that is just kind of commodity content, factual information, and I think that the… LLM, AI systems are doing a good job of presenting that sort of stuff.

And it’s not originating from any type of thing.

So the classic example, as you know, will make people laugh, …but every year we have this little American football thing called the Super Bowl, which is our big event.

…But no one ever can seem to remember what time it’s on.

…Multiple places would then all write their “what time does the Super Bowl start in 2011?” post. And then they would write these giant long things.

…So, you know, and then at some point, we could see enough information and we have data feeds and everything else that we just kind of said, you do a search and …the Super Bowl is going to be at 3:30.

…I think the vast majority of people say, that’s a good thing. Thank you for just telling me the time of the Super Bowl.

It wasn’t super original information.”

Commodity Content Is Not Your Strength

Next Danny considered some of the content people are publishing today, encouraging them to think  about the generic nature of their content and to give some thought to how they can share something more original and unique.

Danny continued his answer:

“I think that is a thing people need to understand, is that more of this sort of commodity stuff, it isn’t going to necessarily be your strength.

And I do worry that some people, even with traditional SEO, focus on it too much.

There are a number of sites I know from the research and things that I’ve done that get a huge amount of traffic for the answer to various popular online word-solving games.

It’s just every day I’m going to give you the answer to it. …and that is great. Until the system shifts or whatever, and it’s common enough, or we’re pulling it from a feed or whatever, and now it’s like, here’s the answer.”

Bring Your Expertise To AI

Danny next suggested that people who are concerned about showing up in AI should start exploring how to express their authentic experience or expertise. He said this advice is not just for text content but also to video and podcast content as well.

He continued:

“Your original voice is that thing that only you can provide. It’s your particular take.

And so that’s what we think was our number one thing when we’re telling people is like, this is what we think your strength is going to be.

As we go into this new world, is already what you should be doing, but this is what your strength that you should be doing is focus on that original content.

I think related to that is this idea that people are also seeking original content that’s, …authentic to them, which typically means it’s a video, it’s a podcast…

…And you’ve seen that in the search we’ve already done, where we brought in more social, more experiential content.  Not to take away from the expert takes, it’s just that people want that.

Sometimes you’re just wanting to know someone’s firsthand experience alongside some expert take on it as well.

But if you are providing those expert takes, you’re doing reviews or whatever, and you’ve done that in the written form, you still have the opportunity to be doing those in videos and podcasts and so on.  Those are other opportunities.

So those are things that, again, it’s not unique to the AI formats, but they just may be, as you’re thinking about, how do I reevaluate what I’m doing overall in this era, that these are things you may want to be considering with it from there.”

John Mueller agreed that it makes sense to bring your unique voice to content in order to make it stand out. Danny’s point treats visibility in AI driven search as a matter of differentiation rather than optimization. The emphasis is not on adapting content to a new format, but on creating a recognizable voice and perspective with which to stand out.  Given that AI Search is still classic search under the hood, it makes sense to stand out from competitors with unique content that people will recognize and recommend.

Listen to the passage at around the twelve minute mark:

Featured Image by Shutterstock/Asier Romero

Ecommerce MGMT 0 Comments
Dec 17 2025
Google’s AI Mode Personal Context Features “Still To Come” via @sejournal, @MattGSouthern

Seven months after Google teased “personal context” for AI Mode at Google I/O, Nick Fox, Google’s SVP of Knowledge and Information, says the feature still is not ready for a public rollout.

In an interview with the AI Inside podcast, Fox framed the delay as a product and permissions issue rather than a model-capability issue. As he put it: “It’s still to come.”

What Google Promised At I/O

At Google I/O, Google said AI Mode would “soon” incorporate a user’s past searches to improve responses. It also said you would be able to opt in to connect other Google apps, starting with Gmail, with controls to manage those connections.

The idea was that you wouldn’t need to restate context in every query if you wanted Google to use relevant details already sitting in your account.

On timing, Fox said some internal testing is underway, but he did not share a public rollout date:

“Some of us are testing this internally and working through it, but you know, still to come in terms of the in terms of the public roll out.”

You can hear the question and Fox’s response in the video below starting around the 37-minute mark:

AI Mode Growth Continues Without Personal Context

Even without that personalization layer, Fox pointed to rapid adoption, describing AI Mode as having “grown to 75 million daily active users worldwide.”

The bigger change may be in how people phrase queries. Fox described questions that are “two to three times as long,” with more explicit first-person context.

Instead of relying on AI Mode to infer intent, people are writing the context into the prompt, Fox says:

“People are trying to put put the right context into the query”

That matters because the “personal context” feature was designed to reduce that manual effort.

Geographic Patterns In Adoption

Adoption also appears uneven by market, with the strongest traction in regions that received AI Mode first. Fox described the U.S. as the most “mature” market because the product has had more time to become part of people’s routines.

He also pointed to strong adoption in markets where the web is less developed in certain languages or regions, naming India, Brazil, and Indonesia. The argument there is that AI Mode can stitch together information across languages and borders in ways traditional search results may not have for those markets.

Younger users, he added, are adopting the experience faster across regions.

Publisher Relationship Updates

The interview also included updates tied to how AI Mode connects people back to publisher content.

Preferred Sources is one of them. The feature lets you choose specific publications you want to see more prominently in Google’s Top Stories unit, and Google describes it as available worldwide in English.

