PMax Channel Controls: What the New Sliders Do to Your Budget

Episode Transcript

00:00:00 - 00:00:35

Chris: Welcome to another episode of Growing Ecommerce. Today we are going to talk about the biggest update to PMax since PMax.

Mike: Wow.

Chris: Yeah, and that is channel controls in a form. So it’s a pretty strange feature where you can control channels. It’s a major concession by Google. The way it’s structured is odd. We’ll talk about that, and we’ll talk about other implications about this which might be beyond the borders of Google.

00:00:35 - 00:00:58

Mike: Exactly, well beyond—at a market level. And also, a friend of the podcast, Jyll Saskin Gales, said that this is proof that PMax was a failed experiment. I’d like to debate that with you, Chris.

Chris: That should be the headline for this podcast!

Mike: We’re proposing headlines now. I think AI writes the headlines. Of course, we are better than AI.

00:00:59 - 00:01:23

Chris: Are we though, in creating headlines?

Mike: Yes. Everything else? No. Creating headlines, I would still say I’m better than Gemini. You can test me on that. I’m arrogantly better than Gemini. But in general… Claude is—no, just kidding, just kidding.

Chris: So, like it’s a failed experiment. But let’s keep that question.

00:01:23 - 00:01:47

Mike: Yeah, let’s wait for that question because this might be a great discussion.

Chris: Let’s start with what we have on our tables right now. There is a massive feature coming in alpha status, which is called channel controls or channels. Just like in the campaign setup, there’s a section called “Channels”.

00:01:47 - 00:02:06

Chris: Yeah, so Mike, enlighten us here. What is this feature about?

Mike: Sure. Before I start talking, Chris, should I send you a link to the screenshot so you can see it?

Chris: Yeah, although without glasses I’m all right. I have to look at the symbols. Hold on.

00:02:06 - 00:02:30

Mike: I’m sorry, it’s fine. I sent it there. I just sent it to the product manager and said they should see it as well.

Chris: Yes, thanks. Yeah, so Mike, let me shed some light on this new feature. What is it about?

Mike: All right. So when you’re configuring a campaign—this is alpha only, so don’t go looking for this right now, you almost certainly don’t have access to it—I know just from screenshots floating around on LinkedIn that this exists.

00:02:30 - 00:03:00

Mike: Let me read to you exactly how Google describes this: “Adjust channel settings if you value a channel more or less based on its importance to your business.” So this is a little different than the way Demand Gen is structured, for example. That has channel controls, but it’s literally just tick boxes. It’s binary.

00:03:00 - 00:03:21

Mike: You’re basically opting in or out of these channels. A while back—we talked about this months ago on the podcast—Google added an opt-in or opt-out for Search Partner Network and Google Display Network. This is something else. Functionally, it’s something else. You cannot just eliminate your spend on these channels. There’s no guarantee of that. You’re not adjusting your budget.

00:03:21 - 00:03:56

Mike: What they do is per channel, they give you this little slider, and the slider tells you to adjust how much you prioritize conversions from a given channel, such as YouTube.

Chris: And there’s an important statement, if I may.

Mike: Please.

Chris: Yes, there’s this slider. You basically can slide the dot from lower importance to high importance.

Mike: Yes.

00:03:56 - 00:04:21

Chris: Google also states here: “Keep in mind that Performance Max automatically optimizes for total conversions across all channels. So adjusting these settings will directly impact your performance.”

Mike: Yeah, which makes sense on the surface. But there’s a contradictory statement when you go to the sliders.

Chris: Yes, because… all right, what does it mean to have an importance slider?

Mike: Weird concept, right?

00:04:21 - 00:04:42

Mike: When you look at one of these sliders, like Chris said, it’s on a spectrum from lower to higher. And there’ll be a default position which is called “Auto”. In the screenshot example that we’re looking at with the YouTube channel, YouTube is automatically at lower performance, pretty close to the bottom.

00:04:42 - 00:05:10

Mike: And I would venture to say that Google Search is going to be automatically at higher importance. And then you can tweak those. So what happens when you tweak importance? It says: “A positive adjustment relaxes this channel’s CPA cost fraction, while a negative adjustment tightens it.” I assume that this will later be adjusted to reflect ROAS as well, not just CPA. But it doesn’t matter—if you’re affecting one, you’re affecting the other implicitly anyway.

