2026.10.09 | GP | Media | Digital Ad Spend: 3Q26 Shifts & Outlook

原标题：Google, Meta, and ChatGPT Ads: $300 Million Media Agency Executive Discusses 3Q26 Budget Shifts and the Forecast for 4Q26 Digital Advertising Spend

#internet #media

GuidePoint · 2026-10-09

Primary:

Meta

Dominant platform with about half of total spending

Scale, consumer demand, strong support and superior creative controls

Algorithm changes lowered returns; aggressive product adoption tactics

Important social platform, especially for smaller advertisers

Broad reach and useful AI troubleshooting

Weak enterprise support, compliance risks and automated creative changes

ChatGPT

Emerging incremental channel

High engagement and substantial inventory

Low click volume, unclear measurement and poorly positioned ads

Amazon

Strongest return profile in the portfolio

Captures consumers near purchase

Performance benefits from highly purchase-ready traffic

The Trade Desk

Still a major connected-TV platform

Broad inventory and lower minimums than some rivals

High fees, difficult interface and transparency concerns

MNTN

Growing connected-TV alternative

Lower fees, easier exclusions and reporting

Lacks display capabilities

Seasonal niche channel

Attractive returns

Very limited scale

Snap and Reddit

Exited

Limited current value

Weak fit and brand-safety concerns

Quantitative Insights

Metric

Result or outlook

Total third-quarter spend growth

6%, within the normal 5%–7% range but below the earlier 6.5%–7% forecast

Google share of spending

Google third-quarter growth

Google fourth-quarter growth outlook

Google return decline after algorithm update

Performance Max share of search budget

Meta third-quarter growth

Potential later Meta reductions

International travel decline

Amazon fourth-quarter growth outlook

MNTN cost advantage

45%–55%

Approximately 6%–6.5%

About 10 percentage points

About 60%

2.8% versus an expected 4.5%–5%

Up to 50%

About 70%

About 10%

Approximately 15%��20%

The Trade Desk connected-TV share

Fell from 85% to 70%; potentially 60%–65% next quarter

10/11/2026 at 11:33 AM UTC

JR: Hi, good morning. This is John Ritucci, and thank you for joining our call today titled Google Meta and ChatGPT Ads: 300 million media agency executive discusses 3Q26 budget shifts and the forecast for 4Q26 digital advertising spend. If you have any questions during our call, feel free to email them to ask@guidepoint.com and I can read your questions anonymously as they come in. And again, that email is ask@guidepoint.com. And with us this morning is Jonathan Kagan. He is presently the director of search and digital media strategy at Emissive Digital. Prior to that, he was at Nine Rooftops Marketing Agency as the director of search and biddable media. And before that, he was the associate director of search marketing at Digitas. And even prior to that, he was the associate search director at MediaCom. So thank you again for being here, Jonathan. Would you like to add anything about your background before we get started?

Jonathan Kagan: I mean, you've kind of hit the nail on the head. I've been doing this for about 22 years agency side for roughly my entire career. My bread and butter is highly regulated industries, so financial services, health care, insurance, employment, credit, etc. But I do touch roughly every vertical under the sun. I made my name in search, but I also handle programmatic, CTV, paid social, basically anything that goes into an online ad auction falls under myself or the team that I oversee.

JR: Awesome. Excellent. And yeah, maybe we could begin by getting your take on how the third quarter shaped up.

Jonathan Kagan: Sure. As a single entity or a specific channel?

JR: Oh, yeah. Maybe could we begin overall just how total ad spend shaped up for your clients and then maybe we could break it down by channel?

Jonathan Kagan: Yeah, so normal growth for us is typically between an aggregate of about 5% to 7% on average. And when I say that, we typically shoot for 6%, 6.25%. Q3 ended for us at 6% even, so roughly right on the nose. A lot of that is largely driven by search. Search as an entity between Google, Bing, and all search platforms I work with accounts for about 50% of my total ad spend, which includes Performance Max. And that grew at the expected 6%. It had been growing slower at 5%. We did see a little bit of an uptick in CPCs, which allowed it to grow a little bit faster and more at a normal level. So we're still growing at a fairly consistent rate across the board. There's certain areas that were a little bit stronger, certain areas that were a little bit weaker.

JR: Right. Okay. Yeah. So I think last time we spoke, you thought maybe the third quarter could be around 6.5% to 7% year-over-year growth. So yeah, what do you think happened that kind of missed your expectations?

Jonathan Kagan: I mean, every quarter we have a different thing to point our finger at, but it's basically the same thing with a different name. And that honestly comes down, right now we're calling it the Middle East conflict and could have previously called it tariffs. Reality is, and this hit anyone that is pushing a tangible physical good that requires shipping, we saw prices increase on that. Now, what was interesting is our ad demand didn't decrease. Our ad demand did not decrease. There was still a huge mass demand for it, but we did see a slowing in ROAS due to substitute buying. And because of that, we're forced because our ad budgets are typically cyclical. Revenue comes in, part of that gets recycled back out for ad spend. If I'm generating lower revenue or in lower ROAS, then despite the same number of purchases, I have to start scaling back our spend and it was only a little bit slower than what we were ideally shooting for, technically it was still within our target area. The one... well, not one, but the area that has really become a little bit

10/11/2026 at 11:33 AM UTC

of a sore spot for us is Meta. But this is nothing new. Meta's had a host of issues. And it's interesting, it's not the SMBs, but it's the large enterprise advertisers that deal with that. And that's what prevents us from going at a faster rate. Ultimately, they have various rules, regulations, lack of true support, and it's just really become difficult to work with them most days.

JR: Okay, gotcha. And in terms of like the industry verticals, are all of them weaker on these macro concerns? Are there any that are stronger than others?

