ChatGPT's New Business Profiles, Google's Silent Review Bans & the War on Scrapers | EP 283

Greg Sterling and Mike Blumenthal test ChatGPT's new business profiles against Google and Apple, dissect an HVAC review ban Google never announced, and trace how Google is gating its results against scrapers. Plus: why no one knows about Ask Maps.

ChatGPT's New Business Profiles, Google's Silent Review Bans & the War on Scrapers | EP 283

Clicking the video loads it from YouTube, which sends data to Google under its privacy policy. Prefer not to? Watch on YouTube.

The player is hosted by Buzzsprout. Loading it sends data such as your IP address to Buzzsprout under its privacy policy. Or listen on Buzzsprout.

Greg Sterling and Mike Blumenthal go hosts-only for a four-part episode. Mike screen-shares a live tour of OpenAI's Businesses in ChatGPT dashboard next to Google and Apple insights for the same jeweler. After that he walks through a two-location HVAC client that Google quietly blocked from getting reviews. The last two segments cover Google's escalating campaign against scrapers and the Ask Maps holiday features almost nobody has heard of.

The platforms are fighting over first-party business data. OpenAI is asking businesses to hand it over, and Google is walling its own results off from everyone else, so local businesses and the tools that serve them end up with thinner analytics, opaque penalties and a measurement stack that has to move toward outcomes.

Takeaways

  • ChatGPT's business profile is easy to claim and hard to scale. OpenAI launched Businesses in ChatGPT on November 10. A business verifies by phone, text or a DNS TXT record and then passes a Persona identity check to get the blue check. Mike calls a single location trivial, but the friction grows with every location; a 50-location brand is "very difficult." He also flags the government ID and face scan as a poor fit for anyone doing political work. (01:28–04:44)
  • The analytics are a thin funnel with 7- and 28-day windows. For Barbara Oliver Jewelry in Buffalo, ChatGPT reported 46 mentions, five profile views and six website visits over 28 days. Discovery topics are broad buckets like "local service providers," not prompts or keywords, so Mike finds nothing actionable. (04:44–06:55) [1]
  • Apple out-delivers ChatGPT for the same business. Over a comparable month, Apple Business showed 65 taps from local search, about ten times ChatGPT's figure. It also broke out brand and category searches, a map of where searchers were, and nine phone calls. Mike's estimate puts Apple at roughly a quarter of Google's local volume, which makes ChatGPT's current referral value small. (07:55–10:05)
  • The real story may be that Apple is a local player. ChatGPT's limited numbers end up showing how much local activity Apple already captures. Greg expects Apple's share to grow as Siri gets better at flexible local queries and Apple moves toward full-funnel search in a single interface. (10:57–11:44)
  • Claim it anyway, especially single-location businesses. OpenAI promises no ranking boost, but the check mark may lend credibility. The Business Agent chat, drawn from the business's own site and inputs, could convert questions like "is this in stock?" Greg sees the Shopify catalog import as the strongest use case, because product search in ChatGPT already outpaces local. (12:08–13:21, 17:11–17:30)
  • Agentic chat is a workflow problem before it is a feature. Google is rolling out its own AI agent in Business Profile messaging that hands customers off when it can't answer. Mike's worry is a business juggling handoffs from Google, ChatGPT and others with no central inbox. Google also has a history of killing communication products businesses built around. (13:21–14:43)
  • The dashboard is a sales funnel and a data pipeline. Greg reads the "GSC Lite" dashboard as an upsell into ChatGPT ads and business subscriptions. He also sees a direct feed of business data as a hedge against Google's anti-scraping moves. Mike adds that the local battlefield "has a lot of bodies from vendors," and adoption will need agencies as a channel, since only a minority of businesses actively manage even their Google profiles. (15:31–17:11)
  • Google review bans switch on automatically and need a human to lift them. A two-location HVAC client (separate cities, two hours apart) lost the ability to get reviews at one location from late September. Google never told them. Support confirmed the restriction only when asked, and removed it only after a second request and an internal escalation. It ended roughly five weeks after it began. (18:25–19:35, 21:59)
  • A ten-day spike plus repeat reviewers plus employee names lined up with the ban. With the unbanned location as a control, Mike tested which review-velocity window matched the ban. A spike against the previous ten days fit best, but that spike was no bigger than earlier seasonal ones. What changed was that customers reviewing both locations and reviews naming employees rose in the same window. Mike stresses this is correlation, not proof. (20:14–21:44)
  • The fix: back off asking, link to the brand, monitor volume. Mike's client paused review requests during the ban and sent the brand URL instead of the review link. Mike's advice for seasonal home-service businesses: watch for a sudden drop in review volume, confirm with Google, and lift the ban after volume eases. Stop prompting customers to name employees, which Greg calls a proxy for incentivized reviews. (21:44–21:59, 25:26–27:25)
  • Google bans reviews for other reasons too. Mike lists geofenced bans around the overturning of Dobbs, wartime Israel, the Olympics and the World Cup, plus bans on businesses being spammed by competitors. Greg, drawing on expert witness work in a review fraud case, says review spikes at odd times of year are likely the first signal that draws Google's attention. (17:36–18:25, 22:28, 24:52–25:26)
  • Google's anti-scraping escalation, step by step. It started with JavaScript in January 2025 and the removal of the num=100 parameter in September 2025. Then came the DMCA suit against SerpApi in December 2025, dismissed twice. After that: google.com/goto redirects from July and poisoned results served to bots. The latest is a sign-in prompt after a single search. (29:55–32:24) [2]
  • A login wall turns "no scraping" into an enforceable contract. Mike's point: copyright and hacking theories failed, but a sign-in gate binds the user to Google's terms, which forbid scraping. Greg calls it hypocritical, since Google built local search on the principle that facts can't be copyrighted, starting with listings keyed in from phone books. Both think it will hit small SaaS tools, ranking grids and agencies hardest, and land as a shot across the bow of ChatGPT, Perplexity and Claude. (32:24–36:36) [3]
  • Rank tracking's days as a client metric are numbered. Mike says agencies that still report rankings to end clients should stop. He would move to conversion metrics such as calls, door traffic and tracked outcomes, and accept that the middle of the funnel will get murkier. Greg expects a login wall to be a non-starter in Europe, and Mike thinks it would inflame Google's US antitrust cases. (36:52–38:48)
  • Ask Maps is Google's best AI product and nobody knows it exists. Google's November 16 post added itineraries, route tips, appointment finding and saved-list search. Greg uses Ask Maps more than any Google product outside Gmail and Docs. Mike polled about 16 family members and friends, many of them heavy Google Maps users in their thirties, and not one knew the feature. He wants it to be the default Maps search. (38:48–41:49)
  • Google markets on TikTok, but not Ask Maps. Greg sees Google TikTok ads aimed at young women about holiday shopping and asks why Ask Maps, with real utility and a public souring on AI, isn't the campaign. Mike's answer is culture: Google launches early and buggy, waits for viral uptake, and takes products down if they don't catch on. He calls it a Darwinian approach that businesses and consumers hate. (42:24–44:22)