00:05:10 - 00:05:37

Chris: However, here is where the contradicting statements come in.

Mike: Yeah, exactly. There’s another statement at the bottom, which is: “Google will optimize performance on YouTube to reach your desired campaign-level target CPA of $150.”

Chris: $150, yeah.

00:05:37 - 00:06:00

Chris: So how does this make sense? At the top, they said that if you adjust these settings, you’ll impact your performance. And then in the slider, it says that it’s going to harmonize towards your target, which are opposite statements.

Mike: Kind of, yeah.

Chris: Look, let me dissect it. I’d love to roleplay here.

00:06:00 - 00:06:21

Chris: The feature is out, we can sign up for the alpha, and I do the following: I increase the YouTube slider, which is auto-set on the lower spectrum of this slider. I increase it to the maximum. I say it’s of highest importance.

00:06:21 - 00:06:53

Chris: What will actually happen then? Because the overall target CPA is not changing, Google will optimize towards that. In my understanding, if I slide the slider to the higher spectrum, it will tighten the ROAS or CPA targets—meaning I will create less revenue and conversions with this channel, I assume, because there are fewer conversions available at a tighter ROAS goal.

00:06:53 - 00:07:23

Mike: Well, no. A positive adjustment would relax it.

Chris: Relax, sorry! Yeah, anyway…

Mike: Okay, actually we have to redo that.

Chris: We do that, okay. One second.

Mike: You have an excuse—three episodes in, you get fuck-ups. You’re not going to get through without a fuck-up.

Chris: My grandpa always said, “There’s no rose without a thorn.”

Mike: In English, you’d say, “Every rose has its thorn.”

00:07:23 - 00:07:45

Chris: So let me roleplay, Mike, because the statements don’t add up.

Mike: Yeah.

Chris: Let’s say I take the YouTube slider and I move it up to the higher spectrum.

00:07:45 - 00:08:18

Chris: Let’s keep in mind, Google will always optimize towards my overall campaign target. But if I move it to the higher spectrum, it means it will loosen up the CPA/ROAS targets for that specific channel, meaning I will probably create more revenue and more conversions at a lower ROAS level than a higher CPA target. So there will be more revenue coming from YouTube at lower ROAS levels.

00:08:18 - 00:08:47

Chris: And at the same time, the overall campaign target stays the same. What does that mean? Is there less revenue coming from other channels? I don’t understand what it actually does or how I can really work with it. Help me out here. It just doesn’t make sense to me.

Mike: I agree. I think some of these will need more documentation from Google over time.

00:08:47 - 00:09:14

Mike: Bear in mind that it’s alpha. But let me take that YouTube example. YouTube—a lot of YouTube is really not conversion-oriented. Or it’ll be view-through conversions. CPA is often quite high on YouTube because it’s more reach and engagement-oriented and upper-funnel.

00:09:15 - 00:09:43

Mike: And so, yeah, if we look at ROAS, the ROAS will be low there. That tells me that you would worsen your overall campaign performance if you try to purposely invest more into that or tweak that, because it’s definitely not going to have comparable performance to the core Google Shopping part.

00:09:43 - 00:09:52

Chris: Exactly.

00:09:52 - 00:10:20

Chris: Higher ROAS… but I talked about if I slide it to the higher spectrum, it will create lower ROAS, right? Did you say higher ROAS earlier? That’s all right, sorry. I actually slept eight hours, but it’s crazy because yesterday I slept so good and had an awesome day, and last night I slept poorly.

00:10:20 - 00:10:49

Mike: All right, so the blooper reel is like three minutes long. Okay, we’re good.

Chris: Sorry. So Mike, let’s do a roleplay here.

Mike: Let me… let me do it.

Chris: Let me do a roleplay, Mike, because for me, the statements don’t add up, and I have a hard time thinking on operational levels about what I do with this feature.

00:10:49 - 00:11:14

Chris: Let me assume I move the YouTube slider to the higher spectrum. What does that mean? It will probably mean, as far as I understand Google’s statements, that it will create more revenue because I’m loosening my ROAS targets for YouTube. All right, so I get more revenue probably at lower ROAS levels.