Jonathan Kagan: I mean, so this is because I handle regulated industries, I get to see the highs and lows of the world because those verticals are very boring. They also give me the vice category. So I'll deal with cannabis, alcohol and fast food simultaneously. Well, that sounds weird and hilarious, and honestly, it is. We've discovered over the years that is a cyclical thing. When there is massive geopolitical concerns or fear of unemployment, inflation, people load up on health care, they reduce their presence for financial services, specifically on the lending side, and then they go skyrocket into their vices and fast food, which is what we've been kind of seeing. Now it's slightly modified. We're still seeing people demand. We're seeing so much substitute buying that the ROAS is just slipping. Fast casual dining is taking a little bit of a beating, which is fairly uncommon for us because so much, especially on the beef side, is imported. They've had to raise the rates. Fast casual dining is no longer a good night out or affordable night out. For us at the same time, so we've started to see some weakness there, but the biggest area we've been seeing weakness, and this has really been emerging since the summer, is international travel. I deal with a little bit of the travel, international, domestically are different categories. International has a much better margin for us. Domestic we get volume, international we get revenue and volume. And I'd say our international travel off the North American continent, or actually even within the North American continent, is down about 70%. Now we have had domestic travel was increasing until about June, and then in June it started sliding. So right now our big weak spot is travel. We are also starting to see weakening or continued weakening in lending and loans because of the interest rate hike. So that hurts automotive, both new and used for us, and it also hurts real estate for us, and it hurts the lending side of banking for us. So banking as a whole is strong, but it's being carried by the retail side while the lending is taking a beating. If you look high level at travel, it looks okay, but in reality you're selling more, but you're not making as much money on it, and real estate is pretty much hit or miss right now, and more so than that.

JR: Okay.

Jonathan Kagan: Yeah.

JR: Wow. All right. So what about the e-commerce vertical? I think last time you noted there was some downward pressure on the value of consumer baskets or price of goods sold. Is that still the trend that you're seeing?

Jonathan Kagan: Still the trend. It's starting to get better. And when I say better, I don't mean a contracting ROAS. So previously we had seen a contracting ROAS. Now what we're seeing is a slowly growing or flat ROAS. Part of it is due to seasonality. Tail end of Q3 is when the early birds really start their holiday shopping and back to school shopping. It's not contracting. It's not growing, though. And I don't know until the conflict in the Middle East is resolved. I don't foresee it getting better, and I do actually foresee it potentially getting worse.

JR: Okay, yes. So is it the high gas prices really and the high interest rates that are mostly the bulk of the macro concerns?

Jonathan Kagan:

10/11/2026 at 11:33 AM UTC

The vast majority. So anything that's a physical, tangible good that needs shipping, 100% it is tied to gas prices or the oil prices. Anything that requires a financial lending, a loan, a credit card, that is where we're seeing the increased interest rates, but I don't think we've truly seen the full effect of that yet. I think we're probably going to see it in Q4. Those are your major players. There's still some uncertainty about the job market. If the job market comes out strong, it'll relieve some pressure. But the other ones is anything that you, and there's not a lot, but stuff that has to be important from certain countries are, not Argentina, Brazil. Australia and Canada, which Canada, hard good lumber typically for us. Brazil and Australia are big for meat for us, especially on the beef and pork side. So if it's consumables, hits restaurants, they start having to raise their prices because of that simultaneously. Not an ideal scenario.

JR: That's true.

Jonathan Kagan: Needless to say.

JR: Okay. Are there any businesses that are, you know, trending higher right now? Any benefit to the stock market is at almost all-time highs? The GDP growth is still, I think, over 2% year over year. And it's just hard to square up the macro concerns with some of the other data points that kind of point to a strong economy.

Jonathan Kagan: So we are in a golden age for healthcare, but that is a highly seasonal business. So Q4, as you go into Q4, it's not so much strong Q3, but in Q4, people are burning off their HSA, FSA cards left and right. So we see a big push on the healthcare side. And then that will typically parlay all the way through mid-February as part of your New Year's resolution. So we're in a good time period for that. There's less concerns about Medicare, Medicaid at the moment for the while. So that is setting it up to be a fairly strong year for us in those categories. But it kind of is not a whole lot of huge wiggle room for that right now, above and beyond that for anything to celebrate.

JR: Okay. Well, it sounds like there's not too many bright spots, but I thought I would ask. I'm sorry, just lost my microphone. For Google, I think last time we spoke it was the bulk, it was the largest platform in your portfolio.

Jonathan Kagan: Yep.

JR: I think it was 45 to 55% of total spend.

Jonathan Kagan: Yep.

JR: What is it currently?

Jonathan Kagan: Still about that range right now. It's not really gaining any more market or share of the pie. It's not losing any, which is good. If any area was to potentially lose, it would likely be paid social and it might kick over to Google because typically when something dies off or goes off, it ends up in Google's hands. That's probably the only thing that we're looking at, though.

JR: Okay, wait, what was that about paid social that was weakening?

Jonathan Kagan: So we're seeing some weakenings in paid social, specifically in Meta. Meta does really good at garnering SMBs and pulling in SMBs. They are hit or miss on the massive enterprise. Now that doesn't mean massive enterprise brands are

10/11/2026 at 11:33 AM UTC

pulling out, but it does mean they're slowing their spend in it.

JR: Okay, and is it benefiting Google? Is that?

Jonathan Kagan: Typically it benefits Google 9 times out of 10.

JR: Okay, and in the third quarter, did you shift any budget from Meta into Google?

Jonathan Kagan: A small amount, I'd say maybe only 1% of Meta got shifted over. It's largely tied to different policy reasons that Meta has been electing to enforce selectively that just make it much more difficult to work with them on certain categories.

