Ep 506: The PMAX Reset Strategy That Delivered a 37% ROAS Lift | AKNF

Summary

DTC Podcast Episode 506: The PMAX Reset Strategy That Delivered a 37% ROAS Lift | AKNF

Release Date: May 9, 2025

In Episode 506 of the DTC Podcast, hosted by DTC Newsletter and Podcast, listeners are treated to an in-depth discussion on optimizing Google’s Performance Max (PMAX) campaigns to achieve significant returns on ad spend (ROAS). The episode features Zaviar ("Zav") from Pilothouse's Google Team, who shares valuable insights and a compelling case study on revitalizing PMAX strategies.

---

1. Understanding the Challenges with PMAX

The conversation kicks off with Zav addressing the common frustrations brands face with PMAX campaigns. He describes PMAX as a "black box," making it difficult to decipher what elements are driving success or causing stagnation.

“Pmax, it's kind of a black box. It's hard to know what works and what doesn't...”
[00:00]

Zav elaborates on the typical lifecycle of a PMAX campaign, noting that initial months often show promising results, which then plateau as cost-per-clicks (CPCs) and conversion rates decline. This stagnation often results from irrelevant traffic and high bounce rates, hindering the campaign’s effectiveness over time.

---

2. Strategies for Resetting PMAX Campaigns

When faced with declining performance, Zav advises taking decisive action:

“The first thing to do is turn your campaigns off. Five to seven days is probably the best time...”
[04:09]

He emphasizes the importance of resetting the bid strategies to prevent campaigns from entering a "death spiral," where poor performance data perpetuates declining results. By pausing the campaign temporarily, advertisers can allow bid strategies to recalibrate.

---

3. Case Study: Revitalizing Benji Sleep’s PMAX Campaign

Zav shares a success story involving Benji Sleep, a Canadian client. Initially, Benji Sleep's PMAX campaign encompassed all product categories within a single asset group, leading to overemphasis on their best-selling sheets and neglect of other products like towels and duvets.

“We have one client specifically here in Canada, Benji Sleep, who basically through kind of a reinvention of their PMAX strategy... continued that growth...”
[02:08]

By segmenting the campaign into distinct asset groups for each product category and separating branded traffic from generic traffic, Pilothouse was able to achieve a 37% ROAS lift. This restructuring allowed for more precise targeting and better allocation of the advertising budget.

---

4. Debunking the Myth of Audience Signals

A significant part of the discussion revolves around the "myth of audience signals." Zav clarifies that in PMAX, audience signals serve more as guidelines rather than strict boundaries.

“Just because you've put an audience on your campaign doesn't necessarily mean that the only people that are going to be targeted are within that audience.”
[10:04]

He suggests broadening audience signals to include a wider range of relevant categories and refining them to better reflect the ideal customer profile. This approach allows PMAX to extrapolate and identify similar high-potential audiences, enhancing campaign effectiveness.

---

5. Feed Optimization and Strategic Manipulation

Optimizing the product feed is another crucial strategy discussed:

“Make titles that are more broadly relevant to different searches... use custom labels to help identify bestsellers, new arrivals, bundles...”
[12:16]

Zav outlines best practices for structuring product feeds, including:

• Broadly relevant titles to capture diverse search queries.
• Detailed product descriptions and custom labels to highlight key attributes.
• Supplemental feeds to include additional product data, improving ad relevance and placement.

These enhancements ensure that shopping placements remain a strong component of PMAX campaigns, driving profitability by targeting users further down the funnel.

---

6. Campaign Structure Before and After PMAX Reset

Before the reset, Benji Sleep operated a single, unsegmented PMAX campaign, making it challenging to manage and optimize effectively.

“We had one campaign with everything in it... it was difficult for us to not only define specific audiences against specific products...”
[14:58]

Post-reset, the campaign structure was reorganized into multiple segments:

• Two primary campaigns: one for brand and another for generic traffic.
• Separate asset groups within each campaign: dedicated to towels, sheets, duvets, etc.

This refined structure allowed for better control over budget allocation and more precise targeting, leading to improved performance metrics.

---

7. Achieving Remarkable Results

The impact of the PMAX reset strategy was profound. Zav highlights the following achievements:

“Our ROAS was up 37% almost immediately. Branding click share was down about 20%, and the volume of conversions more than doubled.”
[19:10]

Key outcomes included:

• 37% increase in ROAS
• 20% decrease in branding click share, enabling more budget allocation to prospecting new customers.
• 2.2x increase in conversion volume
• Slight improvement in cost per acquisition (CPA)

These results underscore the effectiveness of strategic campaign restructuring and targeted optimization.

