Name: Unlocking the Power of AI Max: Smarter Ads, Better Results
Uploaded: 2025-09-18T16:18:09.353Z
Duration: 59 min 52 s
Description: Unlocking the Power of AI Max: Smarter Ads, Better Results

Transcript for "Unlocking the Power of AI Max: Smarter Ads, Better Results": Hello, everyone, and welcome to the webinar today. Big thank you to everyone who is here, bang on time for the start of the session, and I'm sure we'll have a few more people joining us over the next couple of minutes. So while they continue to join, I'll quickly pass over to the excellent panelists that we've got lined up for the session today and let them introduce themselves. So, Kat, do you want to, start us off? Sure. Hello. So I'm Kat, one of the cofounders of House Performance. I love everything about paid media, have worked in it for over fifteen years, and just love everything about the way it evolves and the data you can get and all the different things we can do. So, yeah, I'm excited to be here. Is that enough of an intro? Yeah. Yeah. That's good. And, yeah, you like I said, just before we came on air here, you've been sharing a lot about AI Max recently, so it'll be good to get into some of those insights that I've seen in in some of your LinkedIn posts. Diana, do you just introduce yourselves for us there? Yeah. Hi, everyone. I'm happy to be here, be part of the conversation about Google Ads and the AMX specifically. I am a senior search specialist at Zesa. I work with, Kirk PTC Kirk, as you know, for four years. But, my journey with Google Ads started back in 02/2015, and I worked as the, agency account strategist through Google Partner. So, had a lot of experience with, accounts and strategies across the board. But, as Google evolve every year, it's never a dull moment. So, it's it's exciting. It keeps you on your cheek to toes to stay in touch with the reality of digital marketing. Yeah. It is getting quite crazy. With AI, it's just getting faster and faster. But, yeah, we'll we'll get into that later on. And, Scott, so I'll pass over to you lastly. Cool. Thanks. Yeah. I'm Scott. Senior paid search director at Journey Further. We have a team of 27 PPC analysts across, Leeds, Manchester, London, and New York. Super talented team, and they've been helping dive into a lot of the analysis that I've been pulling this week on on AI Max. But, yeah, excited to get into it with you guys. Yeah. There is there's a lot to get through. And like I said, Scott, you've you've been sharing a lot of interest and stuff on LinkedIn recently. Some of you in the audience may have seen some of those posts from Scott, some of those posts from Kat. We will be getting to that. But, yeah, there's a lot to cover today, and we have had a few more people join us now. So we will get going. Welcome everyone again. Very glad to have you here, whether you're watching live or if you're watching back on demand. For those that don't know me, I'm James, the content lead at Lunio. I will be hosting the webinar today, and we're aiming to get through the content in around forty. Maybe we might run over and go to forty five minutes, but we'll do, the best that we can there. And we will leave some time at the end for a dedicated q and a session, which I'll say more about, in just a second. Just a few other very quick bits of housekeeping before we get into the agenda for today's session. I know there was a question about this earlier in the chat, but, yes, we are recording the session, and it will be available to watch back on demand. If you have any questions about anything that we cover, and anything the panel's discussing today, please pop them into the dedicated q and a tab so they don't get lost in the chat, and then we'll try and get run to as many of those as we can at the end of the session. And, also, you can upvote the questions you're interested in, which will help me prioritize which ones to ask first. So please do make use of the upvote button as well. I'll also let you know that we've got some resources, and links to LinkedIn posts in the docs tab here in Goldcast, and I'll be directing your attention there at relevant parts throughout the session today so you can get more context and insights, on some of the things that relate to AI Max that we'll be covering today. And in the docs section, you can also find the links to the LinkedIn profiles for Scott, Kat, and Diana so you can connect with them if you haven't done so already. We'll also be running some audience polls during the session, which you can vote on in the polls tab. So please do give us your responses there. You should see a notification pop up whenever a new poll is open for responses, and you can respond by clicking on the polls tab and submitting your answer. So on that note, I'm just gonna bring up our first poll on screen here. I will open that and then share it. And just to clarify, the other polls won't appear on screen. You'll have to go to the polls tab, which will be next to the docs tab. So the first poll is very relevant to the topic of AI Max. So the question is, do you trust Google's automation, IE, Broadmach, PMax, and AI Max more today than you did two years ago? And the options that we have there are yes, much more, slightly more, about the same or less. So while we're waiting for more responses to come in from the audience there, I'd love to get a a few quick reactions from you guys on the panel. So, if I was to put it to you on the panel, guys, that the audience is very much going with slightly more here. So I'd be interested to hear it if you guys agree with that. Definitely. I mean, I would be much more if, AMX wasn't, kind of punched into this question, but, definitely, Broadmach and Pmax campaign have evolved over the past two years that, producing much better results than initially we were. Nice. Are you feeding similarly, Scott? I would say much more. I think probably about two years ago is when we started taking away, a lot of the, search query volumes and search query reports and stuff like that. And now now you get a lot more in this insights into what's going on. So you can you can see those PMAC search queries. You can see the AI Max ones. You can see what queries are matching with what headlines, all of that good stuff. So I think that the the transparency coming through, leads to a a much higher level of trust. Nice. And, Kat, any thoughts on that before we move on? I'm also in the more camp, definitely. But I think, as Scott just said, like, we've got more controls now than we ever had. So, like, the PMax, when it first launched, you had very little insight and control, and now you can do loads to control it. So has the tech got better, or has it just allowed us more levers? Don't know. Interesting. Interesting. Well, the audience is a