JR: Okay, and would those be like the healthcare, the financial, which verticals are mostly impacted by these policies?

Jonathan Kagan: So, let me tell you, maybe not categories, but more so the size of businesses. Meta is number 1 thing right now is pushing AI, especially AI enhancements, which has been an absolute nightmare for any advertiser of any size because Meta rolls out a new 1 on average 1 a week or they modify 1 a week. When they roll 1 out, they don't tell the advertiser and they activate it. If you have any legal and compliance team or your creative has to go through any legal and compliance, which all of mine do. And 1 of those legal team members or compliance team members sees a creative in rotation that they did not approve, you have to pull it down and pull it out. I go through this on average 2 times a week. And it's gotten to so bad that we have started to slow down, if not pull out some of our ad spend. And on top of it, we've had to staff up our QA operations just to go regularly check to make sure Meta hasn't activated something new, which we find on average 1 time every other week. And because of that, a lot of regulated industries or teams of legal and compliance teams are saying, we can't continue to do this because we're finding so many things that are noncompliant. Now, SMBs typically don't have to worry about that and they're always down for it or ignorance is bliss. They don't care. And one could always say, well, why don't you just tell Meta to stop? Well, Meta has gotten rid of almost all support except for the major holding companies and a handful of major brands around the world. So there is almost no Meta support anymore. It's been a documented known issue for about 3 years now. So every time they have a round of layoffs, I've had 3 reps laid off in the past 5 years. The last 1 I had was 2 years ago for our agency and we're dropping $25 million, $30 million. We're not the biggest spender on Meta, but we're not nothing. And there's no more support left for us.

JR: Wow. Okay. That's interesting. And the non-compliant examples that some of the brands see with Meta's features, could you just run us through a couple of examples just to illustrate what that looks like?

Jonathan Kagan: Yeah, so the most common thing that gets turned on by Meta by itself is adding music to your creative. So if you have a static image ad, Meta might go and turn on music that it selects. Humorously, yet terrifyingly, I was working with a chain of ophthalmology offices across the country on Meta. Out of nowhere, they turned on the audio, not the words, but just the tune of I Can See Clearly on our creatives without notifying us. Which was terrible. And that was one of the better situations. We had another one where it just started writing headlines and descriptions for us and putting them live, but it was pulling it from content on our site that some of that content was only applicable for certain states. And when they're doing it across the board nationwide, that's a problem. Another situation we ran into is it decided to create a video creative for us without telling us. The problem was it pulled from all of our images and created a 3D animation video. However, the issue is the models that we have for those images, we only have the rights for imagery. We don't have the rights for video. So it was spotted by compliance, and we then had to go do research to find out royalty rights and what we were going to have to pay off because of that, because while we didn't technically create

10/11/2026 at 11:33 AM UTC

that video, it still showed the model in motion as a video or animation that we don't own the rights for, which is entirely separate, different area of compliance. And this problem has been happening all the time. These are some of the better scenarios.

JR: Wow. All right. I mean, those are some really great examples. So you have no one to reach out to at the moment. Is that right?

Jonathan Kagan: Do you have a general chat support phone number? Or... not phone number, general chat support. If you're lucky, they get back to you within 72 hours, and then the problem is it's not the same person that stays on with you the entire way through, so they'll hand off the case. I've had stuff trying to get resolved for up to 3 months. And Meta has just reduced their staff for support. They have plenty of salespeople, but these are salespeople focused on SMB operations that aren't there to be able to provide any technical support. They'll sell you anything, but they won't support you.

JR: Wow. All right. So in the third quarter, what was Meta's share of overall spend?

Jonathan Kagan: So I don't know the exact share. Their normal growth for them is about 4%. We grew at 2.8% for Meta in Q3.

JR: Yeah, all right. It sounds a little bit low based on other checks I've had with advertisers.

Jonathan Kagan: Mm-hmm.

JR: Is that lower growth than you would have expected in the third quarter?

Jonathan Kagan: Yeah, we would have expected it to be about 4.5 to 5. But JR: Yeah.

Jonathan Kagan: we are hitting. Now, I will note the Meta's big moneymaker is a function in an ad unit that not a lot of advertisers, specifically SMEs, realize is activated, which is the Meta's audience network. It works very similar to Google's Google Display Network or Content Network, whatever you want to call it these days. Most advertisers don't realize they're activated into it. We intentionally opt out of it because when you're in that, that is the highest source of spend, typically. So we intentionally remove ourselves from that. To put it nicely, it's a blind spot. You have to really know what you're looking for to be able to find it and deactivate it. And a lot of advertisers don't.

JR: Okay, gotcha. And do you think there's more of Meta, like share of spend, that could migrate away to other platforms? Like, how do you think this shapes up in the fourth quarter for you?

Jonathan Kagan: I think for the fourth quarter, they're going to be stable. I think once you enter low season for most verticals, which is typically the back half of Q2, is when you're really going to start seeing some telltale signs of operations considering abandoning Meta or just grossly reducing it by as much as 50%.

JR: Okay. And what are your thoughts around MetaMuse? Is there any excitement or optimism with advertisers on that product?

10/11/2026 at 11:33 AM UTC

Jonathan Kagan: Not really. For the past 3 weeks, I've been talking to someone about twice a week now that it's a thing. 1 person thought it was the national chain of paint and wine bars. So not a lot of conversation about that so far.

JR: Okay, gotcha. Yeah, it seems more like a consumer hyped up tool right now, I guess. But all of those tools, what about just agentic commerce in general? Do you have any thoughts or do the clients that you speak with have any concerns that agentic commerce could really reshape the advertising funnel, possibly impacting the lower ad funnel products?

