AI vs Human Content, Missing Google Reviews & the Evolution of Q&A | E16 The Whitespark Local Update

We’re wrapping up the year with a special episode of the Whitespark Local Update! Darren is joined by the one and only Mike Blumenthal from Near Media. In this episode, they discuss:

📝 Google Removing Q&A Feature In Google Maps For AI-Powered Ask Button (Barry Schwartz)
📝 A Guide to Web Guide: Our Hybrid Search Future (Dr. Peter J. Meyers)
📝 AI-generated VS Human-written SEO Content (Andrew van Rossenberg)
📝 People Prefer AI-written Content (Kevin Indig)
📝 The Global Wave of Deleted Google Reviews: What Our Data Reveals About AI, Law, and Trust (Michel van Luijtelaar)
📝 Complaints About Review Takedowns Are Rising (Mike Blumenthal)
📝 AI Mode and Agentic Local Results (Mike Blumenthal)
📝 Future of the SERP Is Dynamic (Near Media)
⭐Near Media Podcast

Creators and Guests

Guest

Mike (Michael) Blumenthal

Mike Blumenthal is the co-founder and analyst at Near Media where he researches and reports on reputation, reviews and local search.

What is The Whitespark Local Update?

The Whitespark Local Update is the go-to podcast for Local SEOs and Marketers who want to stay ahead of the curve in local search and the local visibility space.

Join industry experts Claire Carlile and Darren Shaw for a lively, insightful roundup of their carefully curated selection of top “must-read” and “must-watch” links, including news, trends, and can't-miss resources.

Darren (00:00.046)
Welcome to another episode of the White Whitespark Local Update. And I have a very special guest with me today. The infamous Mike Blumenthal.

Infamous, no less. I will take that.

Definitely infamous. Mike Blumenthal is my guiding light in local search. Mike is absolutely the, I'm going to just say it, the smartest guy in local search. He has been a major mentor to me. He has been the guy that I've always been looking up to. So it's just a pleasure to be able to record a podcast with you today, Mike. Welcome.

Thank you for having me. I'm tempted to call this last week in local. It's just old habits time. I would say, having worked with David Mihm Near Media, I think he's smarter than me. Just saying.

Exactly.

Darren (00:52.046)
Yeah, you know and actually let's talk about Near Media so Mike it works with Near Media David Mihm and Greg Sterling at Near Media They are so if you like podcasts like this one There is one other podcast you must listen to and it is the Near Media podcast that is the Really the key podcast that I subscribe to and listen to all the time it's where I get all of my information about what's happening in local search because those three guys at Near Media are

just the very best. so and Mike, you guys have some kind of marketing program at Near Media. Explain that.

So we have been doing consumer behavior search research. So we watch consumers on Google. We put together panels of people. give them a very broad task. then we watch what Google delivers and how they consumers interact with Google, how Google responds to consumers. the idea is that we're working a lot with law firms and hope and we uncover

how in uncovering how real clients search and evaluate and choose a lawyer, we are able to develop insights to improve their, dramatically improve their conversions because we see what happens before the click. Google wants to tell you what happens after the click. They want to take credit for the attribution, all the other stuff. We want to know what goes on, what Google shows. our data just gives us really clear understanding of which messages resonate.

which touch points matter most, which barriers stand in the way of gaining new clients. And we turn complex consumer behavior into strategic and tactical guidance that helps BI lawyers and other lawyers attract and win more of right cases.

Darren (02:40.962)
I think it's really fascinating and I don't think anyone else is doing this. So they're like, you'll get like 200 people to run queries and then you watch their videos and see what they do. And they narrate these videos and then you get to see how are they making decisions? Are they clicking on a local pack? Are they clicking on the ads? How far do they scroll down the page? How much time do they spend reading reviews? And then all those insights are just really powerful to, cause you might be thinking the wrong thing.

yeah, this local pack is really important, but you find like 70 % of people are kicking the ads and I don't know, I'm making stuff up here, but it's just really fascinating to understand how real consumers are using the search results. And the thing that's crazy about it is that the search results are in so much flux, they're changing drastically all the time, the different interface on mobile, different interface on desktop. So the insights that you guys are providing in that program is really fascinating. So anyone listening is.

worth checking out. Go check out Near Media. Take a look at the program that they've got running there. It's good stuff. Are you ready to talk about some links? Okay, that's the point of the podcast. and actually, just want one more you mentioned last week in local. We are definitely tried to model this podcast after Mike's famous last weekend local podcast that he ran with local you for how many years did you do it, Mike?