clear resounding, winner there from the the options that were available, and most people are trusting Google's automation slightly more than they did two years ago. So interesting to get that perspective and take there before we move on. So that is most of the housekeeping out of the way for me. We've got five main themes to cover today before we move on to the q and a sec section at the end. So firstly, before I pass over to the panel, to make sure that just everyone in the audience is on the same page, I'm going to very quickly provide some basic foundational context as to what AI Max is and also address some of the common myths and misconceptions about it. After that, we're going to get into one of the most interesting and revealing parts of the session today where Kat, Scott, and Diana will be talking us through the results of their experiments with the iMac so far. So we'll have some exclusive insights from them there. Then we will move on to look at some of the guardrails and controls you can put in place to keep spend on the control when you are using AI Max. So we'll be considering things like opting out of the search partner network, excluding competitor terms, excluding your own brand terms. And we also have some early exclusive findings from our own data science team here at Lunio, showing the effect enabling AI Max has on invalid traffic rates for search campaigns. I'm not gonna give it away completely here, but I will say that from what we've seen, it does have a significant impact on the amount of bad traffic you're exposed to, but we will come on to that later. Next, we'll explore where AI Max fits into your overall account structure, especially if you're already running other automated campaign types like Pmax. And we'll also get into some of the situations and circumstances where it probably doesn't make sense to use AI Max because it's definitely not something that you need to use, in all cases. Then we'll close with some predictions from our panel on what AI Max means for the future of search and what you can do today as a marketer to prepare for those changes. And as I mentioned before, we will have our q and a section at the end. Again, if you have a question that you want the panel to address, please put it into the q and a tab, and, also, please upload for your favorite questions. That's everything on the agenda for today. So as I said, first of all, you're gonna have to listen to me ramble slightly longer while I go through some basic foundational context, and then we're gonna get into the, results and the experiments that our panel have been running. So one of the summaries or descriptions that I like of AI Max, comes from Mike Ryan, from Smarter eCommerce, and I just have a relevant LinkedIn post from him there, which I'm gonna share on screen. And in Mike's post, I'm just gonna read the the relevant section of it. So he said, quote, the earliest known reference to AI Max is a leaked email from May 2024, days after GML, where a Google exec calls it Searchmax or PMax for search. That's exactly what AI Max is. Earlier in November 2023, Google held focus groups asking people questions such as, how do you feel about migrating regular keyword search campaigns to Pmax campaigns? The response, presumably, was not good. This was about a year after Google completed a forced migration, of Smart Shopping to Pmax and seemingly haven't learned a lesson, they were already offering voluntary migration of DSA and GDA campaigns. That's when someone at Google had a brilliant idea. What if instead of bringing search to Pmax, we bring Pmax to search? So they pull the final URL optimization, which is DSA like and search term matching or keywordless capabilities from PMax into Searchmax. As the Corning touched, riding the wave of AI overview and AI mode hype, they renamed it AI max for search campaigns. Ladies and gentlemen, dear esteemed colleagues, a star is born, end quote. A shout out to Mike for that summary. If you'd like to read his post in full, you can find a link to it in the docs tab. But I very quickly also wanted to make you aware of an episode I recorded with Jill Saskengales, who I can see is there in the audience today and in the chat. So thank you, Jill, for joining us. Some of you may have already heard this episode, but if not, it's well worth a listen. Jill does a great job of debunking some of the most common myths and misconceptions about AI Max. For example, she explained that AI Max is not a new campaign type. It's an optional setting within search campaigns, and Jill also explains why AI Max isn't simply just an example of Google taking away transparency. And in fact, it's being released alongside new transparency features like the ability to see headlines and landing pages in your search terms reports. So if you wanna hear more from Jill, do check out that episode. And if it's useful, it's worth considering subscribing to the Lunio YouTube channel. We're gonna be releasing season three of the podcast, in late October, and we've got some great guests lined up for that already, including Lewis Kemp from Lightbulb Media here in Manchester, and also Ram Fischkin, who is the cofounder of Moz and SparkToro. Also, very quickly, at the end of each month, we also release a new episode of our paid media news show, where we run the commentary and analysis from the top five paid media headlines from the previous thirty days so you can stay up to speed with the latest developments. That is more than enough for me now, and apologies to the panel there for rambling on. So we're gonna pass it over to the experts to get the results from their experiments. So I'm just gonna stop sharing that slide. And, Kat, we will we'll start off with you. So I know I've mentioned that your LinkedIn post already, you have been experimenting quite a bit with AI Max. And for those who might not have seen some of the post that you've been making on LinkedIn, could you kind of broadly talk us through, the the key findings that you've come across so far? Yeah. So to be honest, testing so far, it's not the scary but also game changing feature that I thought it maybe was gonna be. So the first post I shared was about a client where we switched it on, six weeks ago from the data that I shared, and I just measured we didn't do, like, an AB test. We just switched it on one fairly high volume campaign to see what would happen. The client is a b to c client. They're in the health space, and it's a 20 k a No. I think Kat's Wi Fi has dropped out a little bit for me there. Oh, you're back now, Kat. Sorry. Okay. Sorry. That's unusual. So we switched it on, and nothing craze nothing that crazy happened. It spent a little bit more like, it spent more on broader terms. But you can see the search terms come