Jonathan Kagan: So yes and no. They do believe that it's going to lead to fewer shopping. It's basically the concern and the excitement is kind of one and the same. There's a belief that the agentic will be very helpful for upselling on a website. And we've seen a couple of scenarios where it has, where people go to a site and the agentic will help upsell. But there's also a fear that if Google ever goes into an agentic model for e-com, it could hurt advertisers in the long run, specifically ones that are trying to make money because it's no longer going to recommend them or utilize that function. We're still a ways out from that. Right now, I've yet to see anything beyond being on-site of a retailer, but it's something that's being closely watched over time.

JR: Okay, gotcha. And then is there any interest for the Meta business agents for some of your enterprise clients? Like, do you think they'll maybe use those for customer service or maybe trying to get higher?

Jonathan Kagan: Are you talking about the Meta business AI tool that's in the ad UI?

JR: Oh, well, I'm not sure if it's in the ad manager, but it's a, I think a tool targeting more small and mid-sized businesses, but I was curious if enterprise is possible.

Jonathan Kagan: Yeah, it's called Meta AI Business Assistant. Yes, I think that's what it's called, or business agent.

JR: Yep.

Jonathan Kagan: So right now that is probably the most helpful thing at Meta. Like I will take that tool over any other staff at this point.

JR: What does it do for you?

Jonathan Kagan: It answers questions. At A, it answers questions and B, it also diagnoses problems and can walk me through solutions. I can confidently say. In the past 5 months, that tool has been more useful for me for troubleshooting than any person at Meta.

JR: But it hasn't been able to solve that issue, though, just turning on features, right?

Jonathan Kagan: No, it is specifically designed not to do that, actually.

JR: Oh, okay.

Jonathan Kagan:

10/11/2026 at 11:33 AM UTC

It won't even acknowledge it. It'll help me with a lot of other technical bugs, a lot of function settings, but it will not touch anything with enhancements, which are these little AI features that get turned on.

JR: Gotcha. And it hasn't impacted... oh, sorry, go ahead.

Jonathan Kagan: No, go ahead.

JR: Yeah, I was going to ask, has it impacted your ad spend or do you think maybe made it easier to deploy?

Jonathan Kagan: Yeah. No. It's made it easier for me to deploy because it helps me troubleshoot problems, but it's not grown my spend or anything.

JR: Okay. And do you see any potential for a tool like that to make it easier for you to spend more on Meta?

Jonathan Kagan: I mean, there's plenty of tools out there to get me to spend more. When I talk to my colleagues across the industry and internally, we all use it as a technical support tool because it operates like tech support. It doesn't act like a selling tool, which is what we appreciate. That being said, it's not doing anything to upsell me. It's just helping make it easier for me to deploy things.

JR: Okay, gotcha.

Jonathan Kagan: Which is both a godsend for an advertiser, not great for your bottom line, but it is what it is.

JR: Okay. Yep. And you mentioned the fourth quarter you expect to be stable for Meta. Does that imply just like flat growth year over year? What did you mean by stable?

Jonathan Kagan: Flat growth.

JR: All right. Can we circle back actually to Google? I rushed through that one. So 45% to 55% of total spend currently on Google in the third quarter, is that correct?

Jonathan Kagan: Yep.

JR: All right. How would you describe the performance that you've seen on search, PMAX, and shopping?

Jonathan Kagan: Stable is the best way I can describe it. It's remained stable, which we all need some degree of stability in our lives at this point. For the most part, it's been stable. It's still growing. We're seeing a declining... well, actually, let me rephrase that. There has been a change in August that we still have not recovered from fully. So I actually lied. It's not stable. August 17th, Google rolled out a change in their algorithm for anyone that's doing targeted cost per acquisition or cost per lead and or targeted ROAS. They made these small... they described them, I really feel that they understated the impact of what these things were doing... where it was readjusting your reach targets and all that. And if you're running one of these algorithms, it wreaked havoc on your return. Only about 50% of the industry has recovered from this change. I'm in a better spot; I'm probably about 85% recovered. It doesn't impact my spend.

10/11/2026 at 11:33 AM UTC

What it does impact is my return. I'm seeing higher cost for acquisitions, lower ROAS, and while the lower ROAS was happening before this change happened, the cost per acquisition was not. It's not going to reduce my spend at all by any stretch of the means. I still need to spend to get there. It's just my return is not being as stellar as it had been.

JR: Wow. Okay, so how much has your ROAS declined since this change?

Jonathan Kagan: Specifically since this change, my ROAS has declined about 10%. Sorry, let me say, 10 percentage points.

JR: And is that out of line seasonality wise?

Jonathan Kagan: Yeah, 100% out of line. Had this change not happened, I'd be hunky-dory. I might even be growing in my ROAS or growing above my expectation.

JR: What drove the change at Google, do you think?

Jonathan Kagan: I would love to know that myself. That being said, Google doesn't make changes like this unless it's going to make them more money. But the industry is not pleased. We have spoken to our reps. We've spoken to the head of industry at Google. We've spoken to the product experts. And this change happened August 17th. It really started coming to light the impact around August 19th. Google has formally acknowledged multiple times and documented that they recognize that the algorithm is still settling and still working and they don't have a timeline for when it will be fully considered normalized, which is a huge concern.

JR: Yeah, so how much further would you let ROAS fall before there will be changes to how much you allocate to Google?

Jonathan Kagan: I'm saying not until it becomes negative.

JR: Oh, yeah.

Jonathan Kagan: It's a necessary evil. That's the reality.

JR: Okay. And is this like for branded and non-branded campaigns?

Jonathan Kagan: Yep. Everyone, everything under the sun. It's hitting non-branded a lot harder because brand is just so cheap to start off with, but everything under the sun.

JR: Oh, all right. So what was your growth year over year in the third quarter for Google?

Jonathan Kagan: Google as an entity was 6%.