I am.

Mike (04:02.99)
Probably close to 10, I don't know exactly. Yeah.

Yeah, and it was sad when it ended. And so I've always been wanting to do something similar. And so this is basically modeled after last week in local. And so thank you, Mike, for blazing the trail for this podcast. So my first link is Barry Schwartz on Searching for the Land Table. People have been talking about this because Google.

is basically getting rid of the Q &A feature. actually, Mike Blumenthal on Near Media was the one to kind of break this story back in January, where he talked about Ask Maps. So Google has this new AI-driven Ask Maps feature. So rather than looking through the Q &A, you can just ask their AI on a specific listing, hey, when are they open? What is the parking like available here? And so Google's leaning into this new.

methodology for answering questions. And so they recently published a new communication about what's happening with Q &A. And Barry covered it on SEO Roundtable. That was in his videos and stuff, too. But basically, as far as I could tell, and I would love to get your thoughts on this, Mike, this is what I think is happening based on the recent communication. Q &A on the front end is going to go away. You won't have that Q &A section on the front end. And so that's going to be replaced with the

you know, ask maps question or ask a question about this business. Then when you ask that question, it looks like the you know, the AI will answer based on what it knows about the business from their website, from the Google business profile, it'll from the broader web, everything it knows about the business will be answered via AI. As far as I could tell from this recent communication, there is a nuance that is maybe happening now, which is they're going to keep the Q &A section in the back end and potentially

Darren (05:53.612)
I'm trying to read between the lines here. Google might prompt you and be like, there is a question that people keep asking that we don't know the answer to. And then it'll ask you to answer that question. And now AI will have that information, can answer it for future customers. So I feel like the Q &A section will stay in the NMX. That's just gut feeling. Mike, what do you think?

Yeah, I think Google always loves more data and I clearly Q and A had become unwieldy. It didn't work well in the front end interface and search. So moved over to just maps, but they people need answers. Certainly AI with a quick query is a better way to get it. But you're right. I think they were missing data and they're relying on the business to provide it. And they will do that as long as the data they get from the business is helpful. And as soon as it isn't though.

stop supporting that feature, right? It's like everything with Google is a test. This is one amongst many, I'm sure. And I think this is probably more valuable to businesses that have weak websites and don't have a lot of information on their websites, know, nor that kind of thing. But I don't see it as a strategic, playing much of a strategic role. Again, we don't know exactly how AI is going to roll out and where it's going to get all its data.

Yeah, so we'll see. We'll see. Right now, that's my vibe. I feel like they're going to keep it in the back end, and it'll change to, instead of it being your old questions, it'll be like, they'll just prompt you and be like, hey, people here are asking, can you, it's almost like the know this place. Like at the bottom, sometimes if you've been to a place, you'll get this old tag in the Google Maps that says know this place, and you're going to ask her questions about it. I suspect it's going to evolve into that on the back end for Google's, and you'll get an alert. Be like, hey, Google has a new question about your business. That's my suspicion.

Anyways, what kind of links you got to talk about, Mike?