through quite quickly. And some of them were not irrelevant, but a lot broader than we would have bid on, definitely, which is what I would have expected. But they did quite quickly learn that they didn't convert. So although we have the option to exclude them, but, you know, irrelevant terms, actually, it was learning that they weren't gonna drive leads quite quickly, so that was really good. I haven't tested on that account the, I've forgotten the terminology now, which is a bit embarrassing, but the one where it automatically can tailor the ad copy, just because that client well, every client really likes their ads to look a certain way, and they don't wanna go rogue with it. So I'm not in a rush to test that feature. That obviously might make it go slightly more wild. But, yeah, so far, I would say I don't wanna say underwhelming because I'm really glad we didn't turn it on and it went completely wild like Pmax did for some back in the day. But equally, I'd love to see more because I feel like we've been sold a dream. You know? Yeah. The hype the hype is real. Kat, I just wanted to bring one of the slides up here, and this was something that I think you shared recently. Looking at the impact of AI Max on exact phrase and and broad match keywords. So for for people that might not have seen this, could you maybe, summarize what people are looking at here? Yeah. So this is that that campaign I'm talking about. The big arrow shows when we switched it on, and this is just, impressions segmented by match type. So the amount we were showing on phrase match went down quite significantly. You can see in that middle chunk, and Broad kinda spiked after a few days. But, actually, it's really leveled off since then. Like, if I was looking at this campaign just for the last couple of weeks and didn't know it was switched on, I wouldn't notice straight away it was, I think, which is I think it's quite interesting. Yeah. Yeah. For sure. And, Scott, we'll come on to you next then, you know, it's AI Max has been keeping you busy as well. I know, Scott, you were one of the one of the first people to share some of the experiments that, you have been running, and I think there was a post that you made. I think it was about three months ago now. And if I just bring that up on screen, I think that was the the kind of inconclusive results of your initial testing. And could you maybe talk people through, you know, that initial test and anything you've noticed interest in, since you ran those early experiments? Yeah. I think that was a it was a very early test. Like, it was as soon as we could we could get on the, be it for Air Max to get it switched on, and we didn't see much of a difference in in performance. There was a notable increase in impressions. This was with all features, so clearly the text customization feature wasn't working that well because click through rate dropped on on that increased volume of impressions. And, ultimately, there was largely no change between the, control and the test that we run. As I say and as you've said, that was three months ago. We've done a lot more testing since then. So we've tested it across 12 different accounts, and seen a huge mixed bag of results. A lot, a lot of change in impressions versus control versus test. Of the 12 accounts we've seen, six are better ROAS, six are worse ROAS. So you can imagine, yeah, it's just a a complete mixed bag across all the tests that we've been running. Yeah. Very interesting stuff. And, we'll we'll we'll maybe get into some of those other kind of, semantic scoring tests that we're doing and I'm posting about recently later on in the session today. But lastly, Diana, it'd be great to get your insights as well from what you're seeing at Zaydo marketing, and whether what, Scott Scott and Kat are seeing, align with what you've seen at Zaydo. Pretty much. Yes. So, very similar results, with some variations. Of course, in, my first, thought of, seeing a month of data from AMX when we just switched one of the campaigns was, well, that nothing ground shaking, or, I would say, underwhelming because of the expectations we had of, much more AI expansion that actually didn't happen. I'm seeing that CTRs are improving over time, and I'm noticing that actually the AI matching, side is not doing as great as the DSA part of it, the, URL extensions, actually. So, I'd say it compensates for lack of, perhaps AI expansion because it's still, sort of limited, and not a lot of progress for overall overall uplifting campaigns. I'm not seeing, any big changes, but, we are running another a b it's not it's an through an experiment test for AMX, which is just rolled out recently, at least for, kind of where I'm testing things. And, very interesting setup, quite different results. Right off the bat, it's spending more, than we had previously tested. So depends on the industry. It depends on your structure of the campaign. So there's a lot of factors that go into how AI will behave. Nice. So so on that topic of spending more, that gives us a nice transition into our next theme, which is keeping control of spend. So before I pass over to the panel to get their thoughts on this, again, just quickly wanna run through a few relevant LinkedIn posts that I've come across over the past few weeks to set the context for this topic. So we've got some insights, again, from Mike Ryan, some from Sophie Logan, and some recent commentary from Jill Saskengales, as well. So starting with Mike Ryan, some of you may have seen the data that he shared on LinkedIn, which, if I have the slide here, this is indicating that AMX seems to drive a very large number of impressions on the search partner network, which is not necessarily always a bad thing in all cases, but something to be mindful of if you are using it. And then he shared other data in a different post, which is showing that AI Max driving a lot of impressions against competitor terms relevant to relative to broad match or phrase match, which again isn't always a bad thing, but something to be aware of. I don't have time here, and we don't have the time on the panel to get into all this data from Mike, and the smarter ecommerce team. But if you do wanna check it out, and read it in more detail, the links to Mike's post are in the docs tab. Before we come on to the next post from Sophie Logan, I just wanna read a very short excerpt from one of Mike's posts where he said, personally, I'd exclude search partner network, exclude competitor terms, exclude own brand, not rocket science, but dear lord, is it necessary, end quote. So Mike with some top level recommendations there for keeping spend under control. And then next, we've got this post from Sophie Logan. She's been experimenting with the AI Max for a month