Practitioner Notes

  • Add review-volume monitoring to every home-services and PI account. Mike's case shows a ban can run for weeks with no notice. Include a weekly review-count check per location in GBP audits; a sudden flat line is the trigger to open a support case.
  • Strip employee-name prompts from review-request templates. Technician-specific links and "mention who helped you" asks create the exact proxy Mike found in the ban window. For multi-location clients, route repeat customers to the location that actually served them.
  • Claim ChatGPT profiles for single-location clients now; hold multi-location rollouts. The Business Hub can't handle many locations yet. Use the DNS TXT route where you control hosting, and log the 28-day mentions/visits baseline for comparison against GBP and Apple Business.
  • Rebuild client reporting around calls and conversions. With /goto redirects, poisoned SERPs and a possible login wall, grid and rank data is getting less reliable. Call tracking and GBP actions are already the core of our PI and GBP audit work and should lead client reports.

Pick your starting point:

00:00 Intro
01:28 OpenAI launches Businesses in ChatGPT
02:31 Claiming a profile: easy for one, hard for fifty
04:44 ChatGPT's analytics for a real business
06:55 The same business on Google and Apple
10:57 Apple is a bigger local player than you think
12:08 Should you claim your ChatGPT profile?
13:22 Google's AI chat agent and the workflow problem
17:30 Google's silent review bans
18:25 Case study: two HVAC locations, one ban
20:14 What lined up with the ban
25:57 How to avoid and lift a review ban
28:37 Google's war on scrapers
32:24 A login wall as a contract
36:52 Rank tracking's days are numbered
38:48 Ask Maps: Google's best AI product
42:24 Why isn't Google marketing it?

Full Transcript -->
The Near Memo, EP 283 — Transcript

Editor's note: This transcript follows the final edit of the episode, and timestamps match the published video and audio (including the 10-second intro). It has been lightly cleaned for reading. Fillers, stutters and crosstalk backchannel are removed, and clear transcription errors are fixed. Speaker-side errors are left as said and listed at the end.

Greg Sterling [00:10]
Hey everybody, welcome back to the Near Memo with me, Greg Sterling, and Mike Blumenthal, as always. And today we don't have a guest. We're just going to rant and rave as we often do here. We'll be talking today about the new ChatGPT business profile and analytics tools that were just introduced, what I'm calling ChatGBP. And yes.

Mike Blumenthal [00:32]
Just to confuse everybody. Particularly our transcription software.

Greg Sterling [00:39]
Well, gotta do something to keep life interesting. Then we're gonna be talking about how to review bans on Google, what might be triggering them and kind of how to prevent them. If that can be done. You it can be done by getting no reviews, but if you've got a regular review program, how can you kind of minimize the likelihood that you're gonna get a review ban? We will also be talking about Google's new strategy of blocking scrapers by maybe compelling people to log in and how that might impact the tools ecosystem and search in general. And then finally we'll be talking about Ask Maps It's a great product that keeps getting new features, but is anybody actually using it? And what does the future look like? So Mike, well, I'll just I'll jump in because you're gonna show the screens. So about a week ago, I think at this point, OpenAI introduced two things which are tied together: a business profile tool, and then a kind of light lite analytics, which is sort of a very thin version of Google Search Console. And The idea is that you claim your business, similar to GBP, you verify your control over the business, and then there's a personal identity verification step. You can add some information about your business, and then this is going to give you a blue check mark. There's no cost involved at this point. And it and there's also the option to create a branded business agent, which will allow users in ChatGPT to ask questions, to chat with your business. And that information comes from stuff that you supply and also your website. And then there's some other features here. So I'll let Mike give us a tour of this. So this is the live UI. Yep.

Mike Blumenthal [02:29]
Well, firstly, right. The claiming of a business is fairly trivial. Particularly if you have domain level access. You go to a list. There is some friction. Right. And it's and

Greg Sterling [02:40]
There is some friction there, but it's pretty straightforward. Yeah.

Mike Blumenthal [02:45]
The friction accelerates. I've talked to a number people who are have multi locations. The friction gets harder the more locations you have. Right.

Greg Sterling [02:52]
Yeah, it's not it's not set up for multi locations right now.

Mike Blumenthal [02:55]
It's not set up for multi locations. If you have a fifty location business, it's very difficult. So they give you this pitch, show customers what you offer, see how businesses show up, help customers make decisions. Clearly this is meant both as a well, as a funnel, I think, for ChatGPT both to sell advertising but also to sell their upcoming agent tools. You confirms my email address and the business that I'm verifying. And then you can have it give you a phone call or a text, or you can do the DNS, which is what I do because obviously the businesses I'm associated with, I control hosting, which might be a little complicated for typical business. The business phone

Greg Sterling [03:42]
It's it's definitely that step if you choose to go that route is complicated for an ordinary small business.