00:11:14 - 00:11:37

Chris: At the same time, my overall campaign target stays untouched, and globally Google is optimizing toward that campaign level. It’s hard for both of those things to be true, actually, because what does it mean if I get more revenue at lower levels from YouTube? They have to tighten up in all other channels.

Mike: Yeah, exactly.

00:11:37 - 00:11:59

Chris: But why would I do this in the first place?

Mike: You know, I don’t even think YouTube is capable of delivering that much more volume. You could end up shifting away some of your spend from somewhere else to there, and it will come at the expense of that. Speaking from experience, YouTube definitely has higher CPAs.

00:11:59 - 00:12:24

Mike: And, you know, you’re probably talking about view-through conversions. A lot of YouTube inventory is based on a CPM basis, not even a CPC basis, because it’s more reach and engagement-based, so it doesn’t typically deliver. The conversions that do come are much more through remarketing. So if anything, you’d also be telling it to do more remarketing at that point.

00:12:24 - 00:12:50

Mike: So it is a bit mind-blowing for me to think through the tactical considerations and what would happen when you’re moving multiple of these sliders at once. I think you could end up messing up your campaign a bit, for sure.

Chris: And again, for me, I think then it gets smart and multi-dimensional if I change the sliders for each and every channel.

00:12:50 - 00:13:11

Chris: But let me assume I’m just changing the slider for the YouTube channel. There has to be a tradeoff with regards to the other channels, which I can’t predict and I don’t understand.

Mike: Yeah. You’d have to be able to have more budget become available or something to even that out.

00:13:11 - 00:13:43

Chris: Okay, so this is not clear at all.

Mike: No, not yet. There will be more documentation. We have to bear in mind it’s in alpha. But Chris, a very common use case will be that many people are just going to take these channels like Maps, YouTube, and Display Network and drag them lower because they’re trying to make it more like Shopping or focused on Search…

00:13:44 - 00:14:07

Chris: Yeah, on Search and less on remarketing and stuff like that.

Mike: Because this would be my next question to you—and you are my… you know, I always call you the smartest e-commerce person I know. What use case could I have, since it states “according to your business importance”?

Chris: Yeah.

00:14:07 - 00:14:28

Chris: What use case is there where I say, “Hey, YouTube revenue is more important to me because there are more new clients in there”? What use cases could you think of where I make this decision?

Mike: Interesting question. I mean, I guess if you feel confident in your assets and you’re wondering why it isn’t spending more on YouTube, that could be an example.

00:14:28 - 00:14:53

Mike: If you’ve got brick-and-mortar retail stores, maybe you want to see Maps be a bit more prominent or stuff like that. But the cool thing here—or funny thing—is people demanded channel reporting for years and tried every kind of workaround, and as soon as Google made it available, no one was interested anymore.

00:14:53 - 00:15:11

Mike: I swear to God, no one was interested at all anymore. People looked at the report a couple of times, said “That’s neat,” but there was nothing they could do about it. Now there is something you can do about it, which is a good thing. And because there’s now a time-series view available for the channel report, you’ll even be able hypothetically to see the impact of changes you make on the sliders.

00:15:11 - 00:15:35

Mike: But I think the use cases are very dependent, and probably for most people, the most popular use case will be to make PMax campaigns less like PMax campaigns, which is funny. Which maybe is a good segue to the statement from Jyll and others.

00:15:35 - 00:16:10

Chris: Yeah. Before we talk about the implications of what’s to come with regards to PMax and ad platforms in general, I want to come back to it because on a tactical level, I would be so nervous with the slider thing. It’s so arbitrary. I love “high/low”, but it feels like you’re doing heart surgery because you’re interfering with the bidding algorithm, right?

00:16:10 - 00:16:40

Chris: And I don’t understand the impact and tradeoffs with other channels.

Mike: We’re really working from a screenshot here, and as soon as I get access to this… because what I’m desperately missing are numbers on that slider. That’s the thing, desperately.

Chris: If this is really… bear in mind it’s alpha, but if it plays out like it feels right now…

00:16:41 - 00:17:05

Chris: Man, maybe I’ll step on some Googlers’ toes here, sorry for that, but maybe this is a placebo. It gives you the feeling that you can do something.

