SEO + AI: Best Practices for 2026

00:01:19.000 –> 00:01:31.000
Hey everyone, thanks for joining us today. This is our last webinar of 2025. It has been a year, to be sure. Um, but thank you all for joining us, for hanging in with us.

00:01:31.000 –> 00:01:43.000
Um, I have been excited about this webinar. for, like, 6 months, because, um, I first got to see a version of that, an earlier version of this session at the annual meeting in Atlanta.

00:01:43.000 –> 00:01:51.000
And, um… The whole idea of what AI is doing to search engine optimization, what it will do.

00:01:51.000 –> 00:02:03.000
That what it does to search engine marketing is one of the most interesting things I think we could be thinking about relative to generative and agentic AI right now. So, um, I’m just psyched that.

00:02:03.000 –> 00:02:07.000
Uh, Richard and Emily agreed to do this session, and we were able to get it on the books here.

00:02:07.000 –> 00:02:16.000
Still for 2025. So, um, I will let them do a little bit of intro to explain sort of the differences between SEO and SEM, but just quickly, Richard.

00:02:16.000 –> 00:02:29.000
trailer that handles SEO at PBS, Emily Clark handles SEM. And, um… I guess the last thing I want to do is, as we always do, thank Nita for their support of Public Media Innovators.

00:02:29.000 –> 00:02:50.000
Uh, and then also, thank you to David and… Uh, Amber, for always having my back with these things, and some of the times they produce them, some of the times I produce them, but it’s always a team effort. So, with that, uh, I think I’ll turn it over to, uh, Richard. Everyone, please use the chat to ask questions, and I think, as we said earlier.

00:02:50.000 –> 00:03:06.000
Even if the questions aren’t, like, specifically AI, like, you know, you’ve got two, uh… search engine pros, uh, here available to you, so, you know, it doesn’t have to be limited to AI. So anyway, um, I think we said, Richard, you’d kick it off?

00:03:06.000 –> 00:03:15.000
Yeah, sounds good. Uh, thanks so much, Chad, and uh, thank you also to David and Amber for inviting us. Glad to have this opportunity to speak with you all.

00:03:15.000 –> 00:03:30.000
Uh, like Chad said, my name is Richard Traylor, I’m the SEO Manager at PBS. Uh, I’ve been with PBS for… Uh, about a little over 4 years now, I’ve been doing SEO full-time for about 9 years, coming up on 9 years.

00:03:30.000 –> 00:03:38.000
Uh, and specifically within PBS. my role is, uh, I’m a subject matter expert on SEO.

00:03:38.000 –> 00:03:45.000
And, uh, working to improve the organic visibility and discoverability of public media content.

00:03:45.000 –> 00:03:50.000
on Google, from an organic lens, so that non-paid lens.

00:03:50.000 –> 00:04:00.000
Uh, but also, you know, there have been a lot of changes with the integration of AI and other features into search recently, so that’s been a big focus of ours as well, is, you know.

00:04:00.000 –> 00:04:03.000
How is AI going to disrupt the search landscape?

00:04:03.000 –> 00:04:10.000
Uh, and then determining what the impact on public media might be, and developing strategies, and then sharing that within the system.

00:04:10.000 –> 00:04:16.000
Uh, and I’ve been joined by my wonderful colleague, Emily Clark. We come together to create a very.

00:04:16.000 –> 00:04:24.000
great and cohesive, collaborative, uh, search engine marketing strategy, so I will hand it over to her to introduce herself as well before we.

00:04:24.000 –> 00:04:25.000
dive into things.

00:04:25.000 –> 00:04:39.000
Thanks, Richard. Hi, everyone. As Richard said, I’m Emily. I’m a manager on the marketing strategy team. Um, I’ve also been with PBS for about 4 years, and I work on the paid marketing side, so my team tends to focus on.

00:04:39.000 –> 00:04:43.000
media buys and promotional strategies across platforms like Roku or Samsung.

00:04:43.000 –> 00:04:52.000
Hulu, um, and of course, Google. Um, so I spend a lot, a lot of time optimizing our paid search campaigns and working very closely with Richard.

00:04:52.000 –> 00:04:59.000
To make sure that our organic SEO efforts and our paid search engine marketing efforts, um.

00:04:59.000 –> 00:05:13.000
work together, um, in an efficient way. And so, um, on our side, paid search campaigns, uh, we like to try to maximize streams for our programming, just to give people a heads up. Um, so in the recent years, we’ve seen kind of continuous changes.

00:05:13.000 –> 00:05:18.000
to the paid search landscape as Google develops new AI tools to make buying ads.

00:05:18.000 –> 00:05:30.000
easier, which is helpful for us, but obviously makes them a lot of money. It has also made the marketplace quite a bit more crowded and competitive, as I’m sure all of you have seen.

00:05:30.000 –> 00:05:36.000
And they have also done this, of course, on purpose to make sure that they’re ramping up their own ad revenue.

00:05:36.000 –> 00:05:42.000
Um, so thank you all, and I look forward to, uh, talking with you about this.

00:05:42.000 –> 00:05:47.000
Awesome, thanks Emily. Uh, just to give a little bit of context about this presentation.

00:05:47.000 –> 00:05:56.000
Uh, so this is… we consider sort of a continuation of a conversation that started at annual meeting in 2025.

00:05:56.000 –> 00:06:05.000
Uh, and we want to cover some of the same ground that we covered at Annual Meeting, just because we know not everyone is able to attend, and not all of the sessions are recorded.

00:06:05.000 –> 00:06:14.000
So, if there is anyone in the chat who was able to attend our session, there may be some content that you’ve seen before, but we just want to make sure that we’re.

00:06:14.000 –> 00:06:20.000
sort of covering the basis for everyone. Uh, you’ll also see that, uh, our.

00:06:20.000 –> 00:06:25.000
Slideshow is branded with the official 2025 annual meeting branding.

00:06:25.000 –> 00:06:35.000
But just, uh, wanted to preface by saying we have updated the content of the spreadsheet with new information since then.

00:06:35.000 –> 00:06:39.000
We just wanted to build off of what we started previously.

00:06:39.000 –> 00:06:48.000
But I will go ahead and share my screen, and we can get started.

00:06:48.000 –> 00:06:59.000
Awesome. So, we introduced ourselves. So, first, I want to start just by going through, uh, and highlighting some of the changes that have been made.

00:06:59.000 –> 00:07:04.000
Uh, and search, uh, specifically regarding AI over the last couple of years.

00:07:04.000 –> 00:07:11.000
Uh, because things have changed a lot. There have been a lot of new technologies and features that have been added, both in terms of.

00:07:11.000 –> 00:07:18.000
brand new competitors in the space, as well as AI features that are being integrated into traditional search systems.

00:07:18.000 –> 00:07:26.000
So, we just want to start by just running through a list of, like, what’s new, what should you be aware of.

00:07:26.000 –> 00:07:34.000
But it’s 2025. These are the public media innovators, so we have to start by asking ChatGPT to help us out.

00:07:34.000 –> 00:07:45.000
Oh, no, my video didn’t play. Oh, sad. There is a video that we’ll play if you do this presentation on your computer, where basically what I’m doing is I’m prompting ChatGPT.

00:07:45.000 –> 00:07:53.000
To outline, uh, new AI features. Uh, but the joke, which I sort of spoiled a little bit with my, uh.

00:07:53.000 –> 00:07:59.000
my presentation is that we’re going to ask Richard GPT for a better list.

00:07:59.000 –> 00:08:12.000
I haven’t seen this image in a couple of months now, and it never gets easier to look at. It’s always extremely uncanny valley, uh, so… Let’s be done with that, uh, and move on for now.

00:08:12.000 –> 00:08:23.000
So, one of the most prominent AI integrations that you’ve seen, I’m sure everyone has encountered them by now, are Google’s AI overviews. You can see a screenshot of one here on the left.

00:08:23.000 –> 00:08:30.000
And these are AI summaries that are generated by Google, specifically using their Gemini large language model.

00:08:30.000 –> 00:08:34.000
Uh, and these AI summaries appear at the top of search results.

00:08:34.000 –> 00:08:40.000
And as Emily will talk about later, they’re prioritized over both organic and paid results.

00:08:40.000 –> 00:08:47.000
Uh, and basically what these AI overviews do is they provide a summarized answer to the user’s search query.

00:08:47.000 –> 00:08:55.000
Uh, so you have the AI-generated summary, and then on the right-hand side, you’ll see, uh, there are links to some websites for users that want to learn more.

00:08:55.000 –> 00:09:05.000
Uh, we’ll talk a little bit more about tracking later, but I do just want to preface at the beginning that clicks on these, uh, links referenced on the right side.

00:09:05.000 –> 00:09:09.000
are considered organic search traffic when it’s logged in your Google Analytics property.

00:09:09.000 –> 00:09:14.000
And like I said, we’ll talk a little bit more about that later.

00:09:14.000 –> 00:09:20.000
But in addition to AI overviews, we also have chatbots, which many of you are also likely very familiar with.

00:09:20.000 –> 00:09:27.000
ChatGPT is the most popular by far. I think they’re up to something like 700 million weekly active users.

00:09:27.000 –> 00:09:34.000
Uh, but there are also a variety of others. Uh, you’ve got things like Google’s Gemini, DeepSeat, Grok, Claude.

00:09:34.000 –> 00:09:38.000
Probably a dozen more, uh, that I can’t even think of.

00:09:38.000 –> 00:09:44.000
Uh, so we all know how chatbots work. A user enters a prompt, and then the chatbot provides a response.

00:09:44.000 –> 00:09:50.000
This could be in the form of text or images, depending on the chatbot or the prompt that you’re asking.

00:09:50.000 –> 00:09:59.000
Uh, the chatbot may or may not include links to reference websites, it depends on the service you’re using or the level of subscription that you have.

00:09:59.000 –> 00:10:02.000
Uh, we’ll talk a little bit more about how chatbots.

00:10:02.000 –> 00:10:13.000
interact in the search landscape. Uh, later on in this presentation, but I do just want to preface that users tend to use chatbots differently than they use standard search.

00:10:13.000 –> 00:10:22.000
So what we’re seeing is that. people tend to ask more complex and longer queries to chatbots than they would to Google.

00:10:22.000 –> 00:10:27.000
So, think about, uh… With ChatGPT, you might ask.

00:10:27.000 –> 00:10:33.000
A complicated question like, where can a family of four go on vacation for.

00:10:33.000 –> 00:10:39.000
$5,000 that’s within a 5-hour flight from Washington, D.C, for example.

00:10:39.000 –> 00:10:49.000
That’s a relatively complicated query that Google just can’t really handle with traditional search, and that’s where we’re seeing that these AI chatbots tend to excel.

00:10:49.000 –> 00:10:53.000
And like I said, we’ll talk about that more later.

00:10:53.000 –> 00:10:59.000
Uh, so additionally, there’s also, uh, a new tool that has been, uh.

00:10:59.000 –> 00:11:04.000
introduced over the last year called Google’s AI Mode.

00:11:04.000 –> 00:11:11.000
I would have a video that would play here that shows how it works, but it seems like there might be an issue with, uh, with the videos.

00:11:11.000 –> 00:11:16.000
Um, so, uh, this is live, so you can test this for yourself, actually.

00:11:16.000 –> 00:11:25.000
Uh, but generally what AI mode is, is it’s a new tab within Google search engine, so think about how there’s separate tabs for Google Images or Google Shopping.

00:11:25.000 –> 00:11:36.000
mode is an additional tab similar to that. And basically what AI Mode does is you search a query, and then AI mode will comb through websites for you.