Mike Blumenthal [03:49]
Right. And it may take a little bit of time to propagate through the DNS system, which is the internet s location system and so

Greg Sterling [03:55]
I it took it took me yesterday, I did it and I have I had a domain with GoDaddy and I did it in GoDaddy and it was literally about a minute and a half before it went through for me. But other people have complained about longer wait times.

Mike Blumenthal [04:09]
And this obviously retrieves the business phone number from some place doesn't l and that is a legitimate landline, but it obviously can't receive a text at that landline. So there are other complicating factors here, but it's not really complicated. Although if you do verify your existence, you know, with government ID, you've got to do facial recognition, put your face in a database. So if you're involved in political work, probably not a wise move. But

Greg Sterling [04:35]
Well, if you've ever if you've been on a plane these days, you're probably in a facial recognition date database.

Mike Blumenthal [04:40]
Probably, although not required. And then you get a very lightweight analytics that does only two date frames, seven days and twenty eight days. Right.

Greg Sterling [04:51]
Strangely, why

Mike Blumenthal [04:54]
Very inflexible, very limited. And obviously some calls to action for upgrades. This is Barbara Oliver jewelry, which is one of Buffalo's leading sources. Buffalo's a market of maybe three hundred thousand, four three hundred, four hundred thousand, including the whole metro area. So it's not tiny, but it's not huge. It's not it's a working class city, so not a lot of wealth, but you can see over the last twenty-eight days her profile is seen five times. She got mentioned forty-six times and Theoretically what

Greg Sterling [05:30]
So that's so that's interesting. What it so the forty six is dimensions and five is what clicks?

Mike Blumenthal [05:35]
Profile views or actually clicks within to see the profile in more detail, and then six website

Greg Sterling [05:40]
Okay. And then

Mike Blumenthal [05:42]
Visits, which I have not validated via analytics.

Greg Sterling [05:48]
So there's a there's an interesting this is just an anecdotal example, but there's an interesting ratio there that seems to support the idea of the, you know, people seeing or interacting with the business content but not clicking through. Which is

Mike Blumenthal [06:02]
Well, although click throughs are greater than the actual profile views, right? So there's

Greg Sterling [06:06]
Right, but it's like it's eight to one or something or whatever the ratio is there. Not quite.

Mike Blumenthal [06:10]
Right. So it's a funnel. Obviously the discovery topics, it's a ver maybe this is just a funk and they don't get down into phrasing, they get to in broad topic areas.

Greg Sterling [06:22]
Right. So it's not c it's not prompts, it's not keywords, it's just sort of cons little categories, right? Essentially.

Mike Blumenthal [06:29]
Right, loc local service providers and comparisons, products are just let me just blow this up so people can see it.

Greg Sterling [06:34]
Yeah, that's not that's that's not particularly helpful.

Mike Blumenthal [06:38]
Not particularly helpful, very broad. So it's not actionable, I guess, is the thing. So the problems with this are limited functionality and not very actionable other than perhaps buying into their ecosystem. When you compare so this was tw last twenty-eight days. Google Insights typically, as is typical. Is a little bit quirky right now. So I can't even I don't even think I can pull up, just October say it's not giving me any results here. So I don't know what that's about. So let me just go back and bring up the so this is June, July, August, September. This is five and a half months or something. But you know, 1400 interactions when you go into engagement, reviews, you get search trends, you get s you get specific phrasing, you get to see, you know, audience type. So it's much richer and obviously many, many more. And Apple too is even Apple is significantly more than let me just log in real quickly here. Passkeys are a wonderful login system, particularly with a digit. But so you go into the account. One of the things that's interesting about Apple's interface, in this high-level view of the account, they don't make it obvious where you find the business information, the location information, right? Business at a glance. It's not clear. Set up a it's not clear what that is, right? If you click up on brands, You can then come in and find Barbara's location information and then click in further to go to insights. So clearly these insights, this is a week. Let's see, let's go to a month so we get a little bit better comparison. So taps into the local search results 65, which was almost 10 times what we were seeing on ChatGPT from them. And what people searched on. Before, you know, Nate, how many of these were brand searches, category searches, and other types of searches? You get a map showing where these people were

Greg Sterling [09:04]
The user location.

Mike Blumenthal [09:05]
User location, where they searched from. You see how you did comparison to the competition. You see how many times

Greg Sterling [09:12]
Think yeah. Much more data.

Mike Blumenthal [09:16]
Your profile got looked at. So the data is both more detailed and richer. And obviously there's many, many more interactions on Apple than there is on ChatGPT. And Apple is probably a fourth of what Google is. So we're talking about ChatGPT at least at this point not providing a great deal of benefit. Although Apple over the month you know had 10 website visits versus ChatGPT six. But people got driving directions, they looked at photos, they made nine phone calls, so other interactions went on beyond those six.

Greg Sterling [09:57]
So you have to believe that ChatGPT is looking at these competitive platforms and has a whiteboard with a list of features that they're gonna be building out or introducing over time. Because I think the current analytics that they're showing, which are just very superficial, are gonna frustrate people. I mean, one of the reasons, one of the motivations at least for agencies or SEOs to add businesses to ChatGPT is excuse me to get access to that data. I don't know how it compares to the ad data. I haven't run an ad, so I don't I haven't seen that data, you know, ChatGPT gives you some ad related analytics and I would imagine that has more detail, but I don't know exactly what they're showing there. But you've gotta think that they're gonna improve it improve the improve it over time.