JR: How did that compare to your expectations for third quarter growth?

Jonathan Kagan:

10/11/2026 at 11:33 AM UTC

We put a range of 5 to 7.

JR: Oh, okay.

Jonathan Kagan: So it was the search sides, what, 5.75, 5.5. It was carried over 6% or up to 6% courtesy of growth in YouTube and display on Google. So net net is still growing.

JR: Okay, and any plans to change your allocation to Google moving forward?

Jonathan Kagan: Not at the moment. Everyone's been asking me with the court saying there has to be some degree of breakup and sell off and all that jazz. And honestly, it does not matter because Google's growth and success and usage is consumer driven. And so unless there's another platform out there that people want to shift to because they've heard about it enough, we have to stay the course of where the consumer is.

JR: Hmm. Okay. And so for the fourth quarter, what do you think growth might look like for Google year over year?

Jonathan Kagan: For the fourth quarter growth year over year, we're probably looking, I don't know, probably about 6.5% to 7%. The YouTube and the display side are going to be elevated up until midterm election day. Usually midterm elections don't have that big of an impact on cost, but we're seeing a lot of regionalized spend go in heavy in certain geographies, Texas, Maine, New York, what have you. And those are escalating costs. So we're seeing YouTube grow at a faster rate than anything. But then day after the election, it'll drop by about 25%. So I think it's going to come out probably 6%, 6.5% for the end of the quarter.

JR: Okay, yeah, that's fair. What are you seeing in terms of the trends in CPCs and conversion rates on Google search?

Jonathan Kagan: Conversion rates remain constant. We're not seeing them grow. We're not seeing them shrink. CPCs, it depends on the vertical and the gust of wind. Healthcare, financial services get pretty hit hard just due to macroeconomic factors and seasonality, but for the most part, they're only growing about 2% to 3%, which is a normal level for us. Yeah.

JR: Okay, yeah, and how does that compare to prior quarters this year? Is it normalizing?

Jonathan Kagan: Yeah, for the most part, it's normalizing. We haven't seen anything obscure. It was coming in a little bit lower at the beginning of the year at only 2% growth. Now it's moved up to 3%. We started seeing that happen in late June. So I'm not expecting it to go any higher. So this is about as normalized as it can be.

JR: All right. And then how would you describe the performance on the performance max campaigns and the AI max campaigns?

Jonathan Kagan: So I'll start with AI Max. AI Max is very difficult to measure. Part of the reason is there's not actually a formal report for it. Another problem is Google's, and this was an issue we addressed with Google in person during roundtables with them. Their sales team and their reps sold us AI Max. None of them were trained in it. We found out there was a major disconnect between the product and the education. And so much so that Google has gone back and changed the documentation and their sales process of the platform or of the function. AI Max can be done for search queries and keywords. It can be done for ad copy. It can be done for landing pages. 95% of my clients only use it for the keyword

10/11/2026 at 11:33 AM UTC

part. And when we're running it, we find the performance slightly worse, not substantially worse, but slightly worse than if we had not been running it, but that is only on the AI Max delivery. The regular delivery does fine, and the AI Max only accounts for about 5% of our search volume. So basically, it's not enough for us to stop doing it because it keeps being sold into us as incremental. We have no way to prove that it's actually incremental, but that's how it's being sold into us as. It's not a dumpster fire because I can't prove it's a dumpster fire, but we can't actually confirm that it's doing anything beneficial for us. So...

JR: Okay, Jonathan Kagan: yeah, JR: we're kind Jonathan Kagan: of sitting in this limbo, if you will.

JR: Yeah. And without any more transparency, do you plan to allocate more or less to AI Max campaigns moving forward?

Jonathan Kagan: You can't allocate specifically for AI Max. It's either you activate it or not, and if it's activated, Google will decide how much money goes into it.

JR: Okay.

Jonathan Kagan: So there's no true activation for... or there's no true budget allocation, just kind of does its thing. And then Performance, the other one you asked was Performance Max.

JR: Yeah, P-Max.

Jonathan Kagan: So P-Max is doing great. They are rolling out some more pilot programs on it. There's still the same level of skepticism that's been around since 2021 when it first rolled out. You're definitely letting the fox watch the henhouse mentality to it, which isn't stellar, but we don't have any ways to disprove a lot of their things. Now they have rolled out... they've been getting a lot of backlash, they being Google, have gotten a lot of backlash from the industry, and they've tried to address some of that. They've added negative keywords, audience targeting, or exclusion capabilities. Right now in Europe, they're piloting a program where you can actually weight each channel of Performance Max... your Gmail, Maps, shopping, search, what have you... and say, "I want to put more money towards this one, or I want to put more money to this one" based off weighting. It's in an alpha pilot; it hasn't made its way to the US yet, which is interesting because that pretty much defeats the purpose that they've sold P-Max in on. But the jury's still out on that functionality; it's still in pilot mode right now. 90% of the clients that I work with that can utilize it have adopted it. The ones that can't utilize it, typically it's because of creative reasons. They lack some display creative or they lack some video creative. If you don't add in video creative when you launch it, Google will create their own and it's pretty bad. So that's the extent of it. And basically anyone who has the assets and can adopt it has for us.

JR: Okay, yeah. How would you rate the creative automation on Google versus Meta?

Jonathan Kagan: Google is much better than Meta, hands down. Google doesn't really get into... Google will adjust layouts of creative from imagery. It might change videos to make them shorter. It might change the framing, so it's a horizontal versus

10/11/2026 at 11:33 AM UTC

vertical, but it doesn't go change creative outright, like beyond that from a creative interpretation. It also, rather than scraping the site, unless you're utilizing AI Max, it takes your other content and your search queries to help write creative. More importantly, it asks you if you want to utilize it before it deploys it. So I would say that Google's AI creative development is far superior.