now, and she put together a post recapping her findings. Again, the link to her full post is in the docs tab. But, again, just to quote a short part of the quote here. She said, quote, mid month, I did start to rein in some control by adding more negative keywords and excluding app traffic. I tried to be loose and free, but I shouldn't be surprised that I can't handle giving over that much control, end quote. So, again, some further exclusion recommendations there when it comes to negative keywords, and excluding mobile app traffic. Big thanks to Sophie for sharing the results while we're following her on LinkedIn if you're not already doing so. And lastly, before I shut up and pass back over to the panel, I just want to read part of a post from Jill Sask and Gail's. We have got the full post on the slide here, but the text is tiny, so I'm not gonna make you read it. But as always, Jill has a very balanced and fair take on the situation, urging against knee jerk negative reactions, saying that AMAX is simply going beyond your existing query set to find new opportunities that will help meet your goals. She said, quote, like all AI, AI Max starts stupid before it gets intelligent. So if you turn on AI Max for three days and then analyze performance, you're gonna think it's a hot stinky mess of garbage because it is, because that's exactly what it's supposed to do, testing out new queries you're not currently reaching in order to learn what might help you, achieve your goals, end quote. So thanks again to Jill for her perspective on that. And with that context set, I'm gonna put a question to the panel here, and I'll just let whoever wants to jump in, and kick us off first on this one. Feel free to jump in. So what advice would you give to marketers who are concerned about spend getting out of control? Do you think it makes sense to follow the recommendations from Mike Ryan, or do you think it makes more sense to put less constraints on the system and at least try to give it the budget runway needed to find new opportunities and optimize performance? So, whoever wants to jump in on that, I'll let you open up. If you think it's if your biggest concern is about spending a good way, just don't run it. So it's probably, number one piece of advice. No. I think it's it depends on, what your current setup is, what you're trying to achieve, how comfortable you are with risk and, testing, really. It's like expansive thinking versus reductive. It's like, do you want to go out there and get all of the terms and find out what works and what doesn't and then slowly pull things back in? You can do that in different ways. Like, you can add the negative keywords. You can add the, URL exclusions. You can even now add the text guidelines in to stop it going so broad with the copy that it's writing. Like, you can continue continually pull things back in, or you can do the historic way of running things and build it out slowly, but that depends entirely on what it is that you're trying to achieve, the account setup that you currently have, and, again, you're you're comfort with with risk. Nice. And, Diana, from from what you said with with the accounts that, you've seen spends obviously go up a little bit, is there anything that you're thinking about when it comes to, keeping that in check? Absolutely. Just good old negative keywords, can come in handy. Making sure if you have a pretty strong, competitors like Mike Advise ad, some of them as negative. So, though, I am not a huge fan of pushing everything to exclude right off the bat because system needs to work out, and some of the things can come through as, some positive, conversions, some some results. I think that to start with, you have to carefully select which campaign is in the test because I don't think every campaign should be on the AMAX, the way we tested on AMAX, and depends on the combination of things that you already have on the account. So choose it carefully, see how aggressive it can go. Do you have a headroom for this campaign to go higher? For example, when I was testing, one of the campaigns, it was limited research volume. So I was I was curious to see if AI, Max can give me a little bit of boost of impressions and how it's gonna fit into whole performance. So, those kind of things and and that campaign was more, restrictive in terms of spend, so it didn't blow out. It was quite modest increase in cost and with no concern. So I think considering this, would give you at least a better guidance. And then from there, you can start based on your approach to experiments, you either can expand or you, kind of scale back on what you want. One piece of advice I would suggest, because from my experience, what I'm seeing that, it's URL expansion, it takes a lot of budget, where the spend goes from that whole AI max piece, dual page fee. So instead of opening it up and trying to kinda growl it into, like, with exclusions, just set up a a small page feed that you're comfortable with driving your traffic to, and that would be, way more targeted and way more controlling so you're not gonna expand to the entire website. Super helpful piece of advice there that, yeah, everyone should get on that and start building those page feeds. But before we move on, Kath, is there anything else that you'd, add on this point? I think just if you've got strict parameters like, I don't want this campaign to appear on brand, set it up as a negative before you turn it on. Like, I know it's all been said before, but it's only going it's only doing what it's promised to do, really. And it's Google's giving you all the parameters to set those guardrails. Right? Like Diana said, set up a specific page and let it run off that. So I think, yeah, people wind a little bit about these things going rogue, but you have got the controls to control all of that. So just really think about it before you switch it on. That's probably my advice. You know, you wouldn't run a generic search campaign and not exclude brand anyway. Right? I would hope because that's the whole point of it being generic. So just use your normal search best practice hat. Just keep doing it, basically, is what I would say. Yeah. And look look what levers of control you do have available to you and make sure that, yeah, those are being used. Yeah. So as I mentioned during the agenda, I have some world exclusive data to share with you all here in relation to the, impact of AI Max on invalid traffic rates. So I'm just gonna bring the relevant slides on screen here. So this is early preliminary data that we've looked at from one of our larger clients in the luxury retail space where entry level price for their products is around £500 going all the way up to £10,000 on the higher end. The business had been using the new for twelve months prior to January 2025, and that had basically caught their overall average IBT rates, in half across their Google search campaigns, bringing it down from just over the 7% mark to a stable baseline average of 3.7%. And from January to May 2025, we saw a steady maintenance of that baseline average, IVT rate of 3.7 or 3.7%. But then in May, they took part in the open beta that Google ran for AI Max. And as part of that, AI Max was enabled on around half of their sales campaigns. So that gave us an opportunity to take a look at what impact that had on IBT rates. So in total, 895 sales campaigns were included in the analysis. 