AI Max vs DSA, PMax vs Shopping, tROAS vs Max Conv. Value | Google Ads World Cup

Episode Transcript

00:00:00

Mike: Welcome to another episode of Growing Ecommerce. As always, I’m one of your hosts, Mike Ryan. And with me is Chris.

Chris: Thanks for having me.

Mike: Thanks for being here, Chris. We’ve got a good one. In the spirit of the World Cup, we’re making it competitive.

Chris: Okay, yes.

00:00:16

Mike: So, we are doing a feature bracket showdown in Google Ads. We’re going to talk about AI Max for Search versus Dynamic Search Ads (DSA)—of course, DSA is going away. We will also talk about target ROAS (tROAS) versus Maximize Conversion Value, and of course, Standard Shopping versus PMax. It’s basically an elimination game, right? The better campaign advances.

Chris: Yeah.

00:00:43

Mike: Well, are we going to end up doing all of them today, or should we just do this one round?

Chris: Exactly. We’ll do one, then let’s see where we end up.

Mike: Yes. I don’t know if it’s a good sign that I’m rocking the Austrian jersey.

00:01:01

Chris: Yeah, I want to say, how should I win this today? Austria’s already been eliminated, sadly. But yes, we fought our butts off and we won against a superior team.

Mike: Yes, which was Spain. Yeah, you guys did great. It was a really good show.

Chris: We did all right. I hope that when this episode airs, the US boys are still in the race.

00:01:25

Mike: I hope so, we’ll find out. I’m positive about them.

Chris: Yeah, and I’m also positive about my chances today.

Mike: Really? Okay, that’s crazy to hear because I feel pretty positive about my chances, Chris. Look, you’re Spain, I’m Austria. Let’s see how this goes.

00:01:52

Chris: All right, so Mike, with which one do you want to start? And by the way, we have to be clear on which side we are, right?

Mike: Yeah, let’s start with AI Max for Search versus DSA. Let’s go for it.

Chris: All right, I’m AI Max for Search.

Mike: Ladies and gents, I’m on Team DSA. Even though this is a moot point due to a technical disqualification since it’s being migrated, but let’s face it.

00:02:15

Chris: Yeah, let’s face it. But Mike, I mean, how in the world can you support a dying technology?

Mike: Well, again, it’s a moot point. But on the pure technological premise, I think that DSA is tried and true. It’s been working for years. It’s more strictly bounded than AI Max for Search because it’s not going to look at your assets and stuff like that. It’s really focused on your landing pages, it doesn’t have this final URL expansion component, and it’s been reliably generating new keywords for a whole generation of search marketers. It’s a workhorse.

00:03:07

Chris: Kudos on that. For me, it’s kind of like an old reliable Audi A3, but it’s certainly not the self-driving Tesla I’m rooting for now. I think AI Max for Search is basically incorporating all the positives of DSA, but with this extra layer of AI magic, which I think is getting better and better. And what I’m referring to here—and I want to be fair—with AI Max for Search, we have been talking about how the claims Google made were not completely true, which was basically that you get massive incremental revenue at the ROAS levels you face right now.

00:03:39

Chris: But what I see across the board now is that this technology is getting better, and it really helps you as an online retailer to get into this incremental revenue everyone is looking for. And that’s clearly an advantage AI Max for Search has over DSA. That’s my claim. It’s the more up-to-date technology, and especially when it comes down to this incrementality, I think Search is the better campaign if set up the right way. This is the disclaimer I want to make.

00:04:03

Chris: Yeah, what is your take on that? I mean, do you prefer the old A3 versus the self-driving Tesla? You can’t be serious about that.

Mike: I’m not as familiar with your European automakers.

Chris: Let’s rub it in, man. Let’s see how long they will exist. But you know what, take a Ford F-150.

Mike: I was going to say the Ford F-150.

Chris: Does that make you feel more comfortable?

Mike: Yeah, DSA is a Ford F-150. No, so DSA, I hear you. I agree that in terms of doing new stuff and finding incremental opportunities, AI Max is definitely stronger there. DSA was also doing that work—it was also an expansion layer—but I would say it was kind of tethered or on a leash.

00:05:18

Mike: You were just a bit more in the steering wheel, stuff like that. So that would be the case for DSA.