JR: Yeah, okay, that's fair. And in terms of your allocation to Performance Max versus your traditional search campaigns, are you increasing your spend on PMAX? What does it look like?

Jonathan Kagan: Yeah, I am now. Right now, PMAX is probably getting about 60% of our search budget.

JR: Okay.

Jonathan Kagan: It varies from vertical. I don't have any scenario where it's less than 40%. I have several scenarios where it's greater than 70, though.

JR: All right. And do you think it will be increasing or decreasing moving forward as a share of Google's spend?

Jonathan Kagan: So that is a little bit of up for debate at the moment. So when Google first rolled out Performance Max, they always said it was a complementary tool. You shouldn't stop what you're currently doing. This is running in conjunction. About a year and a half ago, Google finally started saying to us, yeah, Performance Max should replace your shopping. Now, they haven't gotten rid of shopping, but they're telling us it should replace your shopping. Just stop running shopping ads and just run them through Performance Max. And whenever they do something like that, that's usually a premonition for a year or so out that they're going to disable one function. You're going to have to make the switch. They haven't disabled shopping yet, but I have this odd feeling that they may. And if that is the case, then it's going to have to absorb more money.

JR: Right. Because what percent of your Google spend is shopping currently?

Jonathan Kagan: For e-com brands, it's about 30%.

JR: Okay. And are you happy with the performance that you see on shopping ads right now?

Jonathan Kagan: For the most part, yeah. For the most part, we are pleased. It's usually the shopping functionality is usually right on par with that of Performance Max, give or take a couple percentage points in either direction. But if they officially drop, like they're officially dropping Google Display Network because they merged it into Demand Generation starting, I think it's January or February of this coming year. If they pull the same thing with shopping and telling you you're going to have to go into PMAX, then yeah, it's a consolidation. I'm not going to be investing more into Google; I'm just going to be reallocating it.

JR: Yeah. And where would you think that that shopping ad spend will land for Google and Pmax or other campaigns?

Jonathan Kagan: Yeah, it's going to end up going to Performance Max.

10/11/2026 at 11:33 AM UTC

JR: Okay. Do you think that their shopping ads could be potentially disrupted by a Generative AI commerce adoption like this Muse and other maybe consumer agents?

Jonathan Kagan: Not really. Basically, you're coming to Google for price quotes for basic research. I don't foresee it becoming a problem anytime soon for them.

JR: Yeah, and it hasn't shown up in the performance yet, is that correct? Like, the performance is in line with... yeah. Then, yes, so for Google's support versus Meta, do you get better support, would you say, than your connections at Meta right now?

Jonathan Kagan: Oh, without a doubt. There's actually connections. There's actually support. There's people that are there. They have changed their business, their customer support model over the years, which has gotten worse, but it is night and day non-comparable.

JR: Yeah, I've heard from some other advertisers that they say that Google seems like very aggressive recently in their pitches. Is that sort of the impression you've gotten as well?

Jonathan Kagan: Oh, yeah. I would use the term predatory, but yeah.

JR: What has changed, like versus, you know, maybe before this trend, I guess, this aggression? When did they begin to show signs that they wanted to get you into AI Max or whatever position?

Jonathan Kagan: So it comes in waves. Honestly, it comes in waves. They do this every year. You see it really ramps up in Q4. Typically, whenever there's a new product rolling out, they try and do this very aggressive push for adoption. Our reps, we know for a fact, have to meet a quota of selling in certain products and getting adoption of certain products. And so whenever a new product rolls out or a new function, they become incredibly aggressive to trying to push it.

JR: Okay, I've heard similar things as well. And for ChatGPT ads, I think you mentioned last time we spoke that you were experimenting with some client budgets there. How are those going?

Jonathan Kagan: Mixed bag. It wasn't exactly what we were sold. The click-through rates are pretty bad. The impression volume is a lot better than we ever expected, which makes us realize that the free version of ChatGPT is much wider used than we realized. The click-through rate is pretty horrendous, but what we've discovered is there's an anomaly that was never disclosed to us, and honestly, OpenAI needs to still work on their sales documentation to begin with, so it's nothing terribly surprising. Just because an ad triggers doesn't mean you see it. And the perfect example is, I was planning a trip to Montreal for a weekend during the summer. I said, here's my criteria. Here's my itinerary. I need 5 days, 6 or 5 days, 4 nights, Montreal in the summer, yada, yada. Create an itinerary for me. It did. It was like 12 pages long. The ad automatically triggers and I get hit with an impression. However, the ad shows up at the bottom of the sixth page or 12th page, whatever, the very last page. If I hadn't been intending to look for it, I would have never seen that image or I would have never seen that ad and that is a massive disservice. We see a fairly weak click-through rate, iffy conversion rates, but we honestly believe it's because of the design of the ad placement.

JR: Okay, they're always at the bottom, is that right?

Jonathan Kagan:

10/11/2026 at 11:33 AM UTC

Yeah, at least from the ones we've experienced.

JR: Yeah, okay.

Jonathan Kagan: I don't think it's going to take any money from Google. I don't think it's going to take money from other platforms. I think it's going to be around for a while. I think it's going to be a minority partner for us, but we don't think it's going to be stealing from Google or anything like that.

JR: Okay, yes. So would that have counted as an impression if you didn't scroll down and actually put it into the browser view in that example you just gave?

Jonathan Kagan: No, it has to trigger typically. So whatever you see at the top of the page on Google is also at the bottom of the page. So they have to be visible.

JR: Okay, it has to be visible to count as an impression, and Jonathan Kagan: Yep.

JR: you're still seeing low conversion rates, even though people are seeing them, but it's because you think they're at the bottom of the page.