404 of those campaigns had AI Max enabled, and 491 of those campaigns did not have AI Max enabled. Our data science team did some statistical analysis on both sets of those campaigns to assess the impact of enabling AI Max on IVT, and the findings will be revealed on the next slide. So there we go. In short, for search campaigns with AI Max enabled, the average IVT rate rose from three point seven percent to five percent with highs of up to six percent recorded. So when we look at that change in the baseline IBT rate from three point seven to 5%, that represents a 35% increase in the amount of IBT those campaigns, with AI Max enabled were exposed to. And for accounts spending large amounts like this particular client, this represents a significant increase in wasted ad spend. So before we move on to our next topic that the panel will be addressing, I'd say, if you are planning to enable AI Max or you've done so already and you're spending quite heavily on Google search, especially, if you're in the region of spending over a million pounds a year, it is well worth auditing your traffic to see what level of exposure you're dealing with. Because, ultimately, unless you look under the hood, you don't really know how much of a problem you've got there. And then once you have that data in hand, you can decide whether it makes sense commercially to address that or not. Depending on the industry and market you're in, the auto may reveal that you're only dealing with 2% or less invalid traffic, which in most cases is gonna be less of a concern. But if your average IBTU rate is consistently hitting 5% or higher across search, there's a strong business case to be made, to do something about it when you run the numbers. So if you are interested in auditing your traffic, there's a link in the docs tab, and one of our team will be on hand to set everything up for you. But now, that is it for the slides. So I'm gonna stop sharing that and move on to our next topic, which is where AI Max fits into your account. And we're going to, get through these last two, themes as quickly as you can, and then we'll get on to the q and a because I know there's been quite a few, questions popping up in the chat. So, I'm gonna pass straight over to the panel, with this question. And, again, if anyone on the panel wants to to open us up and answer first, we'll go from there. So how should people think about AI Max in relation to the other campaign types they may already be running? For example, PMax demand gen, standard search, standard shopping. In other words, how do you see it most commonly fitting into an overall Google Ads account structure, and when does it make sense to use? So, anyone on the panel, I'll pass it over to you. I don't mind starting. I think it to me, it's not a new campaign type. It's just another lever on your current search campaign, really. And so I think, you know, say you were to start a brand new account tomorrow and you you you had a midsize budget, you wouldn't necessarily start with PMax. You probably wouldn't start with AI Max. You probably start with your standard search campaigns. And then once they work for a bit, you would then add PMax. You would then test switching on AI Max. I think in the end, it will probably be on a standard. I don't think we'll we might get a choice. We're getting more controls at the moment, but Google might obviously just force on everyone. I don't know. We'll see. But I don't think it necessarily needs to be treated as this, like, big, new, special, separate thing. To me, it should just be an extension of what you're doing on search already. And it only fits to turn it on if you've got room to grow or appetite to test, appetite for a little bit of risk, similarly to how you would treat PMax, I would say, and and DemandGen in that sense. Nice. Scott, I know we were speaking and and you had some data in relation to Symantec scoring on this. So I'll I'll pass this over to you, but I'm also just gonna bring those results on screen so so people can see them. Yeah. Yeah. Yeah. I I think exactly to Kat Kat's point, it's it's incremental. It it's layered on top of what you're already doing, and it's how comfortable you are reaching further, how how comfortable you are, casting the net wider. So what we did was, got AI to rate each search term that AI Max was pulling through, but also each of the other match types versus the standard keyword set in each account. And I know you're trying to reach further than the keywords, so this isn't to say whether it's good or bad, but just have a control and say whether it match what it matches up to, like, match up to your existing keywords. And the different match types do exactly what you'd expect them to. Exact match matches most closely to your exact existing keyword set. Phrase match goes a little bit further than that. AI max tends to play in the area around where broad matches, like it's reaching much further than your your standard set was. Where this gets really interesting is we decide testing the AI Max on a campaign or on an account that, wasn't running broad match, and it added a fairly strong, semantic score. Like, it was fairly closely matching to the keywords in the account, from a meaning perspective. But then when we switched broad match on, the AI max semantic score dropped off a cliff. So we had to go much further to to find search terms that we weren't already able to match to with the match types existing in the account. So, yeah, it it does what you'd expect it to. It gets incremental stuff that you weren't already matching to. It goes further. It did properly. So, yeah, to, again, to cut the point, it it's to layer on top of your existing campaigns, and I don't think you'd you'd launch with that really. And, Diana, would you be, in broad agreements with with Kat and Scott on that one? Yes. And, I just wanna add that while AMX is not a new campaign, it definitely shifts things around an account because if we utilizing a a a structure of the congress that includes the DSA campaign, Pmax campaign, those are overlapping features. And, we don't it's not an easy way to