Mike (07:41.902)
Well, there's been a long sort of conversation going on in the low community about an increase in review takedowns. Yeah. Many others said it was a bug. I never believe that when somebody says that about a change like this, it's always, it's much more likely to me that these takedowns are a result of a change in the underlying algorithm or an increasing rigor in it. And the volume goes up. So we've been noticing this for the last six months.

particularly hard hit. Claudia Tamina has been talking about it in restaurants. GMB API wrote a recent article analyzing nine months worth of review takedowns, which is very insightful. that link and then Google sent out a notice the other day noting that the number of review appeals has so skyrocketed. They're way behind the eight ball. So leave them alone. They're going to take a long time to get back to you.

review appeal. So clearly the takedowns have gone up. We see that there's a spike in the forum. I posted about this on LinkedIn where we're seeing forum posts largely reflect businesses getting upset about the AI taking down good reviews. So we see an increase in the forum when the algorithm has too many false positives. We saw a bump in November, big bump in, I think it was July. And so

It's an ongoing issue. The article at GMB API shows from the first of the year, a steady increase in review removals with a peak in July. They didn't cover November. and then what was interesting in the GMB API was just how it seems to be impacting in the restaurant industry, older reviews. So clearly Google is doing one of their review sweeps through that industry. And in,

construction seems to be home services, it's affecting newer reviews, which is part of their new algorithm where we know they target high spam categories, make it more expensive and harder for them to get reviews early on, that kind of thing.

Darren (09:56.568)
Do you think that they do these review sweeps where they kind of do a cleanup on a category basis?

I think they might. I don't know. We know they do sweeps. We know that they come through periodically, look at old things by new standards. So we assume that the AI is constantly being trained on new data. Exactly how those sweeps are executed, we don't know. But since we're seeing increases in the restaurant world and a few other categories, it is entirely possible that it's done by category. They need some logic to not have to burden the whole system.

with every business every time they go back and look at old reviews.

So what is the business owner to do if they notice that their review count dropped from 700 to 630? They 70 reviews got removed. Are they supposed to do anything about it?

If you're keeping a log of your reviews, you have a shot at getting some of them back. can go to the sports system and tell them that there's missing reviews, but Google is very, very reticent to restore reviews because they believe they're right in taking them down. is a false positive, but if you have the reviews and you have perhaps some additional evidence that they are real.

Mike (11:15.33)
You can file a claim through the support system. They no longer accept escalations to the forum. Unlike fake reviews, you can file in the support and then you can escalate the case to the forum and possibly get it taken care of. But in missing reviews, it's one and done. You basically file. Now, the other thing that can be done, if it's a one or two or three reviews, you can go back to the person and have them file.

with the Google Local Guide program for review of the review. So if it's a small quantity, you can make a request through sport and then go back to the user, ask them to make a request. If it's a big quantity, you have to have the reviews. You need to be able to submit them to Google and request a return. Chances are low that they'll come back because Google wouldn't have implemented the algorithm. They didn't think they were right. And I'm air quoting right in the video here for those of you who are listening.

All right. So then my next link is Dr. Pete on the Moz blog. He wrote a guide to Web Guide, our hybrid search future. So I was first alerted to Web Guide on the Near Media podcast. You guys invited me to speak about the local search ranking factors. I think it was David that said, I really think the future of search is going to be Web Guide. And I was like, at that time, it was not really in my mind. I didn't really know anything about Web Guide. But Dr. Pete has written a very excellent

breakdown of what Web Guide is, how the query fan-outs work. And he's got all these different types of fan-outs, which I think are really interesting. So in the Web Guide interface, it's kind of like this hybrid between AI results and traditional search results. They're all kind of coming together in what is called Web Guide. And I think he's right, and I think David's right, that this is probably what the future of an AI-generated Google interface looks like. So this is a must-read. You should look at this.

prepare for the future. And it also gives you some interesting guidance on the different types of fan-outs, which I thought was quite fascinating. And that you could use those fan-outs, those different types, to think about the different types of content that you want to have on your pages that are addressing these different areas. So it was a great article by Dr. Pete. Highly recommend you give it a

Mike (13:37.598)
To some extent, it reflects the search journey and micro moments that Cindy Crum speaks about. The, want to know, go, do, and buy. It reflects those. And I think it makes sense that Google is trying to better reflect. also reflects the people also ask type queries that historically seem that users from our research never click on. It sort of pre-populates those with AI information.

Exactly.

Mike (14:07.404)
but it's a fairly boring interface. It has drawbacks in local particularly. It's not a rich, it has drawbacks with video. So there are some things, but I think that Google is gonna be addressing those. I think this is a framework to replace 10 Blue Links. And into this, they're gonna figure out how to plug in other types of results.