Fox also described ongoing work on links in AI experiences, including increasing the number of links shown and adding more context around them:

“We’re actually improving the links within our within our AI experience, increasing the number of them…”

On the commercial side, he noted Google has partnerships with “over 3,000 organizations” across “50 plus countries.”

Technical Updates

Fox talked through product and infrastructure changes now powering AI Mode and related experiences.

One was shipping Gemini 3 Pro in Search on day one, which he described as the first time Google has shipped a frontier model” in Search on launch day.

He also described generative layouts,” where the model can generate UI code on the fly for certain queries.

To keep the experience fast, he emphasized model routing, where simpler queries go to smaller, faster models and heavier work is reserved for more complex prompts.

Why This Matters

A version of AI Mode that personalizes answers using opt-in Gmail context is still not available and doesn’t have a public timeline.

In the meantime, people appear to be compensating by typing more context into their queries. If that becomes the norm, it may push publishers toward satisfying longer, more situation-specific questions.

Looking Ahead

While AI Mode is still in its early stages, the 75 million daily active users figure suggests it’s large enough to monitor for visibility.

Featured Image: Jackpress/Shutterstock

Ecommerce MGMT 0 Comments
Dec 17 2025
Google Gemini 3 Flash Becomes Default In Gemini App & AI Mode via @sejournal, @MattGSouthern

Google released Gemini 3 Flash, expanding its Gemini 3 model family with a faster model that’s now the default in the Gemini app.

Gemini 3 Flash is also rolling out globally as the default model for AI Mode in Search.

The release builds on Google’s recent Gemini 3 rollout, which introduced Gemini 3 Pro in preview and also announced Gemini 3 Deep Think as an enhanced reasoning mode.

What’s New

Gemini 3 Flash replaces Gemini 2.5 Flash as the default model in the Gemini app globally, which means free users get the Gemini 3 experience by default.

In Search, Gemini 3 Flash is rolling out globally as AI Mode’s default model starting today.

For developers, Gemini 3 Flash is available in preview via the Gemini API, including access through Google AI Studio, Google Antigravity, Vertex AI, Gemini Enterprise, plus tools such as Gemini CLI and Android Studio.

Pricing

Gemini 3 Flash pricing is listed at $0.50 per million input tokens and $3.00 per million output tokens on Google’s Gemini API pricing documentation.

On the same pricing page, Gemini 2.5 Flash is listed at $0.30 per million input tokens and $2.50 per million output tokens.

Google says Gemini 3 Flash uses 30% fewer tokens on average than Gemini 2.5 Pro for typical tasks, and citing third-party benchmarking for a “3x faster” comparison versus 2.5 Pro.

Why This Matters

The default language model in the Gemini app has changed, and users have access at no extra cost.

If you build on Gemini, Gemini 3 Flash offers a new option for high-volume workflows, priced well below Pro-tier rates.

Looking Ahead

Gemini 3 Flash is rolling out now. In Search, Gemini 3 Pro is also available in the U.S. via the AI Mode model menu.

Ecommerce MGMT 0 Comments
Dec 17 2025
Google Updates JavaScript SEO Docs With Canonical Advice via @sejournal, @MattGSouthern

Google updated its JavaScript SEO documentation with new guidance on handling canonical URLs for JavaScript-rendered sites.

The documentation update also adds corresponding guidance to Google’s best practices for consolidating duplicate URLs.

What’s New

The updated documentation focuses on a timing issue specific to JavaScript sites: canonicalization can happen twice during Google’s processing.

Google evaluates canonical signals once when it first crawls the raw HTML, then again after rendering the JavaScript. If your raw HTML contains one canonical URL and your JavaScript sets a different one, Google may receive conflicting signals.

The documentation notes that injecting canonical tags via JavaScript is supported but not recommended. When JavaScript sets a canonical URL, Google can pick it up during rendering, but incorrect implementations can cause issues.

Multiple canonical tags, or changes to an existing canonical tag during rendering, can lead to unexpected indexing results.

Best Practices

Google recommends two best practices depending on your site’s architecture.

The preferred method is setting the canonical URL in the raw HTML response to match the URL your JavaScript will ultimately render. This gives Google consistent signals before and after rendering.

If JavaScript must set a different canonical URL, Google recommends leaving the canonical tag out of the initial HTML. This can help avoid conflicting signals between the crawl and render phases.

The documentation also reminds developers to ensure only one canonical tag exists on any given page after rendering.

Why This Matters

This guidance addresses a subtle detail that can be easy to miss when managing JavaScript-rendered sites.

The gap between when Google crawls your raw HTML and when it renders your JavaScript creates an opportunity for canonical signals to diverge.

If you use frameworks like React, Vue, or Angular that handle routing and page structure client-side, it’s worth checking how your canonical tags are implemented. Look at whether your server response includes a canonical tag and whether your JavaScript modifies or duplicates it.

In many cases, the fix is to coordinate your server-side and client-side canonical implementations so they send the same signal at both stages of Google’s processing.

Looking Ahead

This documentation update clarifies behavior that may not have been obvious before. It doesn’t change how Google processes canonical tags.

If you’re seeing unexpected canonical selection in Search Console’s Page indexing reporting, check for mismatches between your raw HTML and rendered canonical tags. The URL Inspection tool shows both the raw and rendered HTML, which makes it possible to compare canonical implementations across both phases.

Featured Image: Alicia97/Shutterstock

Ecommerce MGMT 0 Comments