00:11:36.000 –> 00:11:42.000
And then it takes all of the information from those websites, and then it summarizes it using AI, and it spits out.

00:11:42.000 –> 00:11:54.000
An easily digestible AI summary. So, what AIMode does is it basically does the Googling for you, and then it just provides you with the raw information that you’re looking for.

00:11:54.000 –> 00:12:03.000
AI mode works very similarly to Perplexity, if there’s anyone that’s used that service before, where it basically just does the Googling for you.

00:12:03.000 –> 00:12:13.000
AI mode is relatively new. It was released publicly earlier this year. Actually, funnily enough, it was released about 2 hours before Emily and I presented at Annual Meeting.

00:12:13.000 –> 00:12:24.000
Uh, which was great timing. Uh… While this is still a relatively new tool, it doesn’t seem that it’s really resonated with audiences yet.

00:12:24.000 –> 00:12:31.000
Uh, so, in July, Google stated that AI Mode had about 100 million monthly active users.

00:12:31.000 –> 00:12:39.000
Which is a lot, 100 million is a lot, but when you consider the scale that a lot of these tech companies are operating at, that’s very small.

00:12:39.000 –> 00:12:45.000
Google handles about 14 to 15 billion queries a day.

00:12:45.000 –> 00:12:49.000
Uh, for context. So, it’s still sort of finding its pudding.

00:12:49.000 –> 00:12:55.000
Uh, additionally, similar web data shows that about 50% of the users for AI mode.

00:12:55.000 –> 00:13:02.000
used it once, and then didn’t return. So Google is struggling to get users to use the tool and know about the tool.

00:13:02.000 –> 00:13:06.000
And then also to retain them, uh, and keep them using the tool.

00:13:06.000 –> 00:13:14.000
But despite these initial statistics, uh, responses from Google’s leadership indicates that AI mode is the future.

00:13:14.000 –> 00:13:17.000
of search. This is the direction that Google’s leadership.

00:13:17.000 –> 00:13:20.000
wants to be heading in. So this is something that we need to be aware of.

00:13:20.000 –> 00:13:29.000
Uh, and need to be… cognizant about how this interfaces with our content moving forward.

00:13:29.000 –> 00:13:33.000
But we’ll spend most of the time in this presentation, at least in my portion.

00:13:33.000 –> 00:13:39.000
Uh, discussing AI overviews, as they have the most immediate impact on search visibility and search engagement.

00:13:39.000 –> 00:13:47.000
So, third-party studies of AI overviews have shown that they decrease click-through rates for both paid and organic search.

00:13:47.000 –> 00:13:53.000
So, there are just fewer users that are clicking into websites when AI overviews appear.

00:13:53.000 –> 00:14:06.000
So, in Google, uh, in search, there’s a concept called zero-click searches, which refer to users, uh, searches on Google that don’t result in any clicks to a website, and that’s either paid or organic.

00:14:06.000 –> 00:14:10.000
So, a user makes a search query, and then they don’t click.

00:14:10.000 –> 00:14:19.000
Uh, into anything. Uh, so in zero-click searches. either a user has their question answered by Google directly, and they don’t need to click through.

00:14:19.000 –> 00:14:25.000
Or they have to refine their search results to see more relevant, uh, more relevant content.

00:14:25.000 –> 00:14:34.000
Because AI overviews provide direct answers to users, we’re likely to see a decline in search traffic as zero-click searches increase.

00:14:34.000 –> 00:14:41.000
Because Google is… providing, uh… That immediate summarized answer.

00:14:41.000 –> 00:14:49.000
There’s just a reduced need to click through to learn more, because Google’s just showing you directly what you’re looking for in a lot of circumstances.

00:14:49.000 –> 00:14:58.000
We do want to note that AI overviews don’t appear everywhere, but they do appear more frequently in search than they did even just a few months ago.

00:14:58.000 –> 00:15:03.000
Uh, in addition to the increased prevalence of AI overviews, they’re also getting more sophisticated.

00:15:03.000 –> 00:15:10.000
Uh, so when we presented an annual meeting, it was primarily just text content that was appearing in AI overviews.

00:15:10.000 –> 00:15:16.000
But we’ve recently seen that AI overviews are now embedding video content directly within the summary.

00:15:16.000 –> 00:15:25.000
So you may have a text description, and then there may be an embedded video that appears, but that’s going to depend on the search query.

00:15:25.000 –> 00:15:31.000
Uh, I will add to that, because public media, we operate very much in the video space.

00:15:31.000 –> 00:15:35.000
AI overviews are currently only embedding content from YouTube.

00:15:35.000 –> 00:15:47.000
from what we’ve seen so far. We have done investigations into this, and there appears to be no technical blockers preventing Google from embedding the PBS video player or other video players.

00:15:47.000 –> 00:16:01.000
So, it appears at the moment that they’re just preferring resources from YouTube, but… Uh, we may need to hear if there’s additional, um, details specifically from Google about that, like maybe structured data that we can apply that will.

00:16:01.000 –> 00:16:13.000
Uh, make our content eligible. Uh, so this clearly has very major implications for the public media system, so it’s something that we’re continuing to monitor at PBS, and we can provide.

00:16:13.000 –> 00:16:18.000
updates and guidance on that in the future as we learn more.

00:16:18.000 –> 00:16:25.000
Uh, so all of the data that I’ve shared so far has been from third-party organizations, and while it’s good for understanding the broader industry.

00:16:25.000 –> 00:16:33.000
They may not necessarily match the impact on public media, so I want to share the results of an internal analysis that we did.

00:16:33.000 –> 00:16:46.000
Caveat is that this analysis was from this spring, so… Uh, it’s a little bit older at this point, but we think the, uh, the results of this analysis are still relevant, so we still wanted to share.

00:16:46.000 –> 00:16:55.000
Uh, and this analysis was, um… intended to determine how PBS.org was being referenced in AI overviews.

00:16:55.000 –> 00:17:02.000
And then, what the implications of that mean for how AI overviews might impact engagement and search.

00:17:02.000 –> 00:17:10.000
So, at the time of this analysis, PBS.org was referenced at about 36,000 AI overviews.

00:17:10.000 –> 00:17:18.000
Uh, this was roughly comparable to the number of references to NPR.org and CNN.com, so very strong.

00:17:18.000 –> 00:17:29.000
And it was much, much higher than the number of references we were seeing for traditional streaming sites like Hulu or Disney+, or Peacock.

00:17:29.000 –> 00:17:35.000
The search queries that triggered an AI overview that referenced PBS.org tended to be informational.

00:17:35.000 –> 00:17:40.000
Uh, such as the search query, what is the biggest tsunami ever recorded?

00:17:40.000 –> 00:17:50.000
Uh, more importantly, though, about 95% of the AI overviews that link to PBS.org linked to an article or another text-based page.

00:17:50.000 –> 00:17:55.000
So, these are pages that are primarily on what we call our producer sites.

00:17:55.000 –> 00:17:58.000
which are sites that are produced and managed by presenting stations.

00:17:58.000 –> 00:18:05.000
So, think of the sites for PBS News, or Masterpiece, or Nova, or Antiques Roadshow.

00:18:05.000 –> 00:18:11.000
Uh, these producer sites are excellent at generating organic search traffic and engaging both new and existing audiences.

00:18:11.000 –> 00:18:17.000
But they’re made up of a lot of articles, which don’t tend to convert strongly to, uh.

00:18:17.000 –> 00:18:28.000
video streams or passport signups. So our ultimate takeaway from this analysis was that while organic search traffic may decline as a result of AI overviews.

00:18:28.000 –> 00:18:35.000
This isn’t likely to impact streaming or passport signups or other core KPIs.

00:18:35.000 –> 00:18:44.000
As AI overviews are mostly impacting the informational supporting article content at the moment.

00:18:44.000 –> 00:18:52.000
Uh, however, I will say that this analysis was exclusive to the PBS.org domain, so we can’t promise that this is going to be the same for all stations.

00:18:52.000 –> 00:19:00.000
And AI overviews continue to evolve. This didn’t include information about embedded video, so the situation may have changed.

00:19:00.000 –> 00:19:05.000
Uh, but it’s at least a better indicator of what the impact might be than some of the third-party.

00:19:05.000 –> 00:19:09.000
Uh, studies that we’ve, uh. We’ve gone through.

00:19:09.000 –> 00:19:16.000
But with that, before I transition to handing it over to Emily, I’ll pause there in case there were any.

00:19:16.000 –> 00:19:19.000
urgent questions in the chat that we should address.

00:19:19.000 –> 00:19:32.000
Yes, uh, Richard, it looks like we have a few, um, from Priyanka, it says, in relation to this slide, AI overviews may not have an immediate impact on streaming, but is it fair to say that as organic search traffic declines.

00:19:32.000 –> 00:19:40.000
The visibility and awareness of our sites or brands will have an impact on streaming and other KPIs.

00:19:40.000 –> 00:19:52.000
That’s a really great point. Uh, yes, probably. Uh… It’s fair to say that, as organic search traffic declines, the visibility and awareness.

00:19:52.000 –> 00:19:59.000
Probably. Uh… That is… it just gets a little bit hard to track.

00:19:59.000 –> 00:20:08.000
At that point, uh… What we tend to see is… Um, then you’re sort of talking about, like.

00:20:08.000 –> 00:20:17.000
the funnel as a whole. Uh, as opposed to just sort of, like, that immediate session that follows, uh, an organic search click.

00:20:17.000 –> 00:20:25.000
Uh, and that’s just harder to track. I think that’s something that we’re still sort of trying to understand is, you know.

00:20:25.000 –> 00:20:32.000
What’s the brand awareness and the brand, like, piece of this, because there’s.

00:20:32.000 –> 00:20:47.000
There’s certainly, uh… something to be said about that, uh… And it’s still just a little unclear, as we’ll talk about later, there’s just not a lot of transparency from Google or a lot of these AI companies with how they’re.

00:20:47.000 –> 00:20:55.000
Uh, displaying our content in these AI summaries. Uh, especially outside of the, like, traditional Google search ecosystem.

00:20:55.000 –> 00:21:00.000
Uh, so I think this is something we’ll be focusing a lot more on as, like.

00:21:00.000 –> 00:21:05.000
tools are created that are better at tracking this.

00:21:05.000 –> 00:21:19.000
I’ll also say, uh, Richard, just from… looking at, like, what other people have been doing and such to try to rank for AI overviews, I think now it’s really accepted that appearing in AI overviews is a great brand awareness tool.

00:21:19.000 –> 00:21:30.000
So, while it is disappointing that, you know, site traffic is lower, um, people are saying that by appearing for AI overviews, you are establishing your brand, so it continues to be important.

00:21:30.000 –> 00:21:51.000
To, um, try to format your website and articles to appear in those AI overviews, which Richard will get into later on, um, about how SEO and AI optimizations often coincide. But although it may seem like we’re losing traffic, I think PBS kind of has, like, a very trusted brand that people are.

00:21:51.000 –> 00:21:58.000
looking for that content, it’s helpful when we do appear in those AI overviews, because people tend to remember us better.

00:21:58.000 –> 00:22:05.000
Yeah, I will say, Mark, uh, made a comment that is a much more succinct and better.

00:22:05.000 –> 00:22:14.000
explanation to Priyanka’s question that I set… that I had. Uh, so, thank you, Mark, for covering for me. That was an excellent summary.

00:22:14.000 –> 00:22:18.000
Uh, and then there were a couple other questions. Sorry, I just pulled up the chat.

00:22:18.000 –> 00:22:30.000
Dave asked if we can share the methods, uh… I’ll respond to you, uh, I’ll add a response while Emily is talking, uh, I’ll… share an email address where we can… we can talk more about that.