Jonathan Kagan: The conversion rate is acceptable, but the volume that's actually making it to the site from the ad unit is negligible.

JR: Yeah, okay. And is it being measured accurately? Is that potentially one of the causes?

Jonathan Kagan: We're still trying to determine that because OpenAI wasn't most transparent with telling us how to measure these types of things. So it's been kind of a mixed bag.

JR: Okay, so do you have to rely on their pixel?

Jonathan Kagan: Yeah, and they were delayed actually launching a pixel in the first place. When they first rolled it out, they didn't have a pixel, which was kind of weird.

JR: Yeah, and is the pixel working effectively? I don't know really how this all works.

Jonathan Kagan: So the reality is it seems to be working fine. It's a little bit tricky to set up. Most operations use Google Tag Manager and you try and use a Google Tag Manager plug-in for Pixel. We have to use JavaScript for that one because no one's just created a template for it. It's not the worst thing, it's just not the best thing. But Rome wasn't built in a day.

JR: Okay. Yeah, well, that's true. And then you mentioned last time we spoke that there was a lot of engagement noticed with these ChatGPT ads, I think you said around 90 seconds versus 45 seconds for Google. Is that still a trend you're seeing?

Jonathan Kagan:

10/11/2026 at 11:33 AM UTC

Yes. Still fairly high.

JR: So if ChatGPT made some changes to their ad format, maybe introduced, I think they actually recently announced like visual ads. I think they're just static images, and they might appear just when you are generating longer, you know, answers that require more time. So in the loading period, they're showing these ads now. Would that be helpful, do you think, in terms of capturing more ad spend for ChatGPT?

Jonathan Kagan: I do. It's just a matter of, honestly, I do, but it really comes down to what it looks like and where it sits.

JR: Yeah, yep. And I guess there's so many levers that they potentially could pull in the future because they're so early, I guess, in building out this product. If they gave you video ads, do you think that that would be helpful in allocating more spend to them?

Jonathan Kagan: It wouldn't hurt, but once again, it comes down to where it's sitting visually.

JR: Yeah. I wonder why they don't just put the ads at the top and the bottom or not.

Jonathan Kagan: So I was having a conversation with someone about this the other day and says because it could look like you're giving favorability, which would disprove the ability to give it a confident recommendation if it looks like you're giving favorability. So that is I don't know how to resolve something like that on their end. But yeah.

JR: Okay. And have you expanded the testing to more clients of yours on ChatGPT ads?

Jonathan Kagan: No, we're holding still for the moment.

JR: Okay. And do you think there will always be like a place in the budget for these types of ads just for the visibility and the engagement?

Jonathan Kagan: We always have a testing budget, but the full future state is TBD.

JR: And it sounds like you are pretty confident that this is not going to impact your Google spend or Meta or any other properties. Is that right?

Jonathan Kagan: Correct.

JR: Yeah, it says all incremental budget, if any, will be added to ChatGPT ads moving forward.

Jonathan Kagan: I'm sorry, what was that?

JR: Would it be all incremental budget being allocated to ChatGPT ads if that's the case moving forward?

Jonathan Kagan: It'll be incremental or basically they will come if someone else weakens. It's not so much if they do well, it's if another

10/11/2026 at 11:33 AM UTC

platform starts showing signs of noticeable weakness and we have to start pulling back, we'll probably look to consider that as a reallocation over to them.

JR: Yep. Okay. Have you noticed maybe any potential impacts to Google search as a result of just more agentic search habits or just the usage of LLMs in general?

Jonathan Kagan: Not in my verticals. I have heard there's been some pretty substantial hits on publisher content sites. And the amount of traffic that's coming in and out from them. But nothing that I've witnessed, but I have heard the conversations about them.

JR: Yeah. Do any of your clients have those issues with like declining organic traffic on Google and potentially did they offset that with more spend on Google search?

Jonathan Kagan: None of my clients, just because we don't touch the verticals, I typically get hit by that the hardest.

JR: Yeah.

Jonathan Kagan: I know some online magazines and newspapers where they have started to shift and they're pumping more into social to try and offset it because people like to consume their news socially, but that's about it.

JR: Interesting. Okay. Yeah, last time we spoke, you had a little bit allocated to Amazon-sponsored product ads for, I think, the clients that it was suitable for. Have you expanded spend on Amazon-sponsored product ads?

Jonathan Kagan: Not really, just because it's not right for seasonality for us yet. We're going heavy in November.

JR: Yeah. What do you think the growth might look like on the spend there for Amazon ads in the fourth quarter?

Jonathan Kagan: From year on year, we're probably looking about a 10% growth, which is about normal for us.

JR: And in the third quarter, what did it look like for the Amazon ads in Jonathan Kagan: So it JR: the third quarter?

Jonathan Kagan: was not up by that much only because Prime Day or Prime Week was reduced in time. The duration last year ago was a week this time in October. I'm sorry, you're talking about Q3, I'm talking about Q4. Oh, no, we had a Prime Week in July too. And it was just a couple of days versus a full week. Because it went shorter, we didn't increase as much. So we were doing about 6% year over year, which is roughly normal for us.

JR: Okay. And what about the ROAS on these ad types compared to the other platforms?

Jonathan Kagan: I mean, it's incredibly, incredibly high. It's the single highest ROAS of any platform, but you're cherry picking. So it's not the easiest judgment. I can say, yeah, I have a 10,000% ROAS, which is great, but that's because you have met

10/11/2026 at 11:33 AM UTC

someone at the very end of the journey.

JR: Yeah. And would you say that the ROAS has been improving over the course of this year or not?

Jonathan Kagan: Not really.

JR: Not really. Okay. Is it going down at all by chance?

Jonathan Kagan: A little bit. It's not contracting, it's just not growing at the rate it has been.