Chris: Okay. What about—and I know this comparison is to a certain degree unfair—what I also love about AI Max for Search is that we know for sure that GenAI surfaces are the new game. Google sees the future of search there.

00:05:43

Chris: I want to join this party, and AI Max for Search is the big ticket to enter the chocolate factory. Let’s face it, as a DSA fan, you are not allowed to join. You have to party in your old station wagon. Meanwhile, I’m driving onto the red carpet with my Tesla, and I can train these GenAI surfaces, which I think will become more important. We might talk today about how this search variable will change, and I think AI Max for Search is absolutely the core piece next to how you serve these new GenAI surfaces. So I think it’s a must for every retailer. It’s a great technology, and Google will put a lot of energy and resources into making it even better.

00:06:38

Mike: All right, so I agree it’s better suited for the AI era. And it’s because of what we’re talking about, that it’s a bit more expansionary or untethered than DSA. The final URL expansion is the one that’s hard for me to argue against, although it’s a controversial feature. But we’ve talked in the past about this new relevance model where there needs to be this tight alignment between the query (which, as you said, is becoming very complicated), the AI content, the ad, and the landing page.

00:07:15

Mike: Now, DSA and AI Max for Search are both very much focused on the ad and the landing page. But with final URL expansion, I have to give some points to AI Max for Search because it gives even more flexibility to make sure that the best landing page for that AI content and that very complicated query is selected. So you got me there.

00:07:36

Chris: Actually, we haven’t even talked about the fact that this reliable workhorse, DSA, has always been—and let’s face it, still is—highly relevant. I was just about to bring this up, as it was your strongest argument.

00:08:00

Chris: It shows that this F-150 is still a good car. The sunset of DSA is moved now to early 2027. It was moved from September 2026 to, I believe, February or March of 2027, which is a testament to this campaign.

00:08:22

Chris: However, the campaign type by design has no chance to live forever because access to these GenAI surfaces is just prevented. I don’t know if there’s a specific reason for it, because again, the technology is working, don’t get me wrong. But that fact alone—I’m going to keep mixing our metaphors here, right? Because you’re a Boeing guy.

00:08:45

Chris: You’ll be Boeing because, in that case, DSA is kind of like the Boeing 767, and AI Max is like the Dreamliner.

Mike: I’m a 747 guy, man. This is unfair. You have me right now.

Chris: What can we agree on then? I think for AI Max for Search, there’s still a long way to go for this campaign type.

00:09:08

Chris: Yeah, I think the claims of Google, which are 40% more revenue at constant ROAS levels, are not being met. You have to be very careful with your setup, brand exclusions, and so forth. But I would make the case that this is certainly one of the campaign types of the future, which Google will be doubling down on, and it’s getting significantly better.

00:09:38

Mike: Yeah. I mean, by technical reasons, AI Max for Search is going to have to win this round. But the other thing that I would say is that it’s a bit uncertain to me how some of these campaign types play together. AI Max for Search and PMax, although they do some similar activity, I increasingly come to feel that they’re complementary, whereas I don’t feel that way with Standard Shopping and PMax. And it’s because this gives you a chance to still start with keywords. We know that PMax, if you go all-in on it, is broad and does this DSA and final URL stuff.

00:10:09

Mike: But I still think that the keywords, even if they only act functionally as a signal, give you a chance because PMax by itself is going to be sort of agnostic or very open-minded. It’s going to look out there, see what search volume there is, and do its own thing with no keywords. But that AI Max for Search campaign remains relevant because that’s your chance to put your experience, knowledge, and the things you want to target in there.

00:10:35

Mike: So I think in that sense, the search part inside of PMax won’t be that relevant if you’ve got a good AI Max for Search campaign, which is the superior approach to having your search setup within PMax.

Chris: Yes. Look, I mean, it might be a bit weak, but maybe we can agree on a draw? It’s hard for me, man, because I like DSA.

00:11:03

Mike: There’s no—I’m not calling a draw. No.

Chris: And who advances?

Mike: AI Max for Search.

Chris: Thank you very much, sir.

Mike: Yes, thank you very much. But it was an unfair game, to be honest. You started with 11 players and me with five.

Chris: Exactly, but it’s like a scripted WWE match where the referee already decided beforehand. Exactly.