00:22:30.000 –> 00:22:36.000
Uh, and then Chad asked, how do you see where AI overviews link to your site?

00:22:36.000 –> 00:22:41.000
Currently, that’s not available in Google Analytics. Uh, we pay for a third-party tool called Ahrefs.

00:22:41.000 –> 00:22:44.000
Uh, and I’ll have a slide on this later.

00:22:44.000 –> 00:22:51.000
Um, but we use these third-party tracking tools, these third-party SEO tools. They have.

00:22:51.000 –> 00:22:57.000
They basically scrape Google search results, so they have information about where AI overviews appear.

00:22:57.000 –> 00:23:03.000
And what websites are being ranked, which is what we primarily use for these types of analyses.

00:23:03.000 –> 00:23:11.000
Exactly, and I’ll even say… I’ll talk about this a little later, but even on the paid side, we pay for ads, and supposedly they are appearing.

00:23:11.000 –> 00:23:24.000
You know, above or below or within A overviews, and Google does not tell us this information either. They have still not rolled out any transparency into where these ads are now appearing versus just plain search. It’s all kind of grouped into the same category.

00:23:24.000 –> 00:23:43.000
So they are, you know, definitely… trying to weigh the benefits of revealing how much of your traffic is going to an AI overview versus not, even on the ad side. Um, and they’re hoping, actually, that you’ll double bid on that keyword, just so that you are sure that you’re appearing above, below, within, and all of that.

00:23:43.000 –> 00:23:47.000
Um, so we’ll get into a little bit more detail on some of the AI tools that Google’s using, but.

00:23:47.000 –> 00:23:57.000
it’s obviously disappointing as well that. Even when you pay for it, uh, they’re not showing you how exactly you’re ranking.

00:23:57.000 –> 00:24:01.000
And that, uh, is a good transition, Emily, if you want to take it away.

00:24:01.000 –> 00:24:18.000
Yes, definitely. Um, so hi everyone. In this section, I’ll now be pivoting toward how AI features are changing the paid search landscape specifically. Um, we’ll dive into how AI overviews affect search ad performance, and how to create meaningful paid search ads for websites.

00:24:18.000 –> 00:24:32.000
With enhanced AI tools and Google’s ad suites. I will caveat that at PBS, our paid search goal is to increase video streams, so this strategy focuses on optimizing for streaming KPIs.

00:24:32.000 –> 00:24:46.000
Next slide, please. Great. So, uh, Google introduced AI overviews in 2024, as I’m sure you all saw, and… watched, perhaps weeping. Um, they rolled them out in beta mode to learn quickly how people interact with the new feature.

00:24:46.000 –> 00:25:01.000
Um, they have become, I would say, increasingly more accurate and helpful to users than when they originally were rolled out. Obviously, there are still some caveats to this. I believe there was a study that said, like, 30% of AI overview information is inaccurate.

00:25:01.000 –> 00:25:14.000
So, obviously, it’s becoming more useful, and users are relying on them more for information, but there is still a lot of quirks that Google needs to work out to make sure that the right information is being shown.

00:25:14.000 –> 00:25:26.000
However, with the increased usage and people relying on AI overviews more heavily, it’s no surprise that they are here to stay. Um, they have become increasingly more common, as Richard mentioned.

00:25:26.000 –> 00:25:31.000
And are an essential factor in where your paid search ads are going to appear.

00:25:31.000 –> 00:25:43.000
So, currently, ads appear with AI overviews 87% of the time. This means that almost all ads will appear simultaneously, with an AI overview either above, below, or in.

00:25:43.000 –> 00:25:53.000
in the AI overview, and again, Google will not tell you exactly where they are appearing. Um, Google says the integration of ads with AI overviews is designed to enhance.

00:25:53.000 –> 00:25:57.000
the user experience by connecting them with relevant products or services.

00:25:57.000 –> 00:26:06.000
However, this also means that users are now seeing AI overviews, um, above your ads. So while you may still be paying for that top-of-the-page placement.

00:26:06.000 –> 00:26:12.000
your ad has most likely lost a significant amount of real estate and user attention.

00:26:12.000 –> 00:26:20.000
So, AI overviews hover around 4,300 characters on average, and that means 84% of the time they cover more than half of the search result page.

00:26:20.000 –> 00:26:27.000
Um, with so much of the page swallowed by these overviews, the likelihood of your ad getting seen nevertheless clicked on.

00:26:27.000 –> 00:26:37.000
does drop significantly. In the past 6 months, on average, um, studies have shown that AI overviews significantly eat away at paid traffic.

00:26:37.000 –> 00:26:43.000
We saw in May that there was an estimated 25% CTR drop when AI summaries are shown above ads.

00:26:43.000 –> 00:26:52.000
now there is a 68% CTR drop as people become more accustomed to AI overviews, and Google has increased the amount of time when AI overviews appear.

00:26:52.000 –> 00:27:03.000
However, this is not necessarily the case for all ads, and as I go into the PBS strategy, we have actually seen an increased number of clicks and conversions rather than a decrease.

00:27:03.000 –> 00:27:13.000
because of how we’re setting up our ads, as well as just the, you know, wealth of programming that PBS has published in the past 6 months, um, with AMREV and everything.

00:27:13.000 –> 00:27:23.000
Um, so I’ll go more into that. Um, and to kind of talk about that, we have also seen a divide in AI overviews appearing for branded versus non-branded search queries.

00:27:23.000 –> 00:27:29.000
So for context, non-branded search queries are searches that do not include a specific company or product name.

00:27:29.000 –> 00:27:34.000
For PBS, these would be queries that do not include the word PBS.

00:27:34.000 –> 00:27:38.000
your local station name, or a show title, like All Creatures Great and Small.

00:27:38.000 –> 00:27:46.000
Non-branded terms tend to be generic terms, like streaming service, educational content, best period dramas, or science documentaries.

00:27:46.000 –> 00:27:50.000
So, for these non-branded queries, 25% of the time.

00:27:50.000 –> 00:27:58.000
Google shows an AI overview. Um, with that, Google has not only increased the competition for your content to be seen among other sites.

00:27:58.000 –> 00:28:10.000
But you’re now competing with Google itself. So while AI overviews may pose a new challenge for your ads and content to appear first on the page, we do have some ways to combat this.

00:28:10.000 –> 00:28:16.000
In these screenshots, you see two examples of people searching for the American Revolution.

00:28:16.000 –> 00:28:24.000
In the first screenshot, the query is simply, watch American Revolution. This search generates one of our sponsored ads right at the top of the page.

00:28:24.000 –> 00:28:30.000
We are the first result, and it is likely that a user will click on us because we are the immediate answer to their query.

00:28:30.000 –> 00:28:35.000
And the second screenshot, the query changes to when is PBS American Revolution?

00:28:35.000 –> 00:28:43.000
This query generates an AI overview. Google summarizes when American Revolution premieres on PBS, and specifics about the schedule.

00:28:43.000 –> 00:28:47.000
So, why did these two queries provide very different results?

00:28:47.000 –> 00:28:51.000
This is due to the type of question that the user has inputted.

00:28:51.000 –> 00:28:56.000
AI overviews tend to appear for informational, how-to, and exploratory searches.

00:28:56.000 –> 00:29:03.000
So basically, when a user is looking for a quick and comprehensive answer, rather than a navigational click.

00:29:03.000 –> 00:29:09.000
With this in mind, you can actually frame your paid search strategy based on when you know an A overview will appear.

00:29:09.000 –> 00:29:16.000
So why waste money bidding on a query where your station shows up halfway down the page? Why not maximize your budget for searches?

00:29:16.000 –> 00:29:25.000
where your station appears as the first result. The best time to use search ads tends to be for branded queries. So, like, in the previous example.

00:29:25.000 –> 00:29:34.000
People simply searching for a show title, your local station, or some kind of branded benefit will typically generate your search ad to appear at the top of the page.

00:29:34.000 –> 00:29:39.000
For example, bidding on keywords like PBS Big Cats 24-7 or PBS Passport.

00:29:39.000 –> 00:29:49.000
AI overviews also appear less frequently for local searches. They only appear for 8% of local queries versus 20% for non-local searches.

00:29:49.000 –> 00:29:58.000
And this is a huge advantage to stations where users might be searching for local TV or local news or information about your station or state.

00:29:58.000 –> 00:30:04.000
So, I would recommend leaning into your local queries and ranking top of the page for those.

00:30:04.000 –> 00:30:12.000
Finally, navigational queries, or where someone is trying to find where to do something, tends to bring up paid ads rather than AI overview.

00:30:12.000 –> 00:30:20.000
This is because people want a one-button click to the answer to the query, rather than reading a very long explanation on then how to find that button.

00:30:20.000 –> 00:30:30.000
So, bidding on something like download the PBS app, or donate to PBS Passport should bring up a search ad if you are bidding on those types of keywords.

00:30:30.000 –> 00:30:37.000
While it sometimes seems counterintuitive to bid on your own keywords, it’s actually important to maintain your station’s share of voice in search.

00:30:37.000 –> 00:30:44.000
Share a voice is basically the percentage of time your brand or station is appearing when people are searching for you.

00:30:44.000 –> 00:30:52.000
On a national level, we often bid on branded keywords for paid ads, um, like PBS will fall, PBS American Revolution, Ken Burns.

00:30:52.000 –> 00:31:00.000
Because we want to be the top paid and organic search result. We don’t want competitors swooping in and stealing our traffic.

00:31:00.000 –> 00:31:10.000
We don’t want our brand diluted by someone else appearing. We want to be the only place people think of when they search for our content.

00:31:10.000 –> 00:31:19.000
And now that we have laid out kind of a general strategy on how your ads can appear more frequently in results, let’s discuss how to optimize your search ads.

00:31:19.000 –> 00:31:28.000
using AI tools. Google offers three standard types of AI tools to improve your search ad performance now. Data-powered targeting.

00:31:28.000 –> 00:31:38.000
generative AI for ads, and something very, very… well, somewhat new, called AI Max Expansions that they’ve rolled out over the past few months.

00:31:38.000 –> 00:31:45.000
Um, these 3 tools will actually maximize your results while minimizing the amount of time you have to spend creating an effective ad.

00:31:45.000 –> 00:31:50.000
Data-powered targeting will allow you to show more relevant messaging to the right people.

00:31:50.000 –> 00:31:57.000
Generative AI will allow you to create optimized search keywords, copy and images from your stations, website’s link.

00:31:57.000 –> 00:32:04.000
And then AI Max expansions kind of acts as an additional optimization layer within your search ads.

00:32:04.000 –> 00:32:09.000
It expands your keyword reach by finding more broad match, longer-tail keywords.

00:32:09.000 –> 00:32:14.000
continuously throughout your campaign. So not just in that initial setup period.

00:32:14.000 –> 00:32:19.000
And then it will send users to the best landing page for their search queries.

00:32:19.000 –> 00:32:25.000
Um, I’ll dive more specifically into these tools. Now, so we can look at them.

00:32:25.000 –> 00:32:33.000
So, as mentioned, data-powered targeting is an excellent way to ensure your ads reach people who actually want to digest your content.

00:32:33.000 –> 00:32:41.000
Rather than paying for your ads to be seen by everyone, data-powered targeting allows your ads to be seen by highly engaged audiences.

00:32:41.000 –> 00:32:47.000
To do this, you can import Google Analytics for key events to your Google Ads account.

00:32:47.000 –> 00:32:53.000
This is a very highly recommended by Google, so that you can serve ads based on conversion performance.