access how all of those feature work, across those campaign types and how they overlap and what drives what because there's definitely, at least I'm noticing that my DSA campaign takes a little bit after AI enabled AI Max in one of the campaigns. So there there is definitely gonna be a shift in my account. So you have to account for that, not just, have an AI max, like, a silo vision on this campaign, how it impacts other campaigns also important. So, probably carve the space for this campaign by perhaps including some of the exclusions or adding some exclusions to other campaigns and build a clear picture of what performs. Nice. And before we move on from this section, I know we've kinda touched on it already in few of the the previous sections. But, obviously, if you've got a really tight budget and, you know, you're really concerned about what you're spending, as Kat, you said, it probably just doesn't make sense to use Air Max in the first place if you fall into that, category. Likewise, if you just have a really low appetite for risk, like, it's again, it probably just doesn't make sense. But are there any other scenarios, that you guys can think of where it just doesn't make sense to use Air Max and and just kinda spelling out for people the the the kind of situations where it probably doesn't make sense to use it? Is there anything you you'd say there? Not really. I need just if you don't don't want any level of risk or exploration, I personally think it's a massive shame if people don't use it, but that's that's more me. So I think if you don't like any risk, obviously, down the budget, I can't really think of any other reason. Scott, anything else, or is that pretty much it? I I think, to the risk point, like, highly regulated sectors, particularly because you can't use all of the features of it, so you you wouldn't necessarily be able to use the text cost, all generated assets and things like that. Obviously, that's gonna restrict you in your use of it and not give it a full go of what you what you're trying to achieve. Similarly, the by the terms you can be done and stuff like that. So, yeah, highly highly regulated industries like legal and medical and stuff are gonna struggle with it. Nice. Yeah. That's a good point there. So we're I'm just conscious of time, so I'm gonna move on to our final theme. It is crystal ball time. So let's look to the future and get some predictions from the panel, on what things are gonna look like. It's it's really hard to say. And as we spoke at the the opening of this, like, AI is is causing everything to to change so quickly now. But the question that I've got to pose to the panel is if you were advising a CMO with a big search budget, how they should plan for a future where AI Max dominates the paid search landscape, similar to how Pmax has come to take a dominant position within the ecommerce space even though adoption rates were were low to begin with in 2022. What advice would you be giving to that CMO to to set themselves up to be in a good position to capitalize, on on the future of the search landscape? I'm not sure what AI Max does take over like PMax does, first of all, unless Google just actively decides to scrap the keyword completely, which it would do. But, like, until that point, I don't think it does. But the main thing for me for for voice marketers is what data are you feeding it? Like, what is your conversion set up like? What audience data do you have? Are you looking at lifetime value? Are you looking, like, you get leads, not only how well do they convert, but how well can you convert them? Like, what's the quality of your call center answering the calls? What's the quality of your reps like? What's the intent or outcome of those lead? Like, as much information data as you can feed in to show what success looks like or what success has looked like, the better to guide it because we we just see far too often, just like a generic lead or not to the extent of sales data that we have. They're not looking at things like profit, and people complain about the results that they get. But the more that the AI is taking over, the more you need to guide it with that conversion data with that audience data. Nice. Thanks, Scott. Yeah. Hopefully, Google aren't gonna rip keywords away, in in the coming years, but we we shall see we shall see. Our cats, is there anything that you say on on that point? Don't be scared. Just embrace it because we don't have control of this stuff, really. You Google's great. There's many platforms out there that you can advertise on. It's not a thing to add. There's many ways to make money, not just on Google Ads. But, also, the best advertisers will just adapt. Right? If you embrace it and adapt, you do well, basically, as we've seen with all the developments over the years. So if we go keywordless Yeah. That's that's the way it's going. You you can't you can't fight against There's literally no point sitting there worrying about it. Yeah. So we're not gonna gain anything. I don't think I don't think they'll do that at the moment. I think what they could do is, you know, when you I don't know if this has happened already, but if you start a brand new campaign, AI Max could automatically be on. That definitely could happen. And they could certainly make it easier to what's the word? Like, dupe people into putting it on. Like, they're like the the broad map, you know, make this campaign broad, that whole trick they were doing a while ago. So they'll make it more forcible to uptake it. But, yeah, I think if you just embrace it, use the levers you have, and adapt, you'll be fine. That's all we can do. And that and that is the beauty of this industry. Right? That's what keeps us on in jobs. Yeah. Yeah. Keeps you on your toes for sure. And, Diana, is there anything that you'd say on that point about looking to the future and and what's what your thoughts are on what's coming down the line? I think that if we're not embracing it now, we're gonna be very late in doing so in a couple years. The landscape of, search is shifting to AI, and AI overview is taking notes. One of the largest used AIs right now through Google, browser Chrome browser. And, AI Max is the prerequisite for ads, to start showing up in, AI overview. So I think if, you think of where your, audience is shifting and where you need to go to be, to capture them, there is a mind shift definitely gonna be for some of the just an assumption of how broad match keywords work, how AI match now works. And if you're not testing AI max now, you probably gonna be late for the party. So, my suggestion is, start planning, right now. Start testing right now, and, just be more aggressive in your testing. Just not necessarily pushing more budget, just, develop multiple