00:11:54

Chris: Next one, which is a way more interesting one: Performance Max versus Standard Shopping. Yeah, I’m Team Performance Max.

Mike: I’m Team Standard Shopping because how could you not be? And this is a tough one, man.

Chris: Yes, I’ll let you dig your own hole right away. I don’t want to dig my hole right away, so I think I would like you to start, and I’ll try to return as best as I can. Why is Standard Shopping the better campaign type?

00:12:45

Mike: Standard Shopping now will have an extremely high amount of feature parity—in other words, equivalent features—with PMax. Do you really want me to start, Chris?

00:12:58

Chris: Yeah, listen. Look, if it feels like 5-0 for you, maybe I’ll just accept it and we let Standard Shopping win if you want, because I love playing devil’s advocate, so I’ll jump on later.

Mike: So that’s the status quo. There are two huge upgrades coming to Standard Shopping campaigns.

00:13:20

Mike: It’s going to get text customization, which will allow you to have hyper-relevant titles to these increasingly long queries that we’ve been talking about, and it will also power new AI-native ad formats. So those are very cool features right there. Additionally, this is a bit more controversial, but it’s going to also do Dynamic Search Ads and final URL extensions.

00:13:49

Mike: That’s crazy. And this is because, again, Google needs to make sure they’re covering the landing page part of this relevance model that we’re dealing with. We have a slide for that, maybe we’ll put that on screen once too. But they need to make sure that the right landing pages are being advertised. And it makes feed-based DSA, which is different than your landing page-based DSA, not conflict with your other activities. It’s a net-new activity, and this is just insane. So PMax will also have these features—I’m not saying that these are features exclusive to shopping—but it brings it to eye level.

00:14:14

Mike: Another thing that brings Standard Shopping increasingly to eye level is a new attribute in Merchant Center, which is for your video links. So you can now have video assets in your shopping campaigns, which is going to be really important. I have my opinions, but I’ll let you talk before I talk about PMax. I’ll let you say something nice about it.

00:15:02

Chris: Tragically smart, you know my tactic. By the way, you wanted to bury me right away, right?

Mike: Yeah, I forgot to mention that it’s not only getting upgraded like PMax, but it still has everything people love about shopping—more controls, more granularity.

Chris: So I feel like Austria now and you feel like Spain. I will be smart now. Mike, could you bring up one good argument about PMax if you had to?

00:15:25

Mike: Yes, but that’s not my job. I will, I will…

00:15:32

Chris: Look, what I would like to make the case for is that everything you said now is right. And it’s, let’s face it, a substantial upgrade of a campaign type. Probably one and a half years ago, a lot of people—maybe even I—predicted it was going to be sunset.

Mike: Yeah, right. We had conversations about it.

00:15:54

Chris: We saw in our MCC that it’s not going anywhere. You have this cost share of 15% to 20%. I think this, for whatever reason, forced Google to look at this campaign type again and make it significantly more powerful.

Mike: Yes. Are you trying to score goals on yourself here?

Chris: On my own? Where are you going with this? I want to lure you in, man. I will strike back.

00:16:38

Chris: We’re not friends now. No, but look. Everything you said is true. What I still believe is that by default, Standard Shopping has a right to exist now on eye level with PMax. But I think there’s still a use case for probably 60% to 70% of regular online players for whom PMax is a significantly better fit. Why? Because it is still a one-stop-shop campaign type. Standard Shopping will never, ever be that.

00:17:07

Chris: If I’m an online shop manager responsible for paid media, and I want to get the most out of it, I want to understand my funnel—not even in detail, I just want to make sure that I play the full funnel to what I want to achieve with the least amount of effort. For that use case, which is a quite dominant one, PMax smokes Standard Shopping because it’s the one-stop-shop campaign type I can rely on. I think there are use cases for which Standard Shopping is the superior campaign type—for the ones who are advanced and want to be in control. Fair point.

00:17:34

Chris: But the more regular use case—and this is why Google will be doubling down on PMax for the foreseeable future—is the online player who just wants to play the full funnel to the best of their capabilities, and PMax is still a way better campaign type for that use case.

Mike: All right, that’s your pitch. Give me something where PMax has the right to exist.