00:32:53.000 –> 00:33:00.000
Rather than just impressions or clicks. How this works is you connect your GA4 account to Google Ads.

00:33:00.000 –> 00:33:05.000
and simply select a key event to import. It should take probably less than 5 minutes.

00:33:05.000 –> 00:33:11.000
If you have that connection set up. For example, um, this key event could be streams or become a member clicks.

00:33:11.000 –> 00:33:17.000
By adding these goals to your campaigns, Google will serve your ads to people who are more likely to watch your station’s content.

00:33:17.000 –> 00:33:24.000
Rather than people who may click on your ad and then immediately bounce.

00:33:24.000 –> 00:33:30.000
After you have figured out the data targeting portion, uh, you want to use in search ads, you can now create the search ad itself.

00:33:30.000 –> 00:33:41.000
Google continues to make this easier and easier. You can use AI-generated creative to do keyword research, write ad headlines and descriptions, and find images, all from one link.

00:33:41.000 –> 00:33:45.000
All you have to do is start a campaign, input your landing page URL.

00:33:45.000 –> 00:33:50.000
And Google will find keywords based on your page’s SEO optimization.

00:33:50.000 –> 00:33:56.000
Based on that keyword list and your landing page, Google will then offer auto-generated headlines and descriptions.

00:33:56.000 –> 00:34:02.000
and identify images on that webpage from your ad that you can simply accept, and it will include in the ad itself.

00:34:02.000 –> 00:34:07.000
It will then generate a preview of what your ad looks like based on everything it has created.

00:34:07.000 –> 00:34:10.000
So it makes it very, very easy for you to spend.

00:34:10.000 –> 00:34:17.000
Um… and again, it will have done all of this from simply copying your station’s landing page and pasting it into the dashboard.

00:34:17.000 –> 00:34:25.000
Again, I will caveat, you will always want to review that Google… what Google has created, because it may be grabbing information that you do not want.

00:34:25.000 –> 00:34:32.000
Um, however, because it’s so easy, you no longer have to do a ton of research and copywriting and image finding.

00:34:32.000 –> 00:34:37.000
Um, and it has also given you a hint into how Google is reading your landing pages in general.

00:34:37.000 –> 00:34:44.000
So if you were to input a Ken Burns documentary page, and it’s suddenly giving you headlines for Great American Recipe.

00:34:44.000 –> 00:34:49.000
then you might want to reevaluate that Kensburn page and see why Google isn’t recognizing your show content.

00:34:49.000 –> 00:34:59.000
Um, it could indicate further optimizations are required, and gives you that additional insight into how Google is indexing your pages.

00:34:59.000 –> 00:35:06.000
Finally, I wanted to share a new feature that Google has recently rolled out called AI Max for Search Ads.

00:35:06.000 –> 00:35:12.000
This is an AI Max enhancement, which you can toggle on for new search ads or even existing search ads.

00:35:12.000 –> 00:35:18.000
Um, it’s technically still in beta mode, but I believe it has been rolled out across all accounts.

00:35:18.000 –> 00:35:23.000
So you are definitely, um, able to try this if you have a Google Ads account.

00:35:23.000 –> 00:35:29.000
The goal is to expand your campaign’s reach and capture search intent you might otherwise have missed.

00:35:29.000 –> 00:35:36.000
Um, it uses Google’s AI algorithm to find those additional broad match queries beyond your existing keywords.

00:35:36.000 –> 00:35:40.000
And instead of limiting ads to just your chosen keywords.

00:35:40.000 –> 00:35:43.000
It will reach, um, a broader set of relevant queries.

00:35:43.000 –> 00:35:49.000
Um, that are potentially more difficult to find if you’re using a tool or something like that.

00:35:49.000 –> 00:35:54.000
Um, it’s called, like, broad match logic, or keyword list matching.

00:35:54.000 –> 00:36:03.000
just simply finding those keywords based on the context of your ads, landing pages, URLs, and user intent.

00:36:03.000 –> 00:36:15.000
Um, it will then recommend auto-generated copy, which, as we discussed, it already kind of did. However, it is labeling this as an additional benefit, just because, again, as those keywords are continuously.

00:36:15.000 –> 00:36:23.000
added into your campaign as it runs. Um, it’s trying to make sure that it’s also presenting relevant headlines for those newer keywords.

00:36:23.000 –> 00:36:31.000
And then, instead of always sending users to a fixed landing page, um, AIMax can automatically choose from your entire website’s pages.

00:36:31.000 –> 00:36:34.000
the one that’s most relevant to your user’s query.

00:36:34.000 –> 00:36:42.000
The goal is that this will help match intent better, um, so that you can have fewer campaigns, but still have improved conversions.

00:36:42.000 –> 00:36:55.000
So, if you are running a drama campaign, and AI Max is finding all drama show keywords on PBS, they’re going to find keywords that are relevant to Miss Scarlett, All Creatures, Standington, etc.

00:36:55.000 –> 00:37:08.000
And instead, um, of just sending people to that Drama Explore page, it’s actually going to direct them to a specific show page for that search. So it will show them the All Creatures page, rather than the Drama Explore page.

00:37:08.000 –> 00:37:19.000
However, with all of this said, and it sounds really great and can be really helpful, um, we have been experimenting with this, and, um, are very aware that we need to monitor it closely.

00:37:19.000 –> 00:37:24.000
Um, we’ve seen great, um, results. We’ve seen increased conversions and impressions.

00:37:24.000 –> 00:37:29.000
However, it does, um, require a lot of time, because.

00:37:29.000 –> 00:37:35.000
there was a bit of a learning curve as it tried to figure out what exactly our campaign was optimizing toward.

00:37:35.000 –> 00:37:44.000
So we launched the AI Max feature in our drama campaign, um, and Google’s AI began to index all PBS popular keywords.

00:37:44.000 –> 00:37:53.000
across the entire site, rather than just finding keywords that were related to dramas. So we saw a bunch of generic PBS keywords grouping in that drama campaign.

00:37:53.000 –> 00:37:57.000
and even keywords like Ken Burns, Ken Burns documentaries, and.

00:37:57.000 –> 00:38:09.000
Command Burns American Revolution. And obviously, this would be a terrible experience for someone who is looking to watch a historical documentary, clicking on the ad, and then being routed to the PBS drama page.

00:38:09.000 –> 00:38:18.000
So, through that, we had to extensively go through that generated keyword list and begin adding negative keywords and URLs that we did not want to show in our AI Max ad.

00:38:18.000 –> 00:38:27.000
After we kind of consistently did this for a week or so, uh, the algorithm began to learn and understand a lot better what we really wanted.

00:38:27.000 –> 00:38:32.000
And then it started to find those longer tail keywords that we wouldn’t have thought of, like.

00:38:32.000 –> 00:38:38.000
say, TD detectives in a crime series, or a period drama set in the UK.

00:38:38.000 –> 00:38:45.000
and even longer ones, rather than just recommending. PBS app, or science documentary, or something like that.

00:38:45.000 –> 00:38:55.000
So with all of that said, if you would like to test the AMX feature with your search ads, just please make sure you are checking the keyword list routinely to ensure that you aren’t appearing for irrelevant searches.

00:38:55.000 –> 00:39:02.000
As I’m sure you all know, AI can be very helpful, but it does not replace humans. It still needs a lot of supervision.

00:39:02.000 –> 00:39:08.000
Um, and with that said, I would always recommend testing some of these new tools if you do have a little bit of budget.

00:39:08.000 –> 00:39:14.000
Um, but obviously, kind of, the tried and true search ads at the moment appear to be.

00:39:14.000 –> 00:39:22.000
Still, um, worthwhile spending on. So before I shift back to the organic strategy for Richard, I’ll check to see if we have any.

00:39:22.000 –> 00:39:24.000
Questions?

00:39:24.000 –> 00:39:31.000
Yeah, there was one from Dave. Uh, do you know how Google decides if your search is considered a local search?

00:39:31.000 –> 00:39:49.000
That’s a good question. So, um, we tend to do this based on, I think it’s… probably just your location. Um, so Google, um, can read basically where your website is located, as well as indexes your page to see where your address is and all of that. And you can then target.

00:39:49.000 –> 00:39:53.000
Um, local communities based on that, um, down to the zip code level.

00:39:53.000 –> 00:40:02.000
And so, when your ad appears, and you’re bidding on, um, keywords that were within that local domain, so if it was Nebraska and you were bidding on.

00:40:02.000 –> 00:40:14.000
you know, Nebraska local TV or something, and it saw you were in Nebraska running that ad, it would consider that a local search. Now, I don’t know, like, if Google has more high-wire techniques.

00:40:14.000 –> 00:40:23.000
Where they’re trying to figure that out. Um, but I think it is largely just based on the location of your keyword, and if that is, like, considered a place.

00:40:23.000 –> 00:40:27.000
That’s, um, versus a national keyword, if that’s helpful.

00:40:27.000 –> 00:40:29.000
I can actually speak to that a little bit more of the, like, technical logic.

00:40:29.000 –> 00:40:33.000
Oh.

00:40:33.000 –> 00:40:46.000
Uh, so Google understands, uh. certain things as what we call entities, uh, and there can be national or local sort of designations applied to those.

00:40:46.000 –> 00:40:54.000
So, for example, Google understands. that PBS News is an entity, and that can be related to other entities, like.

00:40:54.000 –> 00:41:05.000
Julie Woodruff, uh… There can be categories or things that are applied to that. Uh, so, for example, Google is smart enough to know that PBS News is a national.

00:41:05.000 –> 00:41:15.000
a national organization, it’s a national entity. Uh, and then, by that same logic, it’s able to distinguish what might be considered local entities.

00:41:15.000 –> 00:41:22.000
Uh, the only caveat there is sometimes. It can depend, just because there may be multiple entities with the same name.

00:41:22.000 –> 00:41:27.000
So, for example, one that’s come up a few times, especially where I live, is NOVA.

00:41:27.000 –> 00:41:31.000
If you search for NOVA, like, obviously there’s the PBS show NOVA.

00:41:31.000 –> 00:41:44.000
But, uh, we also live in northern Virginia, which is often shortened to NOVA, so… it can sort of be a little, like, weird sometimes with the designation and what results you might see.

00:41:44.000 –> 00:41:58.000
Uh, and then a question from Tim. Google runs the Google Ad Grant program, which provides qualifying nonprofits up to 10K per month in search ad credits.

00:41:58.000 –> 00:42:02.000
Google Ad accounts are separate from paid Google accounts and must follow, yeah.

00:42:02.000 –> 00:42:05.000
And what you’re more equipped to speak to this one than I am.

00:42:05.000 –> 00:42:15.000
Yeah, I think this is a shout-out from Tim about the Google Ad Grant program, which is a very cool program that Google runs, and it does provide that.

00:42:15.000 –> 00:42:39.000
10K per month. As Tim said, um, it does have very specific requirements that websites have to meet, um, that paid accounts do not have to meet. Um, so if you do join that program, um, it is essential that you go through several of their steps. Um, I also work with stations on their, um, Google Ad Grant program, so I’m happy to talk more about this, um, later, or send any follow-up any.

00:42:39.000 –> 00:42:58.000
information if people have questions on it. Um, are there any limitations on using those funds for AI SEM? Um, there are not for, uh, AI Max, you can run AIMAX ads within your, uh, Google Grant account. Um, they’ve rolled that out, and we have actually tested it.

00:42:58.000 –> 00:43:05.000
Again, the grant is a great program, but I will say that performance is always, always lower than if you’re actually paying for ads.

00:43:05.000 –> 00:43:10.000
Due to a lot of their restrictions, um, around keywords and competitor keywords and such like that.