---

8. Leveraging Scripts for Enhanced Performance Tracking

Towards the episode’s conclusion, Zav introduces advanced techniques for tracking PMAX performance using scripts:

“There are scripts that you can use that will basically track that data for you... I can include some links to some of my favorite, my personal favorites definitely in the notes for this.”
[20:50]

These scripts provide deeper insights into where ad spend is directed, such as specific placements like Gmail or Display Ads, enabling advertisers to make more informed decisions and further refine their strategies.

---

9. Final Takeaways

Zav wraps up the discussion by emphasizing the necessity of proactive management in PMAX campaigns:

“It's something that you need to push in a direction and keep focused... it usually ends up in kind of a stagnation of results and it gets lazy.”
[18:45]

He advocates for continuous optimization and strategic adjustments to maintain and enhance campaign performance, rather than adopting a "set it and forget it" mentality.

---

Conclusion

Episode 506 of the DTC Podcast provides a comprehensive guide to overcoming the inherent challenges of Google’s Performance Max campaigns. Through expert insights and a tangible case study, Zav demonstrates that with strategic restructuring, audience refinement, and detailed feed optimization, brands can significantly boost their ROAS and overall campaign effectiveness. This episode is a must-listen for direct-to-consumer brands seeking to maximize their ecommerce marketing efforts on Google.

For more tactical insights and step-by-step strategies, subscribe to the DTC Newsletter or visit directtoconsumer.co.

---

Notable Quotes:

• “Pmax, it's kind of a black box. It's hard to know what works and what doesn't...”
  — Zav, [00:00]

• “The first thing to do is turn your campaigns off. Five to seven days is probably the best time...”
  — Zav, [04:09]

• “Just because you've put an audience on your campaign doesn't necessarily mean that the only people that are going to be targeted are within that audience.”
  — Zav, [10:04]

• “Our ROAS was up 37% almost immediately...”
  — Zav, [19:10]

• “It's something that you need to push in a direction and keep focused...”
  — Zav, [18:45]

---

For more information on Pilothouse's services and how they can help optimize your PMAX campaigns, visit Pilothouse.co.

Transcript

[Zav]

Pmax, it's kind of a black box. It's hard to know what works and what doesn't or why you start your campaign and everything's good usually for like three to four months. After that time, you'll notice like a real stagnation in results. CPCS plummet, conversion rate plummets. You're getting a lot of traffic that's just like either not relevant or it's bouncing. We have one client who basically through kind of a reinvention of their pmax strategy, were able to get over that hump and kind of continue that growth that they saw in the very early days. If you're noticing that sort of stagnation, whether that be in your ability to acquire conversions or in your ability to just drive whatever results you're hoping your pmax campaign to drive, I would say the first thing to do is to.

[Eric Dick]

You'Re spending real money to drive traffic to your site. But the truth is 98% of that traffic leaves without ever giving you an email. Meta knows who they are, but you don't unless you use Instant. Instant helps brands identify anonymous visitors so you can send more abandonment emails, build stronger retargeting audiences, and actually recover that lost revenue. Brands like Fate, the label saw a 4.5x boost in their abandonment flows. Massco increased Shopify revenue by over 15%. Third Love, Toys R Us and the Oodie. These are all massive stores that use instant to see an average ROI lift of 11.2x. They'll even guarantee you a 4x ROI or your money back. It only takes 15 minutes to set up and you'll start seeing results fast. If you're serious about getting more out of the traffic you're already paying for, now's the time to fix it. Get 50% off your first month at Instant One DTC. That's Instant O N E DTC. It's all killer, no filler. It's a beautiful Thursday morning here and I'm here with Zavir from Pilothouse's Google Team. Welcome back to the podcast, Zav, how are you?

[Zav]

Good. How are you, man?

[Eric Dick]

I'm pretty good, Pretty good. Today we're talking about it. We seem to be talking about on the Google site all the time, but I think it's a really good topic. We're talking PMAX strategy. You've got a really interesting case today. Walk us through what, what you're feeling with pmax these days.