Mike: Yeah, that’s a good one. That’s bleak, but possible, because these multipliers could have an extremely large effect or an extremely minimal effect depending on the numbers, and you just feel good that you slid the thing down.

00:17:05 - 00:17:29

Mike: Wow, that’s quite a thought, Chris. But all right, Google is in the slider game. Let’s find out. It’s certainly interesting how this whole thing will play out. It might give you massive control, or it’s just a placebo. We’ll find out.

00:17:29 - 00:17:53

Mike: It’s a big feature, it is. And yeah, it definitely makes these… I think we’re going to see interest in that reporting side spike because there’s action here now. And maybe Google wants that because channel reporting—shoutout to Google—is a hell of a thing. It can be done better, of course, but you know, they didn’t want to do any of this, and they did it.

00:17:53 - 00:18:12

Chris: They did it because the market wanted Google to do so. So let’s see how this whole thing will play out. But Mike, the implications might even be bigger, because I can remember—and by the way, you most certainly helped me out with the content here—I can remember at a quite big convention when PMax was quite fresh, we talked about what PMax is and how this whole thing will limit your control.

00:18:12 - 00:18:36

Chris: I can remember the talks like it was yesterday. One thing was for sure: the original plan with PMax was certainly not playing around with channel controls.

Mike: No, definitely not.

00:18:36 - 00:19:05

Chris: It was this one-stop-shop campaign type which does all the relevant things for you without taking control—or rather, by taking control away from you. That was the major takeaway. Now, a couple of years later, this whole thing comes across completely different. My question to you is: is there substance to the statement that this is another proof that the original PMax idea or experiment has failed? Because that’s quite a statement.

00:19:05 - 00:19:42

Mike: Well, there’s substance to everything Jyll says. It’s fascinating. I remember back then—because I always cited this—my favorite definition of PMax came from the API. They described it as a unified buying service for all of Google Ads, which was a great way of describing it, actually.

00:19:42 - 00:20:05

Mike: A one-stop shop. And what they’ve done now—because this is still a compromised solution, not like Demand Gen where you can create fully customized campaigns in that way—is it’s still a unified buying service, but they’re exposing the weights to you and letting you play with those weights. And they’re exposing insights to you as well.

00:20:05 - 00:20:24

Chris: The insights that you need.

Mike: Yeah, but the weights are still opaque because as of now we just have this low-to-high scale. But it’s interesting—you can see the default weight of YouTube in your campaign and everything else, which was certainly not planned back in the day, I assume.

00:20:24 - 00:20:49

Chris: Not back then, because this was presumably all backend stuff.

Mike: Yeah. But has the project failed or has it evolved? It’s hard for me to say. I do kind of lean toward “failed”. But I can argue both sides.

Chris: Argue what you feel!

00:20:49 - 00:21:16

Mike: Well, on the failed side, we’ve talked about this in the past: I feel like PMax is starting to lose its identity a bit. There is Demand Gen for people who want asset stuff and a very customizable asset and feed. Besides Search Network, there is Demand Gen, and it is very flexible. And Google touts it as performance-oriented demand generation.

00:21:16 - 00:21:43

Chris: Yes, it’s rather mid-funnel.

Mike: Yeah, they’re increasingly bringing it down-funnel, which is eating into the positioning of PMax. It increasingly overlaps in terms of inventory, performance claims, and feeds with PMax. And for the people who never want that and just want to run a damn Shopping campaign…

00:21:43 - 00:22:16

Mike: Well, Shopping has been pulled out of the fire. There’s no doubt about that. It’s been upgraded with AI Max, so it’s going to have feature parity with Max for Search Network. So these are the things that make me feel like PMax has lost its identity. But if you believe in the Myers-Briggs personality test, I’m a P (perceiving) rather than a J (judging). I hate making decisions unless I know I have to.

00:22:16 - 00:22:39

Mike: Decisiveness is a virtue, but from my inherent character, I hate it and always see two sides to everything. PMax was originally meant to be a one-stop shop, dirt simple. It’s actually complicated as hell now. It’s super cool, but it’s very tactical as well.

00:22:39 - 00:23:04

Mike: If we can better learn how to use these tactically and strategically, it’s in many ways a more interesting campaign type than ever.