Mike Blumenthal [10:55]
Yeah, one thing it does though is it just points out that Apple is a player in local. By how limited both the functionality and the quantity of referrals is, it points out that Apple is a player in

Greg Sterling [11:11]
Well the r the most of Apple's local traffic has historically been about specific businesses, brand search and driving directions or sometimes categories. But the more that they bring, you know, the more that Siri improves and the more that they make it possible for people to do f flexible local searches, we'll talk about Ask Maps later, the more that The more discovery that will happen in Apple.

Mike Blumenthal [11:40]
Right. And it right. And as Apple moves towards full fun funnel processes in a single search interface, which they're doing. So I'm gonna stop sharing this.

Greg Sterling [11:48]
Right. Well s

Mike Blumenthal [11:49]
Oop, good.

Greg Sterling [11:50]
Yeah, so just a couple of things about this. You know, it's interesting that they introduced it. This is something that we've talked about in the past that they needed to do something like this to get better data, to really become more competitive and local. I think it remains to be seen how widely adopted it will be, but people should do it, on general principles.

Mike Blumenthal [12:12]
Single location businesses should do it. Yes. Yeah.

Greg Sterling [12:13]
Yeah. Well eventually they'll do multi-location and they'll build out agency tools and things like that if they're serious about this. Businesses should do it. ChatGPT or A OpenAI says that there's no guarantee. The language is funny. The implication is that if you do this, you're not gonna get any greater visibility or there isn't gonna be some ranking boost if I can use that language. But the check mark might confer some additional credibility. And the chat feature, if you build it, if you create the agent, which is apparently fairly easy to do, most people don't have access yet, but it draws a lot of information from your website. Then the engagement opportunity within ChatGPT is interesting. Like if I see a business and those chat interactions and results are separate from ChatGPT's own content. But if I see a business and I want to know do they have a particular product available, are they open? Whatever, you know. I can I could potentially chat with that result and get that information, which I think could be useful in terms of conversions or encouraging me to contact that business, what have you. So I think that those are interesting things.

Mike Blumenthal [13:21]
The problem there though, Google just introduced that and they're rolling out very much more broadly over the last two weeks. They've actually roll started rolling it out, this agentic answers, client answers that they handed off to you if you can't answer it. Google has a long history of building communication products that they then discontinue. So businesses that committed to them, you know, have to then re reboot when Google discontinues it.

Greg Sterling [13:46]
Right. Rebuild their thing.

Mike Blumenthal [13:48]
The biggest problem with all of these tools is how do you integrate them into your workflow? And if you have Google and ChatGPT and others doing agentic work and then passing it on to you for the questions they can't answer, how do you manage all that? And is there a central way to manage it? It could it both is a workflow and a process problem and it's a long term communication problem that you don't want to be setting up Client-facing communication systems in the probabilistic way that Google builds them and nukes them or ChatGPT that will frustrate people, right? Or are buggy, right? Can you afford

Greg Sterling [14:18]
To c that w that will frustrate people. Yeah. Right.

Mike Blumenthal [14:25]
The Google tolerance for loose building standards? Can a business tolerate that? Because we don't know where it's gonna, you know, show up, right? Is it gonna answer questions correctly? Is it gonna transfer them correctly? Are you gonna be punished if you don't answer the transfer? You know, there's all these questions.

Greg Sterling [14:42]
So all the content comes from the business. It comes from content that the business provides. It comes from the website. So from that standpoint, there's no yeah.

Mike Blumenthal [14:49]
In the case of Google comes from the GBP, comes from reviews, perhaps.

Greg Sterling [14:53]
Well, so that's a little bit more of a wild card because you don't have control over reviews, obviously. But anyway, this is an interesting product. It's something that people should engage with and you know, test. As you showed, ChatGPT's delivering relatively limited traffic, but that's not necessarily entirely representative of the influence that it has on people's purchase behavior. But it's something that can we'll could keep watching and as it evolves. But it's pretty interesting. Okay, so let's unless there's anything you wanna add to that further. The one

Mike Blumenthal [15:23]
Yep, all good. Ready to move forward.

Greg Sterling [15:26]
One point to emphasize, you mentioned at the outset that it that the Dashboard, the GSC Lite is a sales channel for sure. It's a sales channel both to get you into advertising and to get you to upgrade to a higher level of you know subscription level, biz a business subscription level. So it's a it's trying to generate revenue for them. So that's part of their motivation, obviously, here is to create a s a sales funnel for them. And then another is probably to we'll talk about this in a minute. Is probably to create a pipeline of data directly from businesses s as a hedge against some of the anti scraping stuff that Google is doing that we'll talk to. Let's

Mike Blumenthal [16:06]
Yeah, my response to that though is that the field of the battlefield of local has a lot of bodies from vendors on it. Bit small businesses are the toughest audience to deal with. And trying to scale it worldwide is one of the more complicated problems. And scaling it in a way that is consistent and predictable and doesn't shift that much over time is really hard. And Google has come and gone in the space. They are the most successful. They ha don't have a cohesive plan, right?

Greg Sterling [16:38]
Well, and the other point to make is that, based on different data sources, f you know, recently we talked to Myles Anderson about their most recent study. It's a minority of businesses that are managing their Google business profiles. For years, less than fifty percent had even claimed them. But beyond that, the group that action actively managed them is very small. So that doesn't Bode well for this product in terms of adoption. So they're going to really need to use small agencies and others as a channel. You know, you can load up if you're a Shopify client, you can bring your catalog to this, which might be the best use case because people are doing product search different from local business search. They're doing product search in ChatGPT and product comparisons. So that information might be really beneficial. Okay, so let's talk briefly about. Review banning and you've got a kind of case study to talk about and then we'll discuss the implications of that.