Pretty beta.

Darren (14:33.0)
It's a good look at where we're headed, I think.

Yeah, I have it. It shows up for me almost all the time, almost every query now. Wow. Particularly even local brand queries. It drives me very quickly into AI mode because I don't like the results.

often. It's almost like the word guide is very appropriate because like you put in a query and then it tries to predict what you're going to want to know and where you're headed with that query and then guides you with the sort of the fan out. It's really interesting new way of search this conversational search. So what's your next link Mike?

It's over to

Mike (15:07.426)
My next link is an article at our Near Media newsletter, which has now become free, by the way, you just have to sign up for it. But Greg wrote a very good sort of speculation on what the future of search interface really does look like. I mean, over the past three or four years, the search interface has gotten very confusing, messy, complicated, lot of cognitive load to drive through it. Webguide is one way to simplify it.

in a meaningful way. But like I said before, it doesn't include certain useful results like information from the local knowledge graph, or many reviews or that type of thing. So his article speculated and I think I agree with this, speculating about the future is easy anyway, can do it. opinions are like apples, everybody's got one and they all stink, right. But that being said, that being said,

You know, Google has several problems. One is how and where do they integrate ads? How do they reduce cognitive load on consumers to increase the number of searches and how do they give consumers the information they want? And I think web guide is the best thing I've seen to do that. But there are some, and we're going to talk about this later, but there are some things in AI mode that are very powerful, particularly vis-a-vis local that I think.

But AI mode is more expensive for Google to run. So don't think it's going to be the default search environment, but I think there are going to be moments when AI mode integrated into Web Guide makes a lot of sense. think there may be other ways of integrating videos and stuff into it. So, and Greg made this point that it isn't AI mode or Web Guide or what we have now. It's some, likely some amalgamation of that, that

allows them to deliver ads profitably and then solves these other problems, hopefully with a better user experience. Cause the user experience from my point of view on Google had gotten really shitty and I shifted use off Google towards chat GBT, but AI mode and Gemini I think are significantly better. And particularly in local AI mode is I really like.

Darren (17:19.736)
Yeah, that reminded me actually, I remember like pre-ChatGPT, just a lot of complaints about how Google's results, like just every, all the SEO is complaining that Google's results are garbage. now ChatGPT came out, Google's now fighting for, you know, they finally have someone that could take some search volume from them. And so now they're, they're actually really improving things with these new interfaces.

With the announcement of Gemini in 3 last week, they talked about their ability to craft unique user interfaces for different tasks. So when you take the probabilistic nature of AI and the amalgam of all of these, the increased knowledge of user histories and search patterns, their ability to customize interfaces on the fly, the increased understanding of user intent through

long form questions and you mix that all together. What you see is Google's dream. Every user gets a different search results. So rank tracking is out the window, right? Everybody gets a unique experience that isn't reproducible, which causes tracking, you know, from the SEO point of view causes issues of how do I know if I'm doing well?

Mike (18:44.487)
Yeah, I think it's even going to become weaker as of right.

Yeah, yeah, it's still useful to get a sense of like okay is this stuff that I'm doing having some kind of impact but we'll see yeah and actually your post on this topic I really loved the screen share that you that you Recorded or you recorded yourself doing this on mobile and it was very interesting to see how the results locally evolved with your multiple follow-up queries and that I think was a it's such a great example of Had the future of how?

People are going to search conversationally. It's actually really interesting with the research you do at Near Media, right? Like you can start doing this kind of research and watching what kind of prompts people are typing and what kind of follow up prompts are they typing.

this kind of conversation and the ability to deliver results does several things. One, it breaks down the rigid, brittle, three mile radius or whatever it is of local right now. I live rural, so local results for me are meaningless. I need results in Buffalo, Rochester, Toronto. And so with this new inter conversational interface, you can say like I did in that search on

Yeah, exactly.