00:43:10.000 –> 00:43:19.000
But yes, you can try it out, and that’s where we first tried it out, because it is free funds. So why not see how it works without our own money?

00:43:19.000 –> 00:43:33.000
Gotta test with the free Google Bucks. Awesome. Uh, well, thank you so much, Emily. Uh, this next portion of our presentation was going to get a little bit more into the tactical side of the SEO, uh, things with.

00:43:33.000 –> 00:43:38.000
how you can target AI overviews and what SEO looks like in, sort of, the AI space.

00:43:38.000 –> 00:43:49.000
We’re running a little bit short on time, so I might fly through this a little bit quickly, but I will say we do have a QR code and a bit.ly link at the end of this presentation, where you can download a bunch of our resources.

00:43:49.000 –> 00:43:54.000
Uh, which has a link to the slide deck, as well as a transcript, which.

00:43:54.000 –> 00:44:01.000
Uh, highly recommend digging into that if you want to get a little bit deeper into the tactical side of things.

00:44:01.000 –> 00:44:11.000
Uh, but just to keep things moving. Uh, first of all, uh, so how to rank in AI overuse? I know we’ve talked a lot about.

00:44:11.000 –> 00:44:17.000
the tactics and how AI overviews appear in things, but what do you do if you actually want to rank?

00:44:17.000 –> 00:44:26.000
in an AI overview. Uh, well, what we found generally is that if you want to rank in an AI overview, you just need to rank highly in search.

00:44:26.000 –> 00:44:34.000
Uh, there was a study published that found that 80% of the links that were featured in AI overviews also ranked on page 1 of Google.

00:44:34.000 –> 00:44:43.000
And what we found generally is that. AI or answer engine optimization, all of these new, sort of.

00:44:43.000 –> 00:44:48.000
buzzword jargony things. tend to match very closely with.

00:44:48.000 –> 00:44:54.000
traditional SEO tactics. So, what we found generally is that if you want to appear highly in AI content.

00:44:54.000 –> 00:45:00.000
just do SEO. For now, at least.

00:45:00.000 –> 00:45:04.000
But why would you want to target AI reviews if we know that they.

00:45:04.000 –> 00:45:15.000
Negatively impact organic and paid click-through rates. Well, there’s a couple different reasons. I think one of the main reasons is if it’s associated with your brand. Like, in this case.

00:45:15.000 –> 00:45:24.000
Antiques Roadshow locations. That’s something that clearly you want to make sure that the content is correct, and also have a link for those people that want to click through, uh.

00:45:24.000 –> 00:45:29.000
to secure a ticket. For example, you want to make sure that they’re going to the right place.

00:45:29.000 –> 00:45:35.000
So there are some circumstances where it would be valuable to target AI overviews.

00:45:35.000 –> 00:45:40.000
So, going back to one of Chad’s questions earlier, uh, with how to track AI overviews.

00:45:40.000 –> 00:45:47.000
So, unfortunately, there’s no way to track. AI overviews within Google Analytics or Google Search Console.

00:45:47.000 –> 00:45:53.000
uh, traffic to your website from these features is counted as organic search traffic.

00:45:53.000 –> 00:45:56.000
there’s just no way to segment it out to see, like.

00:45:56.000 –> 00:46:03.000
X percent of traffic to your website from organic searches coming from AI overviews versus traditional search.

00:46:03.000 –> 00:46:11.000
There’s no way to do that, unfortunately. However, you can track where AI overviews appear, but you do have to pay for a third-party tool.

00:46:11.000 –> 00:46:14.000
Uh, the one that we use at PBS is called Ahrefs.

00:46:14.000 –> 00:46:22.000
I am an Ahref stan, highly recommend, but I know, of course, not everyone has a budget available for these third-party tools.

00:46:22.000 –> 00:46:27.000
Uh, there are two other ones, SEMrush and Moz Pro, that both have these features.

00:46:27.000 –> 00:46:34.000
Uh, and we have links in our deck if you want to look into them yourselves.

00:46:34.000 –> 00:46:44.000
So, again, flying through some of the tactics. Uh, one of the things that’s really essential when it comes to doing SEO in the current era.

00:46:44.000 –> 00:46:49.000
is going to be conducting keyword research, and just echoing what Emily talked about before.

00:46:49.000 –> 00:46:55.000
It’s really important to distinguish between those branded and those non-branded keywords.

00:46:55.000 –> 00:46:58.000
Uh, so we like to think about branded keywords as, sort of.

00:46:58.000 –> 00:47:09.000
doing defense in the SEO side of things, so these are people that are already aware of your content, already interested in your content, and you just want to make sure that they can get to the right place.

00:47:09.000 –> 00:47:15.000
This is also typically going to be the people that are going to convert to things like passport signups or.

00:47:15.000 –> 00:47:27.000
Uh, or streaming at a higher rate. And those non-branded keywords tend to be a little bit more of the offensive, like, how do you reach new audiences that may not be familiar with, uh, with your content?

00:47:27.000 –> 00:47:36.000
So, that could be people searching something like British period dramas, or local news, for example.

00:47:36.000 –> 00:47:42.000
Something else that’s really important when we think about doing SEO in this current era is search intent.

00:47:42.000 –> 00:47:52.000
And that’s the underlying… thing that a user wants to do when they’re searching on Google, or one of these new AI tools.

00:47:52.000 –> 00:47:59.000
Uh, search intent, at least in the classical, uh, Google search space, tends to have four main intents.

00:47:59.000 –> 00:48:04.000
Uh, so there’s those informational intents, so if someone has a question that they want to answer.

00:48:04.000 –> 00:48:13.000
And this is what we know, uh, is. Being the hardest hit at the moment by these AI overviews. These are primarily targeting these informational.

00:48:13.000 –> 00:48:21.000
search queries. Commercial is a user that wants to do an action, so they want to stream Leonardo da Vinci.

00:48:21.000 –> 00:48:26.000
Or sign up for Passport, there’s an action that they want to take.

00:48:26.000 –> 00:48:32.000
Uh, this is another really important one, especially in the marketing space where you want to get users that will convert.

00:48:32.000 –> 00:48:37.000
Additionally, we have navigational, which is a user that’s looking for a specific page.

00:48:37.000 –> 00:48:41.000
And then there’s transactional, so a user that wants to make a transaction.

00:48:41.000 –> 00:48:46.000
It’s just really important to think about user intent, uh, when you’re optimizing for search.

00:48:46.000 –> 00:48:51.000
Because you want to be matching the right user intent with the right page.

00:48:51.000 –> 00:48:55.000
So, for example, if someone’s searching, like, what do camels eat?

00:48:55.000 –> 00:49:04.000
You don’t want to give them, like, hey, here’s an hour-long nature documentary, because they just want, like, a short little snippet of text, probably, as opposed to.

00:49:04.000 –> 00:49:17.000
Digging through, like, a full documentary. You’re just going to provide a better user experience to your users by thinking about search intent, which is going to help your search profile overall.

00:49:17.000 –> 00:49:29.000
But the biggest tactic that we really want to hammer home is maximizing engagement. Uh, so thinking about your site KPIs and your business goals when you’re creating web content.

00:49:29.000 –> 00:49:42.000
And when you’re optimizing for search. Uh, because if you’re generating a lot of traffic, it’s all sort of a moot point if users don’t have a way to meaningfully convert to whatever KPI is important to your business.

00:49:42.000 –> 00:49:48.000
So, we always highly recommend doing things like embedding short-form video into blogs and articles.

00:49:48.000 –> 00:49:54.000
plug in your newsletter CTAs, incorporating passport messaging, linking to related content.

00:49:54.000 –> 00:49:58.000
And again, this can also tie back to search intent, because.

00:49:58.000 –> 00:50:06.000
You want to think about what the search intent the user has when they’re landing on the page, and think about what action they may be willing to take.

00:50:06.000 –> 00:50:11.000
Uh, like, a user reading an article may not be ready to pull the trigger on donating.

00:50:11.000 –> 00:50:16.000
But they may be interested in watching some short-form video. Uh, so think about.

00:50:16.000 –> 00:50:20.000
what action they may be willing to take, and try to find ways to incorporate that.

00:50:20.000 –> 00:50:32.000
Where you can. But before we get into the Q&A, Emily ended our… Emily and I ended our annual meeting presentation by giving some predictions.

00:50:32.000 –> 00:50:39.000
Uh, and things have changed a lot, uh, since May, since annual meeting. So we wanted to give a little bit of an update of.

00:50:39.000 –> 00:50:48.000
what our predictions were then, and then what’s happened, uh, in the last few months since.

00:50:48.000 –> 00:50:55.000
Yes, so our first prediction was we predicted ads will begin to appear more frequently within AI overviews.

00:50:55.000 –> 00:51:04.000
Um, this is true. We have seen that they have begun to appear more frequently, um, and you can also see this in Google’s revenue stream.

00:51:04.000 –> 00:51:13.000
Um, in Q3, Google made over $56 billion, a 14% increase from the previous year simply from search.

00:51:13.000 –> 00:51:20.000
Um, these overviews are providing a new avenue for Google to monetize its search positioning and AI features.

00:51:20.000 –> 00:51:31.000
Um, and now you can see, actually, that text and shopping ads from existing search shopping, and Performance Mac campaigns are eligible to show within AI overviews.

00:51:31.000 –> 00:51:42.000
I believe originally it was mostly just an e-commerce feed that was appearing, um, an AI overviews, but it has extended a bit, even though e-commerce is still dominating.

00:51:42.000 –> 00:51:48.000
Um, and studies estimate that ads appear within AI overviews 13-20% of the time now.

00:51:48.000 –> 00:51:56.000
However, you can’t directly do that, you can’t directly target ad placement within AI overviews, as we mentioned.

00:51:56.000 –> 00:52:03.000
And Google still does not provide any reporting specifically on when your ads are appearing in AI overviews, even if.

00:52:03.000 –> 00:52:09.000
you’re trying to bid across Surgeon AI. Everything is just grouped under the same ad performance.

00:52:09.000 –> 00:52:16.000
And this kind of brings us to our second prediction, that Google will begin sharing engagement data on AI overviews and Google Ads.

00:52:16.000 –> 00:52:21.000
This has not happened yet. This prediction has not come true. Hopefully, it will come true since.

00:52:21.000 –> 00:52:26.000
everyone is spending and trying to understand more how AI overviews is affecting their actual traffic.

00:52:26.000 –> 00:52:39.000
Um, I will say we’ve spoken to some of our Google reps, and they’re claiming that people are actually using search more with AI overviews because they read the overview and then search something mentioned in that overview.

00:52:39.000 –> 00:52:46.000
However, even though they may be spending time on Google, there is no data to suggest that people are then clicking through to publisher websites.

00:52:46.000 –> 00:52:55.000
Rather than just reading more AI overviews. Um, they’re continuing to gatekeep this information.

00:52:55.000 –> 00:53:02.000
Our third prediction was that Google will begin integrating AI into other features within Google Search.

00:53:02.000 –> 00:53:04.000
And this is another one that has come true.

00:53:04.000 –> 00:53:10.000
Uh, so if you’re Googling, you’ll probably see things like, you know, there’s going to be the Maps feature.

00:53:10.000 –> 00:53:14.000
images, a lot of times there are related questions.

00:53:14.000 –> 00:53:20.000
And a lot of these related questions, especially in other features, like text features that appeared in it.

00:53:20.000 –> 00:53:28.000
in Google search results. used to pull content directly from websites, and what we’ve seen over the last couple of months is increasingly.