Mike Blumenthal [17:36]
So Google does account level. They Google obviously nukes individual reviews if they think they're spamming, but they also do GBP account level bans preventing that account from getting reviews. They've done this in they do it on a geofencing method. They did it, for example, d when Dobbs was overturned by the Supreme Court. All of the birth clinics were prevented from getting reviews. They did it in Israel during up periods of war, but they also do it when your location is being spammed by spammers by a competitor. And if they think a business is overreaching, getting too many, and perhaps buying reviews or getting the soliciting them inappropriately, you can put a ban. So I had a case study over the last couple of weeks, HVAC business with two locations. Both were street had street addresses, so they were not SABs, both in home services that have fall upticks in business. And what we saw wa and they both have roughly the same number of reviews, give or take over time and so one of them got a ban starting at the end of September and it lasted until early this week. Couple interesting things. One is Google didn't tell them about the ban, so they had to reach out, inquire to support what was going on, and support wrote back and said, Your account's under this restriction. Only one account, one location of the two got the restriction. Google wasn't did not do anything until we reached out to them again and asked for the restriction to be lifted late last week. And Google finally elevated escalated it internally and removed the restriction. It appears that the restriction can go on automatically, but has to be manually taken off.

Greg Sterling [19:24]
Well

Mike Blumenthal [19:25]
So you have to communicate with support twice minimally to one learn it and two to get it taken off.

Greg Sterling [19:31]
W s so are these locations in the same city or are they the same? Okay.

Mike Blumenthal [19:35]
Different cities, same state, two hours apart. And a lot of times a client will have their service technician from one location or the other location, depending on workflow and that sort of things. So, and because it's an annual deal, some consumers would have will have reviewed both locations over time.

Greg Sterling [19:56]
Because the service the servicing happens in the home, the consumer can use either location hypothetically. Dep depending on where they are.

Mike Blumenthal [20:03]
Right. And the technician the t and the technician may have come from one or the other over a period of time, so they may be able to do e do a second review. So what was w so I was curious, since I had a very good sort of control group, the one that wasn't banned and a group and a location that did, I thought I could do a comparison be with a year's worth of reviews to see if there were any obvious things. Now both of these bit locations had periodic spikes during seasonal upticks. Both were doing roughly the same work. So it was good control. And what we found was that uptick by its one, what period should you look at for an uptick? What do you compare it to? And the one that correlated closest to the ban period was a 10-day look back, right? So a spike compared to the 10 previous days as opposed to a running average or year-long average or whatever. So That was the closest we had to that correlated with the ban. But in and of itself, that spike wasn't any greater than any other previous spike we'd seen, like in the spring, when they're also called back for regular service. But when we layered over two factors, one was repeat reviews. In other words, same customer had reviewed the other location and had reviewed this during this period, and use of employee names, both those two factors both spiked during the same period. So it is conceivable, although not provable, it's just correlation, that some combination of factors resulted in this ban. And what we encourage the client to do was stop asking for reviews during this period. Let reviews occur naturally, maybe send them the brand URL instead of the review URL. So that the review volume would be a little bit less. And so roughly five weeks after inception, we're able to get it lifted. And it looks like it's a multifactored variable. Although it occurred or well, occurred well before election time. Sometimes elections can you might they might geofence election if it's a controversial election. They might geofence businesses around that time frame too. But we've seen that in the past. We've seen it around we've seen geofencing go in around the Olympics, right? When there was a site holding in Olympics, we've seen businesses being prevented from getting reviews. World Cup, we saw the same thing. So Google will impose this account level ban for a number of reasons. But this appeared to be a penalty.

Greg Sterling [22:43]
So in the case of the business the consumers that had reviewed both locations, or that there were reviews that the consumer wouldn't necessarily know which location they're reviewing. They might, but if the service technician

Mike Blumenthal [22:57]
Right, the technician gives them the URL, so the technician provides the link. So in that sense, yeah.

Greg Sterling [22:59]
Yes, right. Okay. So in the c right. In the in the case of the consumer who's reviewed both locations, did the let's say there's a review you know, in f in February and then a review in September. Which review was being banned? Was it the later it must have been the later review, right? The second review.

Mike Blumenthal [23:22]
Yeah, the ban occurred during the later spike, right? So and but it wasn't

Greg Sterling [23:26]
Yeah. So it's the second review. They were able to see

Mike Blumenthal [23:29]
You're like when we looked at those spikes by themselves, in other words, two part two reviews from one party, just those spikes, the spikes themselves did not seem to trigger a ban, right? The use of employee names in and of itself or spikes in those did not trigger a ban. But some combination of the 10 day look back plus These other two factors which we could measure. I mean, there could have been something we can't measure, right? I mean there's a lot of stuff that goes on metadata that we can't see. So any it was just interesting that we did see a correl correlation between spikes in those two spikes with the in the review volume, and that was the period in which the ban took place.

Greg Sterling [24:13]
And but from your understanding this is an ethical business that isn't doing anything necessarily to manipulate re reviews. And

Mike Blumenthal [24:19]
Untour. Correct. I it yeah.

Greg Sterling [24:23]
W so what's interesting, s so two quick points. The longer the look back window, the less likely Google is to you know the if you did an annual review window, that's not gonna be meaningful in terms of spikes. I mean y it would have to be a relatively short period of time. Ten days, a month, you know.

Mike Blumenthal [24:39]
Right. Because some businesses like HVAC are very seasonal. Spring, winter, or spring and fall, they're gonna get a lot more reviews than, el elsewhere. Or heat wave they get a spike, right?

Greg Sterling [24:49]
Right. We when we were we served as a expert witness in a particular case of involving review fraud. And one of the indicators of fraud that we talked about in our report was review spikes in a weird time of year for the business. And so review spikes is probably one of the leading signals that alerts Google to the possibility of fraud and then it looks at other things, I would imagine. But that's probably a Yeah.

Mike Blumenthal [25:17]
Right. That's the conclusion I came to was that the spike got their attention and something else caused the over the top.