strategies that you need to test over time to find the best combination that works for you. Nice. Thank you very much, Diana. And on that note, just because I'm mindful of time here, we'll leave it there for the content. And with the time that we've got left, up until 05:00 here UK time, we will address a few of the questions that we've got in the q and a tab. So thank you again to all of you who have already submitted questions and to all of you who have upvoted, your preferred questions. If you haven't submitted any questions yet, there is still time to do that now, in the q and a tab. And please also upvote, the other questions you're interested in in hearing an answer to as well. So if we go with our first question here, I'm just going to bring this up on screen. And then, panelists, I'll leave it up to you, whoever wants to jump in and respond to some of these, but I will share each question, and kind of read it out, for the audience there. So we have, this question from Tanner, Parsons. With AI Max, do we need to, be aggressive with URL exclusions or anything? Say if we have a blog or other parts of the site that are more supportive, or will the algorithm learn, quick enough? I guess that might be related to what you were saying, Diana, about, paid feeds. Yes. I would so if you historically run ads on blog posts and you have some of them performing really well, I would still recommend to do, more extensive page feed, and that will help you to have a little bit more, control over if you trust that your side is optimized because DSA is, first of all, relies on the quality of your website, your links, your content. And that's what's gonna be generated on the ads, front. The quality of ads is gonna depend on quality of the pages. So, I would say only use what you are tested before, and that would be my suggestion. Nice. Thank you. This, you know, things then. Testing comes a little bit later with additional URLs. Nice. So we will just to to get through as many questions as we can, once we have an answer, I'm just gonna move on to the next question because because otherwise, we won't, get through a huge amount of questions in the time that we've got, and there are quite a few here. So the next question that I've got is, during your experiments, did you heavily exclude those broad terms, or was it able to learn on its own, not to show up as much in broad terms in the end? I think that might be in direction to you, Scott, or or potentially, Cap. But, if I do, if you want to, take that one. Yeah. If it's about the semantics stuff, we didn't exclude the broad terms. Like, it literally just like, as soon as Broadmax switched on, it took the it took the volume and that, Max stopped matching to the terms, as closely or stopped matching to the queue with us closely. So no. There's no no exclusions on those broad terms. Nice. Then there is another question here from Tanner again. Have you seen anything in terms of a d to c, b to b, or b to c, more or less benefit from AI Max? Does any industry benefit more or less from AI Max? I guess, kinda similar to how Pmax was raised around for ecommerce and and not as strong for for b to b lead gen. Is there, anything that we can say in relation to that? I think the most successful test we ran was on, a retail account where there was tens of thousands of different products, and it was matching to the product providing specific copy. There was writing itself, like, involving the SKU that someone had searched for one click, one conversion. So, like, retail loads products similar to PMax. It was great in that. Nice. This is an interesting one. I don't because this is a question It'd be interesting to understand how AI Max performs against DSA. Do you see DSAs gradually being phased out as AI Max grows? And I remember I had a recorded an episode of the podcast with Nava Hopkins, who formerly worked for Optimizer and not worked for Microsoft, and she was convinced that DSA campaigns were gonna be taken away in the not too distant future. And with AI Max now available, do any of you guys have thoughts on DSA campaigns and whether those are gonna stick around? I would say I haven't run a DSA campaign in years, personally. I think they're gonna go, and I don't really see the need for them anymore. Like, with a combination of PMax and AI Max, I don't see what value they're adding on top. I haven't done it as I haven't done an AB test between them, to be fair, if you're running any. But, yeah, I don't see the need. I don't know if anyone's got anything. Yeah. Diana, is there anything that you'd say in, for DSAs? So it's interesting. I was at that roundtable for AMX, before it was released. And the question I asked directly, Google, their plan to deprecate or phase out DS18 in the future. And the answer was no. But that is no. I treat it like not right now. So perhaps it will go away, in a couple years or five years. I really don't wanna see it go away because and some concept works really well, for us. And, AI math does give, this a little bit of boost in performance on DSA, but I don't feel like it's gonna completely replace it. Maybe it's, part of me speaking, I don't wanna lose control, because you say, has more control. In, AI max, it's a little bit more wrapped into AI features with more customization of the pages, and there's other things that I don't quite know. I haven't actually figured out, the how sensitive that is, to to the changes I make. So, hopefully, it's not gonna go away anytime soon. And by then, AMX will be more vast and improved to substitute it. Nice. Well, we'll pay attention to that. Before we move on on the DSA, Scott, is there do you have a fondness of DSA's, or is it something that you haven't really, used much? You've you've asked me that question before, I do. I'm running with Manchester. That's right. That's right. Yeah. I certainly didn't use to run DSA's and then really got into them as of as of late, but, yeah, I think you can do all your search term mining and stuff too. Things like Microsoft through PMax now now that they've given you back the, the, like, the view of the search term report and stuff like that. But, yeah, all all of the existing tools like broad matching and things, I think, fill the gap where DSA is at. Nice. There is a question here. We've had a few up votes, and it's directed to you specifically, Kat. And I think it was Kat, in the campaign you were just talking about, did you use a lot of broad match keywords beforehand already? Yeah. So that campaign has a mixture of broad, phrasing, exact yeah. A pretty even mixture, to be honest, but, obviously, broad takes over. The the chart was an impression, so broad's always gonna take over, a win in that scenario. But yeah. Nice. And then we have, another