[Zav]

I feel like what a lot of people are feeling, which is mild frustration. You got pmax. It's Kind of a black box. It's hard to know what works and what doesn't or why you put your assets in. You basically start your campaign, you get your tagging going and you're off to the races and everything's good usually for like, I would say like three to four months. After that time, often you'll notice like a real stagnation in results. So like whether that depending on how your PMAX campaign is set up and like what you're selling and who you're selling it to, this can look a little bit different. But it's usually like CPCS plummet, conversion rate plummets, you're getting a lot of traffic that's just like either not relevant or it's bouncing. And whatever you were tracking, whatever you were hoping to get out of your Performance Max campaign, whether that was, you know, purchases, signups, they seem to to stop happening. I found this to be true almost across the board with accounts at that, and again it varies at that to like three to four month point. It's kind of like the first time that they experienced this. And I've tried a lot of different ways of dealing with it, but we have one client specifically here in Canada, Benji Sleep, who basically through kind of a reinvention of their PMAX strategy and kind of a total refresh basically of the Google Ads account, we're able to get over that hump and kind of continue that growth that they saw in the very early days of their new Performance Max campaigns. And this can be really relevant to anyone who has a brand that has existed for some time, especially in the Google space, but is launching, whether it be like a new product or a new lineup, or trying to enter a new space where they otherwise might not have been, it's a really good way to try to continue to grow those campaigns basically.

[Eric Dick]

And basically I haven't actually run a PMAX campaign myself, but you don't have the same levers. You can't sort of like they're harder to tinker with to try to get them back on track. Is that accurate?

[Zav]

But there's a few things that are happening and like this, this is totally a part of the black box scenario. But the first thing is that like it's hard to know when you dump all your assets into an asset group, it's hard to know which things are performing specifically. Like you can know, you know your roas, you'll know how many clicks, you'll know how much revenue that came through that campaign, but you won't necess necessarily know exactly what drove it. And even more importantly, you don't know where it's being driven from. Right. Because Performance Max operates across, I mean, pretty much every Google possible placement, you know, Gmail, YouTube, search shopping, and a bunch of other placements that don't even necessarily have names all the time. I would say the first thing to do is to turn your campaigns off. And like, I know that can be really painful for a lot of brands who are basically operating in a system that needs Performance Max to generate volume. But I would say if you're noticing that sort of stagnation, whether that be, you know, in your ability to acquire conversions or in your ability to just drive whatever results you're hoping your PMAX campaign to drive. Yeah. First thing you really need to do is turn it off. Five to seven days, I would say, is probably the best time. And again, like, it totally depends what you're selling, who you're selling it to. But I found, like, regardless of the size of the campaign or kind of like the scope of it, five to seven days is usually where you want to start. It gives your bid strategies time to just kind of like reset themselves. One of the major problems why a lot of the time you're seeing that stagnation in results is because basically your campaigns enter this kind of like death spiral. And in the sense of like, they, they're using their historical data to basically extrapolate on to try to better understand how it can find and target and move new people through the funnel. Right. So if the data that it's getting isn't necessarily the strongest or isn't necessarily leading to the outcomes that you're hoping for, and it's continuing to see those sorts of outcomes, it's gonna basically feed off of that and you're never gonna be getting the kinds of, the kind of signals, the kind of like customers basically that you're hoping to get, just because those people don't exist in a system that's being trained basically not to find those people through passport results. And so, yeah, the first thing that you need to do, I think, to try to combat that is just turn it off. Yeah, five to seven days is where to start. And then a restructuring of the campaign works really well. For Benji Sleep, what that looked like was basically taking a couple of campaigns that were segmented kind of like by generic and brand, which is how we generally will segment those campaigns, and then segmenting it further. And what, what that's going to do is it's going to give you more information on how People are finding you basically, and then what's effective in the campaign, Right. So if you have, if you're selling, you know, towels, sheets, duvets, linens, as Benji Sleep does, and you have all of those things in one campaign, in one asset group, it's really, really hard to know what's working and what isn't if you can segment those things. And now depending, again, the segmentation is totally going to depend on like what it is you're selling, who you're selling it to. But like, a lot of the time that segmentation can look like, as simple as, like across your products. So, you know, sheets in an ad group, duvet is in its own situation, towels in its own situation. But it could also look like segmenting it by your customer type. So like, for example, if you're selling a product to a specific age group or gender or affinity category, maybe you want to segment your campaign into asset groups in that way. That way you can better understand how the signals that you've put into all of those asset groups are actually affecting the results broadly in your campaign. Now, it's still going to be difficult to know exactly what's happening across all of the different assets in that group, but you'll be much better prepared to kind of like extrapolate on positive data.

[Eric Dick]

And by concentrating the better performance, you're giving it better signals to build off of rather than the weak signals that were causing the downward spiral.

[Zav]

Exactly. And then, then one thing that you'll also be able to do is you'll have a better. You'll be able to better understand where the poor performance is coming from. So you can kind of nip those things in the bud when they do start to show themselves a lot more easily. Because again, you'll just have a better understanding of how everything is broken out inside those campaigns.