00:53:28.000 –> 00:53:36.000
These are using AI-generated responses. So, what this means, ultimately, is they’re… are just fewer and fewer.

00:53:36.000 –> 00:53:44.000
places within the Google Search ecosystem. that are pulling content from websites, it’s increasingly pulling content from AI.

00:53:44.000 –> 00:53:51.000
uh, continued proliferation of zero-click searches. This was true then, and is still true now.

00:53:51.000 –> 00:53:56.000
Uh, as there’s just more AI summary resp… uh, content, uh, within search.

00:53:56.000 –> 00:54:02.000
There’s just going to be fewer and fewer clicks. Uh, so it becomes really important that we.

00:54:02.000 –> 00:54:08.000
are aware of what’s going on. Excuse me. Uh, and we take, uh, a careful approach.

00:54:08.000 –> 00:54:20.000
Uh, to optimizing for search. And our last prediction, our spiciest prediction, was that Google would lose market share as users would begin migrating to other AI platforms.

00:54:20.000 –> 00:54:26.000
This has not appeared to have come true, but I think it takes a little bit of reconfiguring in our brains of.

00:54:26.000 –> 00:54:30.000
What market share looks like in the search landscape.

00:54:30.000 –> 00:54:37.000
Uh, because traditionally, you know, it was. Google versus Bing versus Yahoo and the other, like, really small ones.

00:54:37.000 –> 00:54:41.000
And a user used one search engine and not the other.

00:54:41.000 –> 00:54:47.000
Uh, with these new AI tools, and this new ecosystem that we’re in, that’s not really the case.

00:54:47.000 –> 00:54:51.000
I think what’s more likely to happen is that.

00:54:51.000 –> 00:54:57.000
Users are going to be using multiple. search tools, but for different purposes.

00:54:57.000 –> 00:55:02.000
So, they’re still going to be using Google when they want to find a page on a website.

00:55:02.000 –> 00:55:08.000
But they’re going to be turning to something like ChatGPT when they want to ask a more complicated query.

00:55:08.000 –> 00:55:12.000
Or they’re turning to YouTube, or they’re turning to TikTok.

00:55:12.000 –> 00:55:18.000
People are just going to migrate to different platforms for different needs. So, Google will never go away.

00:55:18.000 –> 00:55:25.000
But I think we’ll see… sort of a diversification, uh, within search.

00:55:25.000 –> 00:55:34.000
Uh, but with that, uh, I’ll flip over to our QR code slide, and we can… I saw… I’ve been… Following along as the chat’s been popping off.

00:55:34.000 –> 00:55:44.000
Uh, but if you have any questions, feel free to drop them in, but… can take a look at some of the questions.

00:55:44.000 –> 00:55:45.000
Oh, go ahead, Emily.

00:55:45.000 –> 00:55:48.000
Yes. Oh, I was just going to say, I think there’s a lot of good discussion about Google Ad Grant Program, and then Moz Pro on how to sign up.

00:55:48.000 –> 00:55:54.000
Yeah.

00:55:54.000 –> 00:56:04.000
Um, I believe the next question was, I see GPT bot and Perplexitybot on the PBS.org disallow list. I’m wondering if you can talk a little more about that.

00:56:04.000 –> 00:56:07.000
If you’ve taken any other steps in this direction.

00:56:07.000 –> 00:56:13.000
That’s a really good question. That was a decision that was made at PBS.

00:56:13.000 –> 00:56:20.000
Early on, when these bots were sort of. First introduced.

00:56:20.000 –> 00:56:37.000
Uh, I will say the decisions to make. The decisions to block these bots is… like, several steps above my pay grade. So… I can’t comment too much on that, uh, just because I don’t want to speak for the organization.

00:56:37.000 –> 00:56:42.000
Uh, but I think that’s a very robust discussion that needs to happen within the system as a whole.

00:56:42.000 –> 00:56:46.000
But, sorry, I can’t really provide too much detail there.

00:56:46.000 –> 00:56:49.000
Can I jump in on that one just real fast?

00:56:49.000 –> 00:56:50.000
And I totally respect that you’ve got a, kind of.

00:56:50.000 –> 00:56:53.000
Yeah. Yeah.

00:56:53.000 –> 00:57:02.000
You gotta manage up with an answer like that. Um, we’ve been having this kind of discussion between the stations and at the national level for a couple of years. This was a question that came up.

00:57:02.000 –> 00:57:08.000
At the Grand Rapids need a meeting a couple of years back with people asking about what gets blocked and what doesn’t.

00:57:08.000 –> 00:57:14.000
And obviously, every station has to kind of make its own decision. Every organization decides what they want to let in or not, but.

00:57:14.000 –> 00:57:19.000
I believe it is a correct statement, is it not, that.

00:57:19.000 –> 00:57:28.000
Google can’t be blocked, because if you block Google’s AI from crawling you, you will kill your search… you just won’t even show up in search. Is that… correct assertion.

00:57:28.000 –> 00:57:38.000
Uh, for the most part, but Google is nice because they at least have different bots for different things, so there is a separate bot, uh, specifically for their AI systems.

00:57:38.000 –> 00:57:47.000
Uh, it’s also a little bit… difficult, because there’s a technology called RAG retrieval Augmented Generation.

00:57:47.000 –> 00:57:51.000
Where, basically, these tools can go back to a search.

00:57:51.000 –> 00:57:57.000
Uh, index, and pull content from that without having to directly scrape your site. So currently, what we’re doing is we’re.

00:57:57.000 –> 00:58:05.000
preventing these bots from, uh… From scraping our content to use to train their AI models on.

00:58:05.000 –> 00:58:10.000
It’s also extra complicated because there’s been a lot of indication that.

00:58:10.000 –> 00:58:21.000
these AI systems are not respecting that anyway, and they’re, like… They’re spoofing their, um… Uh, themselves to, like, circumvent that blocking.

00:58:21.000 –> 00:58:30.000
So, that could be a whole separate hour-long conversation, uh, but it’s… it’s complicated.

00:58:30.000 –> 00:58:34.000
And I see Anasus, uh, chimed in in the chat saying that, uh.

00:58:34.000 –> 00:58:42.000
There’ll be an update coming at some point, so… She’s got your back. Um… Emily, you want to keep, uh, ticking through some of the questions?

00:58:42.000 –> 00:58:43.000
Oh, sure, definitely.

00:58:43.000 –> 00:58:56.000
Uh, we’re… we’re at the top of the hour, but just to let everyone who’s sticking with us know, we’re gonna go through and try to answer everything before we wrap. Sometimes we go long, I think this is… One, we thought we’d go long on, and it looks like we will, so anyway, Emily, back to you.

00:58:56.000 –> 00:59:05.000
Definitely. Um, so the next question is from Jeff. Is there any indication on whether or when AI overview and AI Bode will become the default search experience?

00:59:05.000 –> 00:59:15.000
I think, Richard, you kind of mentioned that. I think Google’s goal is that AI mode will become the default search experience, even though not a lot of people are currently using it.

00:59:15.000 –> 00:59:28.000
Um, so I think it’s… we’ll see, and I imagine it very much depends on monetization for Google. Um, if they can figure out how to monetize AI mode, then I’m sure they will sink more teeth into how to get people to actually use it.

00:59:28.000 –> 00:59:32.000
Yeah, and that’s exactly what I was going to say.

00:59:32.000 –> 00:59:39.000
Google their public policy has indicated that AI mode or something like it will be sort of the default.

00:59:39.000 –> 00:59:43.000
Uh, at some point, they haven’t indicated when that would happen.

00:59:43.000 –> 00:59:47.000
Uh, but to Emily’s point, I think monetization really is the key.

00:59:47.000 –> 00:59:57.000
The vast majority of Google’s revenue comes from search ads, and they make a lot of money. I think they made.

00:59:57.000 –> 00:59:58.000
56.

00:59:58.000 –> 01:00:05.000
Emily said this yesterday. Q3 was their best quarter ever. They made over $100 billion in one quarter. It was, like, $66 billion from search ads alone.

01:00:05.000 –> 01:00:13.000
Uh, so they really need to figure out what the monetization scheme looks like for these new features before they can make that switch.

01:00:13.000 –> 01:00:14.000
But soon.

01:00:14.000 –> 01:00:27.000
Throw a little more context in there. I was at a session on Google Search at South by Southwest maybe 10 years ago, and they were talking about how the goal, ultimately, at that point, their goal with Google was to.

01:00:27.000 –> 01:00:31.000
To be able to answer your questions or deliver you a result before you even can complete the thought.

01:00:31.000 –> 01:00:40.000
which seemed like a very lofty goal, but, like, pretty soon after that’s when you started seeing predictive text, so you started typing, and then the question would start autofilling on the side.

01:00:40.000 –> 01:00:45.000
And so, this is, to me, just very much in line with.

01:00:45.000 –> 01:00:53.000
that intent that has been a very long-standing intent on Google’s part to really get, as you said earlier, Richard, at, uh.

01:00:53.000 –> 01:00:58.000
intention.

01:00:58.000 –> 01:00:59.000
What else we got?

01:00:59.000 –> 01:01:05.000
Great. Um, then from Lynn, how can we boost local content visibility in a zero-click and AI overview environment?

01:01:05.000 –> 01:01:11.000
how do we get our content seen if no one is looking?

01:01:11.000 –> 01:01:12.000
I… yeah, I could take a stab at this one. Uh, I will say, from what we’ve seen.

01:01:12.000 –> 01:01:16.000
It’s a good question.

01:01:16.000 –> 01:01:23.000
Um… zero-click searches are not sort of an everywhere.

01:01:23.000 –> 01:01:31.000
sort of thing, like, not every search will be a zero-click search. I think there are certain types, like, especially the informational stuff.

01:01:31.000 –> 01:01:38.000
That is very at risk. Uh, of moving into sort of a zero-click space.

01:01:38.000 –> 01:01:42.000
So, trying to find the niches that still can get a click.

01:01:42.000 –> 01:01:47.000
and then focusing there. We do a lot of analysis of.

01:01:47.000 –> 01:01:51.000
The… the projects that we do, and trying to focus on, like.

01:01:51.000 –> 01:01:56.000
What’s still working, and then really trying to lean heavily into that, uh, where we can.

01:01:56.000 –> 01:02:00.000
So I think looking back and doing a lot of analysis of, like.

01:02:00.000 –> 01:02:08.000
We did this… this is… we’re seeing a lot of AI overviews and the responses, we’re not getting the clicks that we’re looking for.

01:02:08.000 –> 01:02:16.000
maybe let’s pivot and try something new. Or, like, hey, we’re still getting a lot of engagement here, even if there’s AI overviews that are appearing.

01:02:16.000 –> 01:02:24.000
trying to find ways to lean into that where possible. I will say news is a sector that we have seen has been relatively.

01:02:24.000 –> 01:02:32.000
uh, stable, uh, given AI reviews, because. Google really hedges their bets when it comes to what they call.

01:02:32.000 –> 01:02:37.000
uh, your money, your life. queries, uh, so people searching things that are, like.

01:02:37.000 –> 01:02:42.000
really sensitive and could, like, adversely impact people’s lives.

01:02:42.000 –> 01:02:48.000
Um, they’re more hesitant to show AI results there because they just don’t really want to get things wrong.

01:02:48.000 –> 01:02:52.000
Uh, so news is a space where we’ve seen, uh.

01:02:52.000 –> 01:02:59.000
AI has been less prevalent, so that can be an opportunity, because I know there’s a lot of stations that operate in the.

01:02:59.000 –> 01:03:03.000
the local news space, uh, as well.