Greg Sterling [25:24]
Right. And r employee names in a review is a proxy for incentivized reviews, right? I mean it's something that they that in and of itself is not necessarily meaningful, but it suggests other behavior which they want to prohibit. So if you're an

Mike Blumenthal [25:39]
Correct. And it's a simple proxy to follow too, right? It's easy for Google to say, are there names in this review, right?

Greg Sterling [25:47]
Right. S so if you're an ethical business and you're you know, doing some seasonable seasonal promotion, hey, you know, a hundred dollars off on getting your furnace serviced or whatever, what should businesses be doing or thinking about in order to minimize the likelihood that they're gonna get a ban in this kind of a situation?

Mike Blumenthal [26:06]
Well, yeah, I mean, during your peak periods, hopefully you can smooth off the peak somewhat. So I don't know there's any way to do that. I you know, I think you just need to keep asking. One, you need to be prepared that if you're in a home service business, you're gonna lose reviews. That's like death and taxes. You're gonna get negative reviews and you may be subject to one of these longer term bans. So one, you should be monitoring your review volume to see if your review volume has dropped off significantly to know if you have had a ban. And then you can reach out to Google, confirm it with them. They require you to reach out to them to get confirmation and then reach out again to get it lifted after you've sort of backed off a little bit on the volume. I don't think you can do much about the idea of whether you're going to get spikes or not. You're going to get spikes if you're in a seasonable, seasonal home service business. I think letting Consumers answer naturally and not encouraging employee names to be put in the reviews is probably a better path forward. I think that employee names, whether it's a trigger or not, creates a both an easy proxy for Google to track and some suspicion that the reviews are a little more forcefully asked for than otherwise. So I think letting consumers respond naturally.

Greg Sterling [27:25]
Maybe a maybe a false positive. You determined that this would the review ban was in place and then you a month later said, Hey Google, get rid of this, take it down. If Google made the determination that this was bad behavior, why would they take it down?

Mike Blumenthal [27:41]
Well, one, the whatever patterns either mitigated or didn't accelerate or maybe dropped off to a more r reasonable level. Maybe the ban was put on, maybe it was an edge case and it really didn't meet the criteria, it just sort of triggered the automatic ban. I believe the ban goes in place automatically and it appears that it needs a human to take it off. So it could be that once a human looked at it wasn't as serious at you know, it wasn't a serious situation. So the c all sorts of things. But it seems that a human has to intervene to get it taken off, which is odd, but that seems to be the way it is.

Greg Sterling [28:17]
Okay, any other final advice before we talk about the scraping?

Mike Blumenthal [28:20]
No, just other than repeating the advice that, reviews are like death and taxes, they're inevitable and they're somewhat painful, deal with it.

Greg Sterling [28:28]
But you gotta deal with it. Okay, so Google has been trying to in the AI era and probably well before that, Google has been trying to prevent third parties from scraping its content, scraping its data, its search results

Mike Blumenthal [28:44]
At scale.

Greg Sterling [28:45]
At scale, yes. And which is terribly ironic because Google itself is the greatest scraper of them all. And this has been a long term point of contention in Europe where news publishers tried to use copyright law and governments tried to use copyright law to force Google to pay publishers for to display essentially s snippets of their content. And that went round and round and round and I don't I think that it was resolved with Google making certain kinds of agreements with publisher consortiums, but I don't remember where that stands. But it's part of the long history of Google cat and mouse game. And in the AI era that we're now in, you've got a bunch of third parties. I you've got all these SaaS tools that scrape for local ecosystem purposes, you know, for agencies and right, exactly. But

Mike Blumenthal [29:32]
And you have review tools that's scrape for review management, review analysis, competitor analysis. You have local tools that's great for mapping grids and

Greg Sterling [29:43]
That's right. And Google was had sued a company called SerpApi, which was supplying a bunch of companies, including ChatGPT, with Google data, with you know, Google

Mike Blumenthal [29:53]
So that was December twenty five, they brought that suit against SerpApi under the DMCA,

Greg Sterling [29:56]
Right. And it was

Mike Blumenthal [29:59]
Which is a copyright claim.

Greg Sterling [30:01]
Right. And the court basically dismissed it with leave to amend. Google amended. It was dismissed again. And essentially we can infer that well, we don't have to infer the Google search results are not copyrightable. It's factual information that cannot be copyrighted. So Google has done other things and now one of the things that they're trying to do. Is for is force people to sign in to get access to the search. So talk about that.

Mike Blumenthal [30:32]
So let me put this in a little bit more context. So January twenty-five, Google starts requiring JavaScript, which increases the cost of scraping. It's no longer just a simple HTML scrape, so it adds to the burden of large scale scraping financially. In September twenty-five, they dropped the number one hundred parameter so there it requires more.

Greg Sterling [30:54]
Which messes up the tools.

Mike Blumenthal [30:56]
Messes up the tools, requires more scrapes to get the same amount of data, right? So again, drives the cost to scrapers up. And then the lawsuit to SerpApi, which was servicing ChatGPT and Perplexity and others, December 25, where they determined that these if these results were not copyrightable. And then in July of this year, they tet they started doing these redirects, which again goof up the results. So instead of going directly to the business site in the search results, they were going to a Google redirect URL. Which and then just last month, September, they started poisoning SERP results. So when they saw a bot doing scraping, they'd be throwing in random wrong results. So the accuracy of the results. So and then the most interesting test to me though is was in on Barry's search engine Roundtable first seen by Brodie Clark, where Google is testing that whether they could re historically if you did a search and you weren't logged in, it would take three or four more searches before Google would encourage you to log in, even if you were a bot, right? With this, you did a search unlogged in, and on the next query required a login. So it was clearly a test. Google responded to it that. They weren't doing it as a matter of general practice. But what's interesting about it is once they put that gateway in front of users and that complies contractually for the user to come I'm sorry, it requires that the user complies with contractual terms of not scraping. Not scraping is one of the rules, always been their rules, but they've had no legal way to enforce it. You couldn't enforce it under the, you know, copyright

Greg Sterling [32:43]
Right. So then it becomes they in i essentially impose a contract on the entity or the user, which they can then enforce under co copy contract. Yeah.