Chris: If we’re very objective about it—and I’m not judgmental here…

Mike: You can be judgmental! I’m going to let you.

Chris: Okay, I will be. But I will start with an objective observation: Smart Shopping to its core was probably…

00:23:16 - 00:23:44

Mike: …more automated than PMax is now.

Chris: Yeah, because with Smart Shopping, you couldn’t do shit. We didn’t like that campaign type, but it was working. It was a great campaign type. But if you compare PMax to Smart Shopping now, it’s the second coming of the campaign type with the most control layers you could ever think of, especially given where it started.

00:23:44 - 00:24:10

Chris: You can literally control almost everything again. I know there’s device, demographics…

Mike: Shoutout to something Fred Vallaeys said years ago. He used to work at Google, and he said Google always tries to heavily automate, and then they inevitably find that they have to roll back the automation. They take a big step, roll back, take a big step, roll back.

00:24:10 - 00:24:30

Chris: Fair enough. Maybe this is what we’re looking at right now. So coming back to Jyll’s statement, I think it’s not a failed experiment, but I don’t think it’s an organic evolution either. I think it’s a forced evolution.

00:24:30 - 00:24:54

Chris: And by forced evolution, I mean PMax is the one campaign type next to AI Max for Search which will serve all these new AI search surfaces. So I think they can’t roll back in terms of sunsetting PMax. That would be a cost calculation that doesn’t make sense. At the same time, they saw that adoption rates were stalling.

00:24:54 - 00:25:16

Chris: So I think it’s not failed, because PMax is a great campaign type, but it’s a forced evolution. Nothing is done inherently for the sake of it—it’s forced because adoption rates were stalling.

Mike: Yeah, the market demanded it and Google listened to the market. So I think there’s no bad thing in it.

00:25:16 - 00:25:38

Mike: But this control is structured weird. It’s a concession or a compromise. It’s probably not structured the way you or I would like it, but it’s structured in a way that Google can live with it and advertisers can too.

Chris: I don’t think there’s anything salty about it. Google listens to the market, and the market gets what it wanted.

00:25:38 - 00:26:12

Chris: It’s a coexistence, and both worlds can live with it because I don’t know any retailer who is super critical about PMax anymore. The things Google has shipped and has been working on are very, very good for the advertiser, and PMax has landed in a positive way.

Mike: Yeah, the mood has totally changed.

00:26:12 - 00:26:33

Chris: Which I think is very important for Google, because PMax plays a critical role for the future they’re heading towards. So I think there’s no bad thing in it, but it’s certainly a compromise to what the original idea of PMax was.

Mike: Yes, definitely.

00:26:33 - 00:26:57

Mike: I think there have been changes in management at Google. I think also Google’s market position relative to Meta, for example, has shifted, and they’re willing to make concessions now. We have to remember this lesson and bring it to whatever the next fight will be, like getting visibility into AI surfaces and stuff like that.

00:26:58 - 00:27:33

Chris: But Mike, the question then is… I think you were one of the first who said that PMax is not just a campaign type—a big Google campaign type—but probably something the whole industry will work towards. You called it “the Max”.

Mike: Yeah, exactly.

00:27:33 - 00:27:57

Chris: My question is now: might this be not just a reaction to what the market is demanding, but from a product management perspective, a strategy of creating differentiated value compared to campaign types on other platforms? Google is saying, “Look at that black box over there with Advantage+.” By the way, that campaign type is going nowhere. So Google is trying to create a value proposition that you have enough control here, and it’s not a black box. Do you think that might be a strategy on Google’s side?

00:27:57 - 00:28:25

Mike: Definitely. Google has always been a market leader, not only in terms of ad revenue, but in terms of platform features and maturity. They pioneered one thing after another with smart bidding, automated placements, automated creatives, and automated targeting. They’re increasingly focused now on automated budgeting.

00:28:25 - 00:28:53

Mike: What I used to talk about is this idea that the whole ad platform ecosystem is connected—it’s a whole category of technology. Analyst Eric Seufert says “Everything is an ad network.” That’s his famous slogan. And it’s true, because there are ads anywhere ads can possibly be. We live in an ad economy.