00:18:21

Mike: I hear what you’re saying about the one-stop shop. But in my opinion, that value proposition would be most useful to the whole SMB sector who don’t want to manage ten different campaigns.

Chris: Yeah, I was referring to that.

Mike: But the paradox is that they don’t have the assets needed, by and large, to make PMax campaigns really fulfilled.

00:19:17

Mike: People have hated the auto-generated assets from Google for the most part, but as the tech gets better, this could become a reality. I have always found that the people who would benefit most from PMax are least able to benefit from it. But what I think is PMax’s secret, if you can call it that, is that it will often outperform shopping because it has dynamic remarketing in it.

00:19:51

Mike: At this point, I see the main differentiator for PMax is the dynamic remarketing. Some people don’t like that it has dynamic remarketing baked in, but I look at how cutthroat competition is, and I look at everyone having the same technological advantages with smart bidding. If you’re not getting in front of your customer, Google knows that this user has a high propensity to buy, high AOV, and has been looking at your product before.

00:20:26

Mike: Your competitors are going to be there, and so I think that dynamic remarketing has value. It’s the main thing you can say about PMax.

Chris: Fair enough, and thanks for that assist. PMax versus Standard Shopping is a true battle, and again, I don’t want to come across defeatist, but I really think it’s not a “versus” scenario. I think they are complementary if you set them up the right way.

00:21:10

Chris: What you said is absolutely true. Remarketing is this one big differentiator because it’s coming out of the box and it’s working. I think PMax is clearly the superior campaign type versus Standard Shopping when it comes down to making sure that high-volume products and top sellers are maxed out. Especially with your top sellers, you shouldn’t think for one second that just because someone bought that product before, they belong to you.

00:21:38

Chris: That’s where PMax is superior to Standard Shopping—to make sure that your top sellers are really maxing out the market.

Mike: All right, so there is the whole hybrid topic, which maybe we could call a draw. But the hybrid topic is more important than ever if you want to run these things next to each other because they’re going to be doing even more of the same activity.

00:22:05

Mike: I really think you cannot advertise the same product in both of these campaign types. I think there’s probably nothing you can do to make that system safe enough, in my opinion.

00:22:28

Mike: But you can just decide, “Hey, these are my products that I want remarketing on, PMax is great for that. These are ones I don’t.” Especially for strategically important products, whether they are high-volume products or the whole long-tail thing, Standard Shopping has a true case here.

Chris: I give you that, no doubt about that. Our biggest failure right now, Chris, is we don’t have a referee, so we have to decide. But I’ll decide now, and I’ll decide in your favor. I will give Standard Shopping the pass here.

00:23:10

Mike: I thought you were just going to pull rank on me.

Chris: No, no, I’ll give Standard Shopping the pass because the upgrades are substantial.

Mike: The comeback—it’s such an exciting comeback story, everyone loves that. So I would give Standard Shopping the win here.

Chris: However, I think it’s rather a complementary setup if you do it correctly. And then, I don’t want to come across cocky, but we were right again. That’s why our shopping platform is capable of managing PMax and Standard Shopping at the same time, because we said Standard Shopping is not going anywhere. It’s a comeback for the ages, and I actually like it.

Mike: Me too.

00:23:50

Chris: For me, Standard Shopping is the most successful campaign in the history of Google.

Mike: Well, keyword search covers more than just shopping, so it’s very big. But from an ecommerce perspective, yes, definitely. That’s 1-1. Now it’s crunch time.

00:24:17

Mike: All right, target ROAS (tROAS) versus Maximize Conversion Value.

Chris: Yeah, you’re in camp Maximize Conversion Value, right? Good, because I know you’re a target ROAS fan. We had this discussion in the episode before.

00:24:41

Chris: Look, Maximize Conversion Value literally means I maximize my conversion value without any target constraints behind it. Have you ever gone to a bar, Mike, and just told the barkeeper, “You know what, make it the best night of my life, go for it”?

Mike: Yes.

00:25:07

Chris: Okay, yes, you did it. I’ve never done it because my credit card limit isn’t that high, man. It really feels like that for me with Maximize Conversion Value. What do you think? It sounds great on paper. But one thing is for sure—as much as I’ve shared flowers and roses with Google over the last 12 months because they did a lot of things right—I’m not trusting a NASDAQ-listed company to max out my bids without guardrails.