Mike (19:56.376)
LinkedIn, can say, give me the results for the greater Buffalo area and the suburbs for all of the best custom engagement ring designers. And it will do a query fan out, show you local results for Buffalo, local results for the suburbs. And then there's a little checkbox next to them. So you can then say, gee, compare these three. And then you can say, which one has the best reviews overall around the web?

That was really cool, yeah.

Mike (20:25.93)
And Google will then do a query fan out, find all the results around the web. So it can open up all new ways of, discovery in local, much broader than current local algorithm allows, which I'm pretty excited about. then finally, if you do a restaurant search, you can then, it will actually tell you, it'll go out and look at their reservation modules and we'll bring in what hours this particular restaurant has available for seating.

and allow you to book it right there. So it becomes agentically transactional on top of more flexible, better able to meet intent, more intuitive conversation, all of these things. I AI mode with local strikes out into a much friendlier usable discovery environment.

I was thinking about that in terms of, so yeah, it's definitely finding the bookings and what times are available. It was getting his agentic SEO stuff. And then the same thing for product searches, like who has this product? What is the cost? It's in stock. Can you reserve it right now? Your agents can do all that stuff for you. And I was thinking about this from the service business perspective. Like, I could see a future world where service businesses are smart to put in some kind of booking calendar onto their website.

that then Google and other agents can then see what your availability is. Do you want to book a consultation on Friday at 3 PM? They have that available. It's just automatically. So I feel like some of these models for booking reservations and purchasing products online will come to the world of services because we need it in an agentic world.

Which brings up a point that was mentioned on a podcast several weeks ago by Jess Schultz at your memo. The memo where she said you can use these agents now to test whether they're able to navigate your site to accomplish that task so that you know you are ready when agents do ultimately become more broadly available. Which I thought was a great idea. Using AI agents to user test your website to see if you can accomplish, can a user easily

Mike (22:35.63)
call me, can you easily contact me, whatever can use user easily make an appointment. I think it's a great use of agent to AI today because it's not really mature.

So to wrap it up, have one more combo topic. So this guy, Andrew Van Rosenberg, posted on LinkedIn that they did a very interesting test of handwritten content versus AI written content. And the results, very dramatic, were shocking. So the human written posts brought in 722 clicks and 300,000 impressions. The AI generated posts?

nine clicks and 9,000 impressions. So this is very interesting when you think about what is going to be successful content. A lot of people are just generating content with AI. And so you might be thinking in your head, OK, this is a problem. Our website, we've got too much AI-generated content. We're never going to perform on that. So I wanted to highlight this because it's interesting. But I have another article that I

noticed like honestly the next day and so my wheels were turning like AI generated content sucks. And then the next day Kevin Indig on LinkedIn had posted this and it's like the balanced perspective. And so he wrote on LinkedIn, people prefer AI written content in direct contrast with what I just said. So I thought that was interesting. And he just goes on to speak about how the thing is, is if you just go and put in a default prompt,

and say, hey, write me a blog post on this topic. And then chat GPT gives you the blog post and you copy and paste it on your thing. That's obviously going to be a major failure. But what he talks about is that the key is prompting. If you use just standard prompting, it's predictable. You're to have this garbage output and the human content will win. The results with fine tuning, that's where you really actually get benefits of using AI to generate content. So he talks about this. He said,

Darren (24:44.056)
We need to continue fine tuning the AI. So you give it your tone, you give it your style, you train it on how you speak and how you want the content to be. Then you also can train it on client interviews, pain points extracted from reviews. Now you're generating really high quality information gain content.

but you're using the AI to actually just put the whole piece together at the end. That's a whole different ball game than just garbage AI content. the AI, that process with AI can actually help you accelerate content development and output really good stuff. And he closes his post with, we need to stop judging AI based on generic out of the box prompts. That's like judging a Ferrari by only driving it in first gear. I think it's a really great analogy. And so...

Yeah, AI content. What do think, Mike?

I absolutely agree. know, I've, when I see an article in the newspaper that I miss suspicious of, I don't understand the context or they haven't set the context. What I will do is I will go to AI and say, you know, who wrote this? What, where did it come from? What is the, you know, what is the ideologic bias of it? what are all, is there research that reinforces this or counterpoint? I will then.