00:28:53 - 00:29:19

Mike: I started saying “Everything is a Max” because that’s what happened next: Meta had their PMax with Advantage+, Reddit bought a company to build them a Max, Pinterest had a Max, TikTok had a Max, Amazon had a Max. Everyone had their version of Max. For two years, this was a major trend. Every platform was building their Max.

00:29:19 - 00:29:49

Mike: And if Jyll’s theory is correct, the market kind of refuted this thesis or category of technology that everyone followed.

Chris: So this has big implications, because I think we will see this trickle down. Google ends up in this role model position.

Mike: Oh yeah, I forgot Microsoft has Max—they even call it Max!

00:29:50 - 00:30:25

Mike: Every other platform will probably be forced to go in this more transparent direction.

Chris: From that perspective, Jyll’s statement is probably absolutely true, because the original idea of the whole ad ecosystem has certainly failed if Google is the role model again and others are forced to adopt what Google has done.

00:30:25 - 00:31:00

Mike: Exactly. If Google couldn’t make this work, no one else can. Look at Meta—we talked a couple of months ago about how Advantage+ adoption flatlined way lower than PMax adoption. It’s free-falling. The idea that you can just automate everything—Mark Zuckerberg would like nothing more than you just giving him a target and money without caring about anything besides one single number—will not be accepted anytime soon.

00:31:00 - 00:31:22

Mike: This will be increasingly interesting to see as AI gets involved in new ways. AI has been in these platforms forever, but as these new generations of AI get involved, to what extent will this be accepted by the market? How inevitable is it really?

00:31:22 - 00:31:51

Chris: I think we came closer to the statement, because from that perspective, the experiment to automate the whole ad ecosystem has failed with high probability.

Mike: Yes.

Chris: If Google can’t force it down your throat, no one else can. Then it’s a failed experiment.

00:31:51 - 00:32:14

Mike: Because Max today… at one point I was saying “PMax 2025 is not your granddaddy’s PMax,” but PMax 2026 is radically different than how it was launched.

Chris: I can’t believe it, but we might end this episode on a positive note again. Did we do that two weeks in a row?

00:32:14 - 00:32:36

Chris: We should certainly think about that!

Mike: I told you about those supplements, Chris. Maybe they’re taking effect on me.

Chris: Man, I tell you one thing: the positive thing about this failed PMax experiment is that the advertiser who works with the right partner or has the smarter people in their teams can make a difference again.

00:32:36 - 00:33:06

Chris: Because I can remember that was the biggest fear—your analogy of two Formula 1 cars driven by the same AI, who is going to win? That question was super valid, and I think the version we talk about today gives you the freedom and levers again to outperform your competitor. That’s what I love about it.

00:33:06 - 00:33:33

Mike: Yeah, definitely. It was hard to stand out or use this better than anyone else. There was no alpha or edge. Now that opportunity starts to emerge, because these sliders, for example… they’re designed so an idiot could use them, but an idiot will use them. I think you have to be pretty smart and pretty careful to use them correctly.

00:33:33 - 00:34:00

Mike: You need good data and control, because if you adjust Slider A and B at the same time, you’d better look closely at it. If it’s not a placebo, you’d better be watching it.

Chris: This is very, very positive, and I’m absolutely with you. I can’t imagine a world in which Microsoft, Meta, you name it, are not following.

00:34:00 - 00:34:26

Mike: Microsoft was already trying to be a more transparent, better picture.

Chris: I think it’s good. And about this feature, we will keep an eye on it as always. Let’s see if this slider game is really a fun game. It looks fun—everyone loves to drag sliders around randomly! I love sliders, I love it.

00:34:44 - 00:35:16

Chris: Yeah. Mike, wrap it up.

Mike: Sounds good, Chris. In a way only Mike Ryan can wrap it up… This has been another episode of Growing Ecommerce. Thank you as always for listening. If you want to learn more, visit us at smec, and please leave us a review or recommendation, hit the like button, or move the slider upwards to a higher importance and accept a lower ROAS to see more of us!

00:35:16 - 00:35:27

Chris: I would love that. Yeah, thanks. We’ll see you next time.

Mike: See you soon, Mike—thank you very much.

Chris: Thank you, Chris. The slider game—three in the books!

Mike: In the books, Kevin. Actually…