question. Tanner, like, shout out to you. Your questions are receiving a lot of upvotes. So we've got another question from Tanner here. If you are constantly limited by budget, is it worth turning on, to note this product was described by my Google rep as utilizing all the extra budget toward more broad type, content based on signals? I would say nine. If you're really limited by budget, nine, ten, 10. Yeah. Great. Definitely. I shouldn't limit the, smart bidding in many ways, so I don't think you need to expand more. They already tell you you don't have enough budget, then you want more exposures, so kinda counterproductive. Nice. Yeah. Play it safe. But let me, address this one. It's interesting. Is AI max the only requirement or option for an ad to be up here? Sorry. In AI overviews, if I understood, my Google card manager correctly, broad match alone is sufficient for now. I anyone have any insight on that? So I will yeah. We've been sorry. Go on. Oops. Go ahead, Diana. Oh, thank you. Wait. I'm gonna mute. I'm sorry. It's a little bit of delayed. So I think couple of things to note here. The initial documentation that Google provided about AMAX, overview, they do mention, broad match, on search campaign. They also mentioned there need to be smart, big, and strategy. Now I'm seeing that, actually, in order for you to show up an AI overview, you need to have an, AI, content customization, which means that you have to have this enabled. The broad match alone are not gonna give you, either as much exposure or at all. So AI, customization part of the ads will be important and DSA expansion, that's just two two features that are, basically, in the later determination, I'm seeing it more, being moved towards that. And I think the future with, just broad match and small bidding loan is not gonna cut it. Nice. And, Kat, to to pass over to you on on that one. Oh, yeah. I I was actually gonna say really similar. Yeah. Basically. Fine. Yeah. All good. So let me find the next question here. So we have one question here. When you set up an AI Max, do you prefer to leave room for exploration using only a few negatives and then tightening gradually, or do you start with a very strict structure right away with lots of exclusions and controlled targeting? So what would what would be your approach with that, Scott? I would leave room for exploration. I mean, why you're running this is presumably for that exploration. It's to cast a net wider. It's to try and find those new nuggets, those new search terms to appear in, in those new longer tail searches than you were historically. So, while you should put some, guide rails in place to make sure that you're not going crazy. Yeah. The the entire purpose of what you're trying to do is to explore, so I'd allow it to explore. Nice. Another question that we've got here. Thank you, Deborah, for submitting your questions. Quite a few here. So does AI Max actually work a lot better with broad match, than with exact and phrase match, or does it just increase your impressions and clicks, but not necessarily conversions? I think that's very dependent on the account, the keywords that you're using rather than just, like, boldly saying, or broadly saying broad up phrase or exact are being used. It's a case by case. It's a big old it depends, unfortunately, I think. Classic. We we always have to have at least a couple of it depends answers in here, but it's it's unavailable. Unavoidable. Unavoidable. Sorry. But we were in the time that we've got left, I'm just gonna bring up one more question, and then we're we're gonna move to close just because, we don't wanna run over for time too much. So the final question that I will put to the panel is I expect AI Max to test brands and competitor and loosely related terms. However, how do we account for terms that are nonsensical? For example, a debt is showing up for searches for tire shops and veterinarians. That shouldn't be happening. Interestingly, the data that I shared is for a dentist, a cosmetic dentist, and we didn't have that. Every every search was teeth related, should I say. But, yeah, if that is happening, either it depends again, it depends on your appetite for risk. Either exclude irrelevant things to help it learn or switch it off and try again in, like, a few more months when it's learned more. Interesting. Diana or Scott, have you ever seen anything nonsensical show up, or has it been all kind of broadly what you expected? Yeah. I've seen, broadly expanding. For example, if you're looking for a car rental, all of a sudden, it says, stay in Los Angeles. It's something that is remotely relevant just like veterinarian can be. The best needs from dental care as well. So we can kinda stretch it, across these terms, but, I think it's, you're absolutely sure, and you've tested this before, that those terms are not gonna be, valued to you, but, exclude them and continue to test them. It's not the Google gave us this levers for a reason, so use them. Just don't abuse them. Don't over don't over exclude things. But at the same time, the direction where Google is going that it's gonna be predictive. And people that are searching for, for some reason, they had a reason to show some big reason to to show the ad, to serve the ad for dentist, because, historically, they might have been looking for this, type of target before services. So, depends where your campaign goal is. If you want just to be bottom of the funnel and just go after certain keywords, then by all means, exclude them. If your purpose is to expand, give the system a little bit of time to learn and understand that. Nice. I think we're gonna leave it there in the interest of time so we don't run over too much. But I just wanna say a massive thank you to Scott, Kat, and Diana for joining us today and sharing the results of their AI max experiments. Really appreciate it. And, I'll pass over to you and the panel if there's anything any of you would like to say or mention, before we wrap up and and finish here. Just test it. Worry less. Test more. That's that's the that's the note that we need to end on. I'll take it out for that. So, if you have any questions for us, you can contact me directly via james.dini@lineo.ai, and I'll be happy to help and and point you in the right direction if I can. As I mentioned before, if you're interested in auditing your your search and AMS campaigns to find out your your average IVT rate across those, Just click on the link in the docs tab, and our team will be there to help you out. Lastly, big thanks to all of you for tuning in either here or if you're watching back, the on demand recording. Hope to see you again in our next webinar. But for now, take care everyone and all the best with your AI Max experiments. Cheers. Thanks. Bye.