01:03:03.000 –> 01:03:08.000
We’ve talked about this, the three of us, when we were doing pre-production, but when, like.

01:03:08.000 –> 01:03:14.000
Um, the question about… Getting local content to show up.

01:03:14.000 –> 01:03:26.000
There’s a distinction between a tune-in strategy, right? And, like… You know, do you want to get your local documentary or a podcast, if you’re coming at us from the radio side, perhaps?

01:03:26.000 –> 01:03:34.000
Um, you know, in front of folks versus, you know, making sure that facts from an important local news story.

01:03:34.000 –> 01:03:47.000
Uh, don’t get, you know, lost because they’re not getting… optimized for AI. Can you maybe talk a little just about the distinction between those two?

01:03:47.000 –> 01:03:48.000
Yes.

01:03:48.000 –> 01:03:58.000
Yeah, I think, um… We tend to focus a lot on the tune-in side of things that’s just from our leadership, where we’ve, uh, we’ve seen a huge focus.

01:03:58.000 –> 01:04:07.000
I think what’s… what will be interesting and what… sort of to… to Emily’s point from the presentation.

01:04:07.000 –> 01:04:15.000
We’re still waiting from a lot of these tools to start getting data about how people are using them, and I think what would be interesting.

01:04:15.000 –> 01:04:23.000
is if they come up with some, um, metrics that are less click-focused and more in that, like.

01:04:23.000 –> 01:04:33.000
awareness, like, visibility sort of focus, like, almost a CPM-style, uh, metric, as opposed to, like, a click-driven.

01:04:33.000 –> 01:04:43.000
style of focus, uh, which I think, uh… can speak to some of, like, the lighter touch, like, how you’re influencing the conversation versus, like.

01:04:43.000 –> 01:04:49.000
just how you’re appearing and then converting to a click.

01:04:49.000 –> 01:04:50.000
Mm-hmm.

01:04:50.000 –> 01:04:57.000
Uh, I’m not sure if that… that answers your question, but I think that, uh, potentially could sort of elucidate some of the.

01:04:57.000 –> 01:05:04.000
some of the… the gaps that we’re seeing in analysis, and sort of speak to some of the, like, more mission-driven.

01:05:04.000 –> 01:05:13.000
Uh, focus of PBS of, like, trying to create educational content and sort of be a trustworthy space, uh, for information.

01:05:13.000 –> 01:05:22.000
Exactly, and I think that goes back to how overviews are almost like an awareness play now, since it’s kind of inevitable that, you know, more and more abuse will be created.

01:05:22.000 –> 01:05:33.000
It is really essential to be appearing in AI overviews with certain information, and oftentimes it’s just being more of that trusted brand and optimizing for SEO, um, indexing there.

01:05:33.000 –> 01:05:45.000
Um, so if you can appear within that AI overview when people are reading your content, obviously the goal would be for eventually for them to get back to your website, um, so that they could kind of read the.

01:05:45.000 –> 01:05:50.000
you know, full article, or watch the full video.

01:05:50.000 –> 01:05:57.000
Um, to Lynn Hatter, I’ll say, like, this is definitely not the last time we’re going to be talking about this topic, and I think there are going to be nuances that, as.

01:05:57.000 –> 01:06:13.000
As a body of best practices develop, with some professionalism around this, not that y’all aren’t professional, but like, you know, it’s just early days, we’re not even in version 0.5 with some of this stuff that we’ll keep talking about this, and I would expect this to be.

01:06:13.000 –> 01:06:18.000
You know, a topic here, but maybe some of the other kind of interest groups.

01:06:18.000 –> 01:06:23.000
Let’s jump… I’m gonna jump to the other, Lynn Valakat’s question real quick here. Where can we find the.

01:06:23.000 –> 01:06:30.000
this disallow list to make sure staff know this in addition to station’s own policies.

01:06:30.000 –> 01:06:36.000
Must be referenced to be Jan’s question.

01:06:36.000 –> 01:06:51.000
This is another one where, uh… This also… I think it’s just difficult to talk about the… the disallow policy, because that’s sort of in the… the policy sphere that’s, like, again, not really something that Emily and I.

01:06:51.000 –> 01:07:02.000
Are making policy on. So, I would say that this is something we can follow up with, with PBS leadership to try and get some more clarity that’s available at stations of, like, what the.

01:07:02.000 –> 01:07:10.000
What our policy is, and, you know, reasoning behind that, but… just hesitant to speak on that directly ourselves.

01:07:10.000 –> 01:07:24.000
Sure. I hear you on that. Um… And I actually didn’t know if there was a public disallow list until I saw Vijan’s question, so… but I think Anna said she’d do some follow-up with some of that. So, Lynn, I think you’re… you’re on the record with that one.

01:07:24.000 –> 01:07:37.000
Chuck Road back to Christian Nutsen, uh… What do you think about the idea that these findings on impacts of AI overviews might suggest that the comparative audience generating value of.

01:07:37.000 –> 01:07:43.000
Video content is increasing in comparison to text-driven content.

01:07:43.000 –> 01:07:45.000
Any thoughts on that, guys?

01:07:45.000 –> 01:08:03.000
I think that’s a good call. I mean, we’ve just seen that AI overviews have started to add YouTube video. Well, obviously they’re prioritizing YouTube because it’s a Google entity. The fact that they’re adding those videos now means that there’s definitely value in it, um, and we see on, like, social platforms, TikTok has become kind of the search engine for Gen Z, right?

01:08:03.000 –> 01:08:08.000
And so, as people become more accustomed to vertical video and videos to learn things.

01:08:08.000 –> 01:08:12.000
I imagine, I hope, that PBS will also become kind of like.

01:08:12.000 –> 01:08:19.000
that guiding light toward trusted, safe content, because we have so much video that people can watch easily.

01:08:19.000 –> 01:08:20.000
Got it. And I’m… I’m scrolling, so I’m actually seeing BJN actually said, like.

01:08:20.000 –> 01:08:23.000
Yeah.

01:08:23.000 –> 01:08:33.000
He just checks the robots.txt file, uh, which is… Something some of you know how to do, and some of, like, I’m not sure I would be able to actually figure that.

01:08:33.000 –> 01:08:40.000
But he’s put the URL in there, so there you go. Um, so you can kind of see that, too. Um, let’s see.

01:08:40.000 –> 01:08:47.000
Still going, uh, and we will get through all of these, if nothing else, for the recording, so if you do need to dump off, get the answers in here.

01:08:47.000 –> 01:08:59.000
You’ll be able to kind of scroll and watch later. Um, has PBS done research on SEO and AI in other search engines besides Google? Um, and I’ll just kind of augment a little bit and say, like.

01:08:59.000 –> 01:09:08.000
And also, what about AI browsers? Like, what, OpenAI is rolling out, and, you know, the lines definitely are blurring on a number of fronts here.

01:09:08.000 –> 01:09:16.000
Yeah, that’s a great question. We have not really done much research into traditional search engines other than Google.

01:09:16.000 –> 01:09:29.000
Bing is the number two, uh, but… for PBS.org, I think the stat for 2024, at least, was, like, 93% of our organic search traffic was coming from Google.

01:09:29.000 –> 01:09:34.000
And what I’ve found, not necessarily at PBS.org or at PBS, but in.

01:09:34.000 –> 01:09:46.000
previous roles is that. Bing and other search engines tended to follow Google, uh, in a lot of their guidance and optimization strategies.

01:09:46.000 –> 01:09:55.000
Uh, so we haven’t… we’re focusing on Google mostly because they’re the biggest one, and it has the biggest impact, and we just have minimal resources, to be honest.

01:09:55.000 –> 01:10:04.000
Uh, so, uh, we’re focusing on number one. Uh, to your point, Chad, about, uh, some of the alternate browsers.

01:10:04.000 –> 01:10:08.000
Uh, or even just, like, chatbots themselves as sort of, like, a search engine.

01:10:08.000 –> 01:10:15.000
We’ve done a little bit of preliminary research into ChatGPT. Again, ChatGPT being the biggest one.

01:10:15.000 –> 01:10:22.000
Um, there are tools out there, uh, we’ve met with a couple of tools, um, that can.

01:10:22.000 –> 01:10:28.000
You know, that focus on optimizing for ChatGPT or these other LLMs.

01:10:28.000 –> 01:10:31.000
profound, uh, is… is the biggest one that we spoke to.

01:10:31.000 –> 01:10:36.000
To be honest, they’re just very expensive. And we found that.

01:10:36.000 –> 01:10:43.000
At this point, at least. the ROI, uh, in terms of traffic and streams and conversions.

01:10:43.000 –> 01:10:48.000
is still relatively low, so I think it’s something that we’re looking to do some, like.

01:10:48.000 –> 01:10:54.000
manual, uh, analysis ourselves, and then sort of determining from there if it’s worth, uh, looking into, uh.

01:10:54.000 –> 01:11:00.000
a tour at a later date. So, it’s still early days. Uh, we haven’t really done much.

01:11:00.000 –> 01:11:11.000
Uh, research into the browser side of things, but I think that would be good to do, at least from, like, a… a UX, like, accessibility sort of perspective, just to, like.

01:11:11.000 –> 01:11:14.000
Hey, let’s download them and make sure that all of our content loads.

01:11:14.000 –> 01:11:20.000
In a correct way. Yes. Uh, ChatGPT a little bit.

01:11:20.000 –> 01:11:26.000
other search engines, not so much, and browsers. Probably on the horizon at some point.

01:11:26.000 –> 01:11:39.000
Not that we needed a reason to get back together to discuss this, but it definitely seems like we’ll… We’ll have them. Um, I can tell folks that just from a Nebraska standpoint, I can see referral traffic coming to us from ChatGPT now.

01:11:39.000 –> 01:11:46.000
Probably a few hundred a month, but the line goes… the line is trending up.

01:11:46.000 –> 01:11:53.000
And so, um, I don’t think it’s gonna go down from there. A perplexity… Maybe a dozen a month, it’s low.

01:11:53.000 –> 01:12:01.000
Um, but, you know, you could go and, you know, if you have access to your analytics for your site, you can kind of go and begin to just track those so you can get a sense.

01:12:01.000 –> 01:12:11.000
Kind of like we used to do with mobile back in the day, of like, when is mobile gonna cross that kind of threshold of, like, a significant percentage of site traffic kind of feels the same to me.

01:12:11.000 –> 01:12:22.000
Um, the tools you were just talking about tools, I think this maybe gets to what Ashley’s asking, so I’m going to jump to Ashley’s question real quick, which is there any way we can find out which of our content is linked to AI overview? That’s what you were talking about, Ahrefs.

01:12:22.000 –> 01:12:26.000
Uh, Moz Pro, I… what was the third one?

01:12:26.000 –> 01:12:27.000
SEMrush. Yeah.

01:12:27.000 –> 01:12:32.000
was it SEMrush? Semrush? Okay. SEMrush just got bought by Adobe, I think, too, did they not? Yeah.

01:12:32.000 –> 01:12:34.000
Yeah, for, like, $2 billion. Yeah.

01:12:34.000 –> 01:12:49.000
That’ll be an interesting thing to watch, so keep an eye on that one. Um, and then, uh, we’re just about to wrap up. Uh, I’m gonna dial into Mark’s comment here, and I’m gonna… I’m going to take the answer, you guys can jump in if you want, but I don’t think it’s a real fair question for you to…

01:12:49.000 –> 01:12:57.000
to, uh, to handle, um, and that is, you know, do you think we can say, uh, with any confidence, 100% or less.

01:12:57.000 –> 01:13:04.000
Uh, that deciding to block these bots could potentially have a negative impact on discoverability and LLM results.