Mike Blumenthal [32:52]
Right. So it's been this cat and mouse game where Google put these technical things in, try legal theories like you know,

Greg Sterling [32:58]
Copyright.

Mike Blumenthal [32:59]
Copyright and prior to that, you know, hacking, those legal theories did not play out. The only one that appeared to have legs was this formal gateway of a login requirement. And it looks like Google may be going that direction to slow these down. If this test becomes more than a test. The problem with it, from my point of view, is it There's a whole ecosystem of small SaaS tools and big SaaS tools, small businesses that have grown up around scraping Google results, reviews and map distance maps and ranking maps, as well as ranking grids

Greg Sterling [33:33]
Right. Ranking grades. Yep.

Mike Blumenthal [33:36]
And all this other stuff that have grown up around it that agencies rely on and small businesses rely on as, you know, key parts of their marketing suite. And this action if it were to take place disrupts a whole ecosystem of and makes attribution it just makes the whole process of understanding your marketing significantly harder.

Greg Sterling [33:58]
And that ecosystem of tools and agencies and third parties that help marketers understand business owners understand, this the efficacy of their marketing or what they need to do, it benefits Google ultimately. It's this is kind of a semi self destructive behavior on their part. I mean, it's unlikely, you know, what happens to in incognito mode? I mean, that just would have to go away completely. That wouldn't be you couldn't do it, you know, you couldn't do that.

Mike Blumenthal [34:25]
Right. And it adds friction to the user. It it just makes an already unpleasant experience l less pleasant. More unpleasant.

Greg Sterling [34:31]
More unpleasant. Well they're they're cramming, especially on mobile, they're really cramming the sign in messaging down your throat. If you're you know, they want to tr the if you're a iPhone user using Safari, not signed in, they're you know, they'll give you this message. They try and get you over to the app, they're trying to capture your data, they're trying to, you know, and then what this move is a kind of a prohibitive prohibit the scraping. Google is a hypocrite because this is how they built their business. They built their business on the back of r laws that said facts cannot be copyrighted. And there's the whole local search industry came out of the yellow pages and the ability of companies and third party data aggregators like infoUSA now I'm forgetting their name. Anyway there's just there were just a bunch of companies, locally info USA and others, that did some version of taking phone books and sending them to low wage labor countries, keying in the information and then r reselling that. And Google used a lot of that information. And relies on the fact that factual information can't be copyrighted. If you could copyright facts, it would be much harder for Google to have built its business like business listings as facts. And so now it, having done that, it wants to turn around and prevent other people I understand the business mentality, but it's just a hypocritical move. And Google needs these tools and these third parties to some degree to keep its ecosystem engaged and healthy.

Mike Blumenthal [36:09]
Right, but it what it's likely to do is impact the smaller players. The larger players will still be able to probably circumvent it and will be able to defend themselves legally and have the resources to do it. So there's that and i it will mostly impact smaller players. Clearly it's sh it's a shot over the bow of ChatGPT and Perplexity and Claude, whoever's using it.

Greg Sterling [36:31]
Which right, which may be one reason why ChatGPT has introduced its business profile tool. Right. That's right.

Mike Blumenthal [36:36]
And why it's licensing Yelp data, both. But from a practical point of view, I think that, you know, it really we've from a realistic working world view of a small business working with an agency, rank tracking has long been on the decline. And it's for a good reason, is that it is such a insignificant part of what happens in Google. And a particularly in AI, that it should never be a key metric that a small business looks at. It might be one thing for b for an agency to look at it for trends or to see if there's a algorithm update that maybe impacted some of your clients, but to use it as an end user reporting tool, its days are numbered. I think agencies need if they haven't stopped using it as an end user reporting tool, they certainly should. And I should

Greg Sterling [37:27]
Well, there's lots of challenges with analytics these days, as we know.

Mike Blumenthal [37:30]
Lots of challenges and this will just make it mo greater. And I think that businesses need to move towards more conversion oriented metrics, door traffic, phone calls that can track outcomes. And the middle is gonna that is whatever's happening between knowledge and conversion is gonna be more confusing. But and s you know, so I think that's the reality. I think the agencies and small businesses have to sort of Face that reality and act accordingly.

Greg Sterling [38:00]
It's worth mentioning that this probably is a non starter in Europe for multiple reasons, including privacy, but just the coercive dimension of it is going to be inherently at odds with sort of the European regulatory approach. So Google

Mike Blumenthal [38:15]
And it's at odds with Google's antitrust regulations in the United States or antitrust rulings. I mean, if a judge sees this on top having just ruled on some antitrust thing or another, there I think there's six current concl you know, cases that haven't been where they've been adjudicated antitrust that haven't fully resolved all the penalties, or maybe there's four, you know, if the judge sees them doing this, it would inflame those cases against Google as well.

Greg Sterling [38:41]
Right. It's probably not gonna come to pass and it's in that form. Okay, so let's talk about our final item, which is Ask Maps Google put out a blog post I think this week. Let's see. Yes, November sixteen on Monday, where they're kind of it's kind of a greatest it's sort of thing where they say, you know, here's a whole bunch of Ask Maps features and capabilities that will help you with your holiday travel. And they talk about, some of the agentic features like let you know the calling of business locations and the creation of itineraries and the flexibility of ask map search and integration with you know other tools that Google has. And it for me what it did is it reinforced this idea or brought back up this idea that Ask Maps is a really good product, sort of getting better with more and more capabilities and much more flexible than traditional Google Maps, better than AI mode in many respects for local search, better than Google search. And it's really the one thing, you know, putting aside Google Gmail and Google Docs, it's really the one thing that I use that's a Google product that I use pretty consistently.