JR: Okay. Gotcha. And then I think Pinterest was a little bit of your spend last time we spoke. Is that still the case?

Jonathan Kagan: Right now we're out. We'll go back in in November.

JR: Yeah. Okay. Are you happy with the performance that you did see? What pulled you out of it?

Jonathan Kagan: It's basically a seasonal platform for us. The reason we don't stick into it year round is it's got such a niche audience, and they have a terrible retention system with that. So if I am strapped for cash, yeah, I'm going to get a 5X ROAS versus out of Meta, but it's going to be 1/50th the volume, so I have to go where there's going to be volume.

JR: Fair. All right. And then another small social platform, Snap, I think last time you mentioned it was like very little in terms of gains in Snap.

Jonathan Kagan: Yeah, we evaluated it. We completely exited.

JR: Any plans to come back? No?

Jonathan Kagan: Not really, unless I'm starting to hawk predatory credit cards with obscenely high interest rates, which right now is every credit card. There's no reason for us to go back in.

JR: Yeah. Have they offered you any credits for ad products?

Jonathan Kagan: They have not actually.

JR: Then Reddit, have you had any spend there in the third quarter by chance?

Jonathan Kagan: No, we pulled out back to brand safety concerns, which keeps getting us in and out of it. So we ended up pulling out. They have some brand safety tools, but they're not great. And the last thing I need is my ad showing next to some really bad conspiracy theories.

JR: And is it the majority of your clients enterprise level, like, would you say, or is that sort of brand safety a huge issue?

10/11/2026 at 11:33 AM UTC

Jonathan Kagan: Oh, absolutely.

JR: Yeah, and what percent would you say is not enterprise grade of your clients?

Jonathan Kagan: I'm sorry, what was that?

JR: Yeah, what percent is small, mid-sized business, or would you say total revenue?

Jonathan Kagan: Only 10%.

JR: Okay, good. And then TikTok, did you spend anything there in the third quarter?

Jonathan Kagan: Yeah, it's not a whole hell of a lot. We're maybe doing a couple hundred thousand in Q3, which is actually up, but it's only up because of net new brands being onboarded to it. For like-for-like brands, it's flat.

JR: Okay. Anything about the performance of this quarter that's worth highlighting?

Jonathan Kagan: Nothing really.

JR: Then The Trade Desk, I think you mentioned last time we spoke that there was some share erosion there and it was a lot of your programmatic spend. What percentage of The Trade Desk is currently in your budget?

Jonathan Kagan: So it's specifically the CTV side of the programmatic, and it's gotten worse. I would say when it used to be on the CTV side, that The Trade Desk had 85% of my share of CTV, now it's down to about 70% and it's probably going to go down even further in Q1. It's probably going to drop to about 60% to 65%. We're just finding better rates with other platforms, specifically Mountain, and The Trade Desk platform remains a massive pain in the butt and barrier.

JR: Yeah.

Jonathan Kagan: And their bad press does not help them.

JR: And the bad press doesn't help them? Is that what you're saying?

Jonathan Kagan: Yeah, I've actually had clients come say, like, we no longer want to be associated with them.

JR: Anything in particular? Did they find an audit went bad or something?

Jonathan Kagan: There have been several audits that went bad, but more so the expose smear campaign by the holding companies that put those out calling about lack of transparency fees, all of which I honestly don't disagree with. But I also look at those articles and statements as a little bit skewed because those are companies are building their own networks. So they only have to gain and benefit from a downfall of The Trade Desk. But what they did say is stuff that we've all

10/11/2026 at 11:33 AM UTC

suspected and wondered about for a long time.

JR: Yeah, okay. And Mountain, how much lower are the fees that you see when you migrate that spend?

Jonathan Kagan: It varies on my targeting, but between 15% and 20%.

JR: Okay.

Jonathan Kagan: And when I'm dropping millions of dollars, that adds up quickly.

JR: Sure. Does Mountain have any competitive advantages other than the fee structure compared to The Trade Desk?

Jonathan Kagan: Yes and no. It's a lot easier to exclude networks with Mountain and actually see what networks you placed on. It's a hell of a lot easier to do reporting on, and those are typically the wins for us on that one. The weakness is it lacks display.

JR: So why do you think you're going to keep around, I think, 60% to 65% on The Trade Desk still? What is keeping you there?

Jonathan Kagan: They're still one of the biggest programmatic CTV networks under the sun, second only to DV360, and DV360 has massive minimums to meet where The Trade Desk doesn't.

JR: Yes. Okay. And what about Amazon DSP? Did they win any share of your spend?

Jonathan Kagan: They don't, but that's for complicated reasons that I can't get into.

JR: Oh okay, I see. So, The Trade Desk, what do you think the least you like, where's the floor for them in terms of, like, you got to be there just because they're large?

Jonathan Kagan: Probably about 50% right now. That's what my guess is.

JR: And do you think eventually that remainder will end up going to Mountain or DV360 or others?

Jonathan Kagan: My speculation is someone's going to purchase Mountain. I don't know who. Money and history says typically it's Google, and then they'll get that money, they'll get more money from that one always.

JR: That's one way to do it. All right. Well, I think this was very helpful. So thanks again, Jonathan. Hopefully we could catch up again soon.

Jonathan Kagan: Always a pleasure. Thank you.

JR: All right. Take care. Bye.

10/11/2026 at 11:33 AM UTC

DISCLAIMER: The statements or opinions expressed today are those of the Advisor and not Guidepoint, who disclaims all liability for the content provided. The Advisor may not disclose material nonpublic or confidential information or any information that would cause the Advisor to breach any duty or obligations. Guidepoint is not a registered investment advisor and the information provided is not intended to constitute investment advice.

10/11/2026 at 11:33 AM UTC