You know, summarize what I learned and then I'll say, okay, write me an op ed piece to contest this news article, this bullsh** news article with, you know, what we've learned. Right. And then it will draw on all of the contexts that I've created and produce. then like in, in the same with content, I can add a tone, a political point of view.

Mike (26:41.494)
I can say, gee, I'm a democratic socialist or a social Democrat and tailor your, and it will then take all the facts that we've sussed out, narrow them down and put them in contrast to the original article. And it creates either great outline or a very good op-ed piece. So I agree totally that you can't take this stuff at face value. You have to dig and dig from broadly to very narrow.

come down. In fact, I thought about this in terms of review responses. What we see in our research in your Media, in Europe, 50 % of the people read reviews, in the United States, roughly 25%. But even that 25 % is one of the biggest contributors to conversions. It's bigger than almost any other thing other than reviews themselves. But one of the goals of review responses, given that

Is that you want people to not just see that you responded, but to read your responses. Yeah. So the generic AI response is not going to do that. And so I think there's an opportunity in review responses using AI to create a clever, caring, compassionate, meaningful response that might actually get read. And if it gets read, then all sorts of other good things happen. Time on your listing improves.

conversion rates improve, their perception of your business improves. So I think this applies at a big level like your website, but also at a very little level, like not relying on the standard AI answer response to review that I think you can use AI to really get creative with the review responses and take it to the next level.

Yeah, that's interesting. Actually, it's like any snippet of text doesn't matter if you're writing a whole article or you're writing a social Media post or you're writing a review response or a description for your YouTube video. When you go with the default AI, it's generally trash. If you just say, give me a description for this video, give me a response to this review. It's generally trash. You have to do it. And I actually see it all the time. I was talking about this with my marketing team this morning as we were

Darren (28:58.624)
writing the social Media post to promote Miriam's new guide that we just published. And it's like, I will use AI a little bit. I'd be like, okay, here's my draft of a hook. Give me like five variations. And on this call with my marketing team, I'm looking at these five and we hate them all. They're all like this dramatic AI, they just have this feel, this AI feeling to them. And they're like, ugh, gross. They're all.

We want to throw them all out because this is AI garbage. And so we ended up just like giving up and being like, well, we can maybe pull like those, kind of concept and you have to write it naturally and, and, and human. And so I don't, maybe I could train the AI on being like, don't be gross, be more like the way.

I think what you do is you pull those pieces, which is what I do sometimes with your memos, I'll pull those gross pieces out, take what's left, craft something that's more or less what I'm looking for, crank that back in and say, give me variations on this. And you may hit a better variation, you may not.

In summary, AI content out of the box sucks. Everyone hates it. Don't do it. But AI is a fantastic tool to help you with your writing. All right, think we did it, Mike. We did a whole podcast.

We didn't do it in 20 minutes. That's all right. I was going to caution you. I'm never a 20 minute guy.

Darren (30:18.584)
Thank you, Mike. Thanks for joining us today. You can definitely follow Mike Blumenthal on all the social Media places. Probably not X. You're still on X.

You're still in X. you quit X because he's an a**hole, you'd have to quit Microsoft. You'd have to quit Chet GPT. You'd have to quit Google. You'd have to quit Walmart. You'd have to quit Target. You'd have to quit Amazon. What's left? The local diner where they're also a**holes.

the whole

Darren (30:51.502)
No, those people at a local diner

It's like, it's hard to.

I do I do know what you're saying. So anyways follow my bloom with all and all the socials run by Yeah And definitely check them out at Near Media if you do not subscribe to the Near Media newsletter if you do not listen to their Media podcast You are missing out. Those are if you're listening to this podcast, you're gonna love that stuff

Thanks us.

Mike (31:19.04)
It's called Near memo, by the way, if you're searching for it on.

The podcast is Near memo. Thank you. All right. Thank you, Mike.

Thanks for having me. I appreciate it. Anytime I will talk longer than 20 minutes. Just give me a platform.

Sounds good. Okay. Bye everybody.