Mike Blumenthal [39:57]
And but only it isn't that old and so it required you to change behaviors to use it also.

Greg Sterling [40:00]
Right well, I'm I was a Google Maps user, so it's not that big a change, but it's and we as we discussed before that the interface is kind of awkward. You have to click the button and then you go into it. And we've, you know, talked about how likely that's gonna change over time. But the interesting thing to me is the th nobody's using this. I mean Google will say, we have millions of users, but really People don't know about this. People in the world do not kn yeah.

Mike Blumenthal [40:28]
I did a survey of my family and of a group of friends, so it's 12 survey, about sixteen. And many of them are in their thirties, young thirties, of an avid maps users, go app Google Maps users, and not a one of them knew of the feature.

Greg Sterling [40:42]
And they had no Yeah.

Mike Blumenthal [40:44]
And this was only a month ago. So the awareness of the feature is very low. You have to point it out to people, explain what it does. Google's not doing that marketing, never have. I don't see them starting now, but

Greg Sterling [40:56]
Well, this blog this blog post is an example of that, but it's not being marketed to the general public. It's being marketed to industry insiders that read Google blog posts. And it's a good product that should be promoted. And it's the it's in my view, it's really the best AI tool that Google has. I mean, I'm biased obviously.

Mike Blumenthal [41:19]
In my book it just should be, as I've mentioned, the default search mode and it should

Greg Sterling [41:23]
Should be the maps should be the maps interface.

Mike Blumenthal [41:25]
Be smart enough to interpret whether it's just a data query or needs AI interpretation. It makes that decision, either returns data or it gives you a more complex narrative answer.

Greg Sterling [41:36]
Right. Gives you a Yeah, Google is trying to do a version of that with search, and so there's no reason why they couldn't do that in maps. But I just wonder what they're gonna do with this because they've got so many AI interfaces, so many doorways into it. This is a coherent product with real use cases. They've got AI overviews, they've got AI mode, they've got Gemini, and they're all offering slightly different experiences. And I just Yes, and you get different

Mike Blumenthal [42:02]
And different answers, you can ask the same complicated travel itinerary of them all, and you're gonna get what you know, criticisms of each other as well as totally different itineraries.

Greg Sterling [42:14]
But every everybody, just like your family, your informal poll of your family, everybody that I talk to where I bring it up, nobody knows about it. People don't know about it. And I see Google on TikTok, for example, running ads. Google is running ads trying to reach young women about shopping, holiday shopping. That's the theme of many of the ads that I've seen. I'm not a young woman. But I'm seeing these ads. And I wonder why they don't do this should be the subject of a consumer campaign, right? Here's the new Ask Maps is flexible. It you can do all these great things with it's really useful. That's its chief benefit, is that it has genuine utility. The public has largely gone sour on AI. With the data centers and the broligarchs and all that stuff, there's a there's an there's a lack of trust, there's a lot of and

Mike Blumenthal [43:10]
Although I don't know if you noticed in the news over the week the last couple of days that many of the broligarchs have now rebranded as progressives and not MAGA. So as moderates, yeah. So

Greg Sterling [43:16]
Yes, as moderates. Yeah. I'm no longer Elon Musk. I'm not a Nazi. I'm a moderate. But anyway, this is something that Google should be promoting and it's a positive thing. It's a good use of AI and yet

Mike Blumenthal [43:30]
Promotion of Google products outside of user uptake is not the norm. They made their millions with viral uptake of their products, whether it was search or Gmail, ultimately achieving

Greg Sterling [43:41]
And monopolistic s behavior.

Mike Blumenthal [43:45]
Monopolistic goals through being better, but still, and so the culture within Google is you put it up early, buggy, you see if it gets uptake. If it doesn't, you take it down and you forget.

Greg Sterling [43:58]
It's a Darwinian Darwinian thing.

Mike Blumenthal [44:00]
It's a very, very Darwinian probabilistic approach that consumers and businesses hate because it causes such churn on the side and the business is trying to stabilize over here. But it Google doesn't have this culture of that kind of marketing historically.

Greg Sterling [44:16]
Ex except as I was saying, they are doing so this is what surprised me. They are doing these ads on TikTok in an effort to get people back to Google. You know, there's a whole group of people out there that are basically AI haters and debunking AI. And look at what tiny percentage of Google search AI is and AI hasn't killed Google. You've you we w this is something we don't need to get into in detail, but this is really a product that deserves to be promoted. It's something that is you know would bring people back to Google Maps that have abandoned it. I mean, I was one of those people. And you know, we'll see what it does. I mean, it's but it's just it there's I think your suggestion that it become the default interface for maps is probably where it's headed. But it deserves broader exposure than that,

Mike Blumenthal [45:06]
Yeah, it's a good product. I agree. It's one of the first products from Google in the last couple of years that I actually use regularly.

Greg Sterling [45:12]
That works and it's good and is yeah, it's I mean I don't the number of times when I go to Google Maps, it's usually to use Ask Maps and it's very rare that I'm if I use if I'm getting directions, I use Apple Maps. But if I'm trying to make it, you know, I'm coming back from the airport, show me restaurants that are still open between, you know, here and here, it delivers. It really delivers. So anyway. All right, so next week we is the Thanksgiving holiday in the United States, as everybody or many people know, not everybody. And we will not be doing an episode next week. We're taking the week off, but we will be back the following week with more exciting adventures. Like and subscribe, tell your friends, thanks for listening as always. And if you are in the United States celebrating Thanksgiving, have a great holiday.


Interested in sponsoring this podcast or our newsletters please reach out to mblumenthal@nearmedia.co

E-mail Mike