01:13:04.000 –> 01:13:10.000
Um, it just strikes me that we’ll all have to explain the implications of a decision like that.

01:13:10.000 –> 01:13:15.000
And that it’s more complicated than simply keeping AI companies from hoovering up our data.

01:13:15.000 –> 01:13:24.000
Um, you know, again, I think every organization’s got to make that decision. At Nebraska, we decided, you know, pretty much.

01:13:24.000 –> 01:13:31.000
From the early days of ChatGPT being public, you know, 22, early 23, we were going to allow.

01:13:31.000 –> 01:13:34.000
the data to be hoofered up, because we’re public.

01:13:34.000 –> 01:13:40.000
media entity. We produce our content for the public good, and so if we felt like.

01:13:40.000 –> 01:13:51.000
You know, that was part of our mission. Uh, to make our content accessible, and also just from the standpoint of garbage in, garbage out, right? So, like, we have good content.

01:13:51.000 –> 01:13:57.000
much rather they hoover up good content so that the models get better and at least provide better service if you’re.

01:13:57.000 –> 01:14:04.000
Asking questions about Nebraska. So that was a decision that Nebraska made to not block anything. I think it is.

01:14:04.000 –> 01:14:14.000
Undisputed that Google does. punish you. Like, they have different bots and things like that, but they do punish you. And a lot of this is also just dependent on, like, you can tell them not to… to…

01:14:14.000 –> 01:14:25.000
To scrape your site. What we’ve found is a lot of them ignore that. They took… Bijan talked about the robots.txt file, that’s where you tell sites, hey.

01:14:25.000 –> 01:14:34.000
don’t look at me. I don’t want to be seen. Um, and we know that most of the companies, Anthropic was one of the early ones, like, you know, they have Claude, Claude is the nice guy.

01:14:34.000 –> 01:14:42.000
chatbot and, um… They… we had some instances where they were slowing our site down. The crawls were just grinding us to a halt.

01:14:42.000 –> 01:14:50.000
Um, and… Uh, doing server resets and things like that. So, um, and Anthropic is supposed to be arguably the good guys.

01:14:50.000 –> 01:15:01.000
in the whole cohort of frontier model creators. So, um, you can put that in there, that doesn’t mean they’re gonna follow it, and so we, Nebraska, have just said, like.

01:15:01.000 –> 01:15:12.000
We have mission reasons why we want to not go down that path, so I wanted to just put that out there to say, like, your organization’s leadership, you know, has to make that.

01:15:12.000 –> 01:15:17.000
decision based on your own kind of missions and what you determine is the right.

01:15:17.000 –> 01:15:20.000
Right answer. And PBS has a whole different perspective on this.

01:15:20.000 –> 01:15:27.000
Uh, because they’re national, and they’re protecting national brands, and it’s a different… it’s just a different point of view from the local angle, I think.

01:15:27.000 –> 01:15:45.000
So, I wanted to put that out there. Um… Let’s see, um… Thanks, Nick, uh, for thanking us in Menominee. Um… And, let’s see, Kristen, traffic-induced items on PBSwisconsin.org, that’s really referred to by ChatGPT.

01:15:45.000 –> 01:15:48.000
Has been steadily increasing, so you’re seeing kind of the same.

01:15:48.000 –> 01:15:52.000
thing, and isn’t too far behind Bing. I should check Bing and just kind of see.

01:15:52.000 –> 01:15:59.000
Um, do you guys have any comments on Bing? I mean, we never think about it. Now we talk about perplexity as the other search engine, but, like.

01:15:59.000 –> 01:16:05.000
You know, Bing is technically still. Like, out there for a hot minute, we thought it was going to be a thing.

01:16:05.000 –> 01:16:14.000
And then Microsoft kind of… Deflated, like, with their search and AI, but what do you guys… do you have thoughts on that one?

01:16:14.000 –> 01:16:28.000
On the paid side, we don’t really, uh, do anything with Bing, um, we just see that Google dominates the majority of our searches, so we, you know, spend our budget there. Um, we experimented a little bit with it a long time ago, and it just didn’t seem nearly as many.

01:16:28.000 –> 01:16:32.000
conversions or anything, um, so paid side, no.

01:16:32.000 –> 01:16:41.000
Yeah, not really on the organic side as well. Uh… Honestly, we have paid attention to Bing a couple of times, and it’s usually because they’re, like.

01:16:41.000 –> 01:16:47.000
doing something weird on the technical element that’s having, like, a noticeable impact for the website as a whole.

01:16:47.000 –> 01:16:52.000
Uh, so we mostly only pay attention to Bing when something, like, breaks.

01:16:52.000 –> 01:16:58.000
I will say they have, like, unlimited money, and I know they’re doing a lot of interest… they’re sort of.

01:16:58.000 –> 01:17:03.000
mirroring Google with their integration of Copilot into Bing search results.

01:17:03.000 –> 01:17:06.000
We haven’t really done much testing to determine, like.

01:17:06.000 –> 01:17:14.000
If it’s higher quality, uh, than Google is. I feel like now… I know Google just released, uh.

01:17:14.000 –> 01:17:28.000
Gemini 3, and people are saying Gemini 3 is, like, the best thing ever, uh, so haven’t done any comparisons there, uh, but… Um, yeah, primarily focusing on Google, for the most part.

01:17:28.000 –> 01:17:33.000
Um… I’ve always felt Bing has basically existed to keep.

01:17:33.000 –> 01:17:39.000
Uh, antitrust officials away from Google. Um, and now with perplexity and.

01:17:39.000 –> 01:17:45.000
And ChatGPT Atlas and the other browsers, it’s almost like… It’s not relevant, but to give them their due, like.

01:17:45.000 –> 01:17:49.000
You know, when ChatGPT came out, uh, Google was on the back.

01:17:49.000 –> 01:18:03.000
on the back heel for a surprising amount of time, and, you know, for, you know, again, a hot minute early 23, we were talking about Bing as being the sort of cutting edge.

01:18:03.000 –> 01:18:04.000
Yeah.

01:18:04.000 –> 01:18:13.000
of search meets, uh, meets chatbot. It’s just, like I said, kind of deflated, and… And went away. Um, we should wrap up here, a favor for Chad, we’d love to have a copy of this chat. Oh, it will… the chat will come out.

01:18:13.000 –> 01:18:21.000
in the, um… it will come out as a part of the materials we put on the PMI site.

01:18:21.000 –> 01:18:29.000
Um, for this… to accompany this webinar, um… And then… so just a heads up on that, so we will… definitely that’ll be shared afterwards.

01:18:29.000 –> 01:18:34.000
Um, Mark says, being at a program that was similar to Google Ad Grant.

01:18:34.000 –> 01:18:49.000
But that’s shutting that down on, it’s interesting, I didn’t know that, so… Um, I have one last question for you, Paul, which is, um, who knows when we’ll be able to all get together to discuss this again, hopefully not a year from now, but sometime deeper into 2026.

01:18:49.000 –> 01:19:07.000
Um, what do you guys recommend. as, like, resources that people can, um… follow to self-educate, to stay up on this? Like, what are the… what are the trade… publications or sites that, you know, hit your top 3 list for keeping us educated.

01:19:07.000 –> 01:19:13.000
The one that I check, like, daily is called, uh, Search Engine Roundtable.

01:19:13.000 –> 01:19:21.000
Uh, I’ll drop the link in the… Uh, in the chat. Uh, the guy who… it’s basically run by one person.

01:19:21.000 –> 01:19:30.000
Uh, Barry Schwartz, and it’s basically sort of like… Uh, it mirrors… it’s like search news, so it’s usually, like, up to the minute.

01:19:30.000 –> 01:19:33.000
uh, sort of views of what’s going on in search.

01:19:33.000 –> 01:19:41.000
Um, so, he talks a lot about, like. the Google search algorithm, like, if there’s a core update, or, like, new enhancements.

01:19:41.000 –> 01:19:50.000
Uh, that have appeared in search results. Uh, but he also has increasingly covered, um… AI integrations into search.

01:19:50.000 –> 01:19:55.000
It doesn’t really have as much content focused on, like, if you want to look into, like.

01:19:55.000 –> 01:20:07.000
AI-specific chatbots, or, like, you know. ChatGPT news is not really covered, it’s more sort of like the Google ecosystem, but… Uh, there’s enough sort of related content that it’s really valuable.

01:20:07.000 –> 01:20:13.000
I also really like Kevin Indig. He has a newsletter called Growth Memo.

01:20:13.000 –> 01:20:25.000
Uh, I think is excellent. He does some really interesting deep dives into, um… Uh, into what’s going on in the AI space and, like, optimizing for AI overviews and AI search results.

01:20:25.000 –> 01:20:41.000
I really like his content. And then, uh, Rand Fishkin is… he’s sort of like the… father of SEO. He’s sort of outside of the SEO space at the moment. He operates a company called, um… SparkToro, uh, but he does, like.

01:20:41.000 –> 01:20:47.000
Occasional great, like, wonderful presentations about the state of search as a whole.

01:20:47.000 –> 01:20:51.000
Uh, and whenever they, uh, whenever they come out, they’re just like.

01:20:51.000 –> 01:20:53.000
I drop everything I do, and I sort of, like.

01:20:53.000 –> 01:20:58.000
take it all in, uh, because he does some really wonderful high-level analyses.

01:20:58.000 –> 01:21:02.000
Uh, but Emily, I don’t know if you have other ones.

01:21:02.000 –> 01:21:09.000
Um, those are all great, I mean, basic, I’m sure everyone knows Search Engine Land, I find super helpful. They’re always giving some type of update.

01:21:09.000 –> 01:21:20.000
Um, as well, I think just following Google, uh, in general’s news, which may be an obvious one, but, like, listening to their summit about what they’re rolling out, and then they’re always dropping new announcements and such.

01:21:20.000 –> 01:21:28.000
gives you a real indication into what they’re focusing on, which I find helpful even if they’re not particularly transparent about how everything is happening.

01:21:28.000 –> 01:21:35.000
Um, those are, those are two, and then, um, actually shout out to Mike East Centrela on our.

01:21:35.000 –> 01:21:40.000
AI and product innovation team, but he got us onto this newsletter called The Neuron.

01:21:40.000 –> 01:21:49.000
Um, which gives a lot of AI updates, not just, um, search updates, but that’s great just to keep tabs on, um, because it… Gives a lot of insight.

01:21:49.000 –> 01:22:01.000
Yeah, the neuron is great, and actually, interestingly, they were at Adobe Max. They have typically been in the AI space, but acknowledging that AI is impacting creativity, they actually did a huge.

01:22:01.000 –> 01:22:09.000
report this year on, um, on all of the AI announcements coming out of Adobe Mac, so I thought that was pretty good. So, yeah, the Neuron Daily has been on my.

01:22:09.000 –> 01:22:24.000
reading list for a very long time. Um, so, yeah, um… Well, great, uh, great suggestions. Thank you for those. Um, I’m gonna wrap it up, I guess. Uh, thanks for everyone who stuck with us to the end. I think this is the longest one we’ve.

01:22:24.000 –> 01:22:30.000
done in many, many months. So, um, but again, not surprising. Um, thank you, Richard.

01:22:30.000 –> 01:22:36.000
Thank you, Emily. Um, we’ll see you again in 2026. Thanks again to David and Amber.

01:22:36.000 –> 01:22:50.000
Uh, for the… Playing in the sandbox with me, thanks to NETA, and uh… Everybody have a safe, relaxing holiday season. Have a very happy, joyous new year, and uh.

01:22:50.000 –> 01:22:59.000
We will see you all in 2026.