00:25:37

Chris: ROAS is the one thing I want to control by any means. That’s why, before we talk about technical details, I’m a big fan of target ROAS here. For me, it’s a no-brainer. I just don’t trust Google enough to let this thing go.

Mike: Well, let’s talk about that ROAS. By the way, this is hard for me because I actually am a huge tROAS fan. I’m playing devil’s advocate here.

00:26:17

Mike: But that isn’t the main value of tROAS to me. If we look at Maximize Conversion Value with no ROAS setting, you are never going to hit a ROAS of 0 or 1 anyway. That is not going to happen because this still exists within your addressable market and within the budget you have available. These are constraints.

00:26:59

Mike: So basically what will happen is that you’re going to get as much revenue as you possibly can with the demand that is available, and you will reach kind of the native or naturally occurring ROAS. It’s probably going to be around four or five, and it will vary. It will introduce more day-to-day volatility—I give you that—but I think if you look on average across your month, the number is not going to be that bad or fearful.

00:27:25

Mike: And you are going to make sure, with this demand-led philosophy, that the budget you have available and the demand that exists pace themselves, and you’re going to get as much of it as you possibly can.

Chris: Fair point, I like it. Let’s address two things: the volatility you were talking about. I agree it won’t be a ROAS of 0.5 or 1. It will adjust to the available demand.

00:28:15

Chris: But the volatility is a big issue. Volatility within a month I could live with because if on average the ROAS is fine, I think it’s fair. But who guarantees me that the next month will align with the month before? If I zoom out a bit, one month we have a ROAS of four, another month five, another month three, because demand varies. That’s the counter-argument I want to make.

00:29:01

Chris: The second thing, Mike, is that we have been talking about this for years: ROAS is one of the major signals I give to Google next to my campaign structure and budget. Google is a transactional engine, and I want to feed it precise constraints because I just don’t trust it by default. If I take away one of my three major levers, it’s not going to happen on my watch. I’m old-school here. Those are my two arguments.

Mike: Yeah, you have me here because the signal aspect is what I actually care about. It’s a way you can differentiate your campaigns and keep control.

00:29:35

Mike: Without that constraint in place, wherever the demand takes you might not meet your break-even point. But to your point about zooming out, I actually tend to think this is a productive way to look at ROAS if you don’t have a target in place. If you look at the daily noise, it’s going to spike.

00:30:13

Mike: If you zoom out to the week or the month, it will smooth out. By default, you’ll start to see the average. In month-over-month terms, this is ultimately the manifestation of your seasonality and demand fluctuations because, let’s face it, demand varies. I’m not too troubled by that. I think zooming out is healthy. But on the signal aspect, I have to give you points. ROAS is there, although imperfect, as a safeguard and a steering mechanism. I can’t really argue with that.

Chris: I think you’ve scored a point that you don’t have to fear. And you’ve scored a point that it’s an important lever.

00:31:09

Chris: I think the lever is more important than the fear. However, I’m biased here because I want to win. What I want to say, though, is I learned from the best because the target ROAS signal is one of the major points of view of this company, and we have developed software features to max out this signal.

00:31:29

Chris: For that reason, Mike, I really believe that the demand-led model is great on paper, but as an online player—whether a DTC or multi-brand online store, it doesn’t matter—don’t give away ROAS as a lever.

Mike: We can ask our producers.

00:31:48

Mike: Producer Kevin says tROAS. Well, fuck.

Chris: But you still have the US passport. The 4th of July was last week.

00:32:17

Mike: Yeah, I do. It was quite a battle. You made your points, and you smoked me with PMax.

Chris: Do we have time for one more topic, or do we save it for another episode? We are quite well into the time box, so we’ll save that for next time.

00:32:36

Chris: Yes, it was a pleasure, mate.

Mike: Likewise, I felt the pressure. Good game. Thanks, everyone. This has been another episode of Growing Ecommerce, brought to you, as always, by Smarter Ecommerce, also known as smec. You can learn more by visiting Smarter Ecommerce. Let us know in the comments what goals we missed or what points we could have scored. We would love to see that. Thanks, and we’ll see you next time.

Chris: Bye-bye.

Mike: Bye.