[Eric Dick]

I was just wondering like, so for the turnaround in this case, was it one of the product categories that was bringing things down or was it. It really was, it really was.

[Zav]

So basically we try to sell, we sell things basically not just across product catego, but as bundles as well. Basically what's happening is performance Max will try to take the path of least resistance in any sort of advertising that it's doing, like whatever. The easiest way for it to basically come by a result that you're hoping to get is the way that it's going to go. So if you're dumping everything in one campaign, you might find that it's overemphasizing one particular product, which, I mean, Benji, their Core product is sheets. Right? So if we have sheets, towels, duvets, all in one asset group, and sheets is by far the most popular, that's what it's going to sell. Sell all the time, right. And then it's going to neglect a lot of other things. So it's not so much that a particular thing is underperforming versus others as much as that we're kind of overextending ourselves on a particular product and neglecting our overall offering. And so by segmenting further, whether you're segmenting by campaign or by asset group type, you're going to have a much better control over how you're spending across those groups and who you're targeting. And then, of course, you'll have a better understanding of the results across the different product categories or audiences or however your segments in your campaigns.

[Eric Dick]

Super cool. What do you mean here by the myth of audience signals?

[Zav]

Yeah, so basically the myth of audience signals is like, I think we've been trained by meta, by, you know, Reddit, by all of these different platforms that we advertise on that basically you pick an audience, right? And then you advertise inside that audience. And. And that's basically the extent of it. But in performance max, it's not necessarily true. So, like, just because you've put an audience on your campaign doesn't necessarily mean that the only people that are going to be targeted are within that audience. The audiences in performance max are less of, like, strict groups that you're meant to be advertising to and more of, like, general guidelines that the campaigns can basically try to extrapolate on to get you the results. So I would say depending on the size of your audience, don't be afraid to kind of either make it larger than you think it might be through inclusion of just like a broader scope of people, whether that be new affinity categories that are kind of like tangentially relevant to whatever the product is, or whether that be like, including different signals from your website that you might not have already included. Like, a lot of times your asset groups are going to look like all visitors to your website or something like that. Or like, maybe if you're selling a particular product to like, all of the demographic that is available to you when you don't necessarily need to do that. You can also refine it a little bit more tightly to say, like, I'm really only looking for people who visited this particular product page or like, who've done this set of actions on site so that you can try to make sure that the people that you're targeting are representative of your ideal customer and less of kind of just like a blanket group of people who might be interested in the product. Pmax can then take that information, it'll extrapolate on it and try to find similar people as well. That's a relatively new thing that Performance Max does that I've had a lot of success with.

[Eric Dick]

Very cool. Nice. What do you mean by strategic feed manipulation? Is that about the products that you're serving it or is that something else kind of.

[Zav]

Yeah, actually that's, that's something that, that's a big proponent of it also. So like you want to make sure basically shopping campaigns or the shopping element of Performance Max campaigns operate by taking a feed of your products, right? All of the information, the attributes that are in there, it uses to serve across shopping placements and Google Ads. Depending on how you break out the things in that feed and the information that you provide, it can have a really, really, really strong or not so strong kind of impact on how much you're paying for ad for per ad and who's actually seeing it when they, when they search. Because as you know, like a shopping placement in Google Ads isn't necessarily tied to keywords, right? So it's looking, it's going to look at all of the things that you have associated with the products in your feed and, and try to basically serve people the most highly relevant things. So the things that you can do to try to improve your feed will be like making titles that are more broadly relevant to different searches, making sure that your product type descriptions, all of your different categories that are, that are available to you are well filled out and make sense. Making sure that you're using custom labels to help identify like bestsellers, new arrivals, bundles, if you're selling those, and making sure that you have a supplemental feed with any sort of missing product data that might help drive creative alignment. So that'll be things like seasonality, material, tags, what the product is made out of. Like basically any, anything that a person that might be relevant to a person searching is something that you should try to include because it's just going to broaden the way that people are finding your ads. And the more that you can do that, the more that you can make sure that shopping placements are kind of a prominent part of your campaign. Because shopping placements are so far down the funnel, it's ultimately going to help you a lot in terms of driving profitability and making sure that the Performance Max campaign isn't just following the path of least resistance, which is to say, you know, a ton of high level placements that don't necessarily result in the actions that you want people to take on site.

[Eric Dick]

I'm just curious, like from a high level, I guess you think you have a few more points you're going to, you're going to talk about pmax, but at some point I'd also just like to hear what the camp like paint me a picture of what the PMAX campaign looked like before and after. Is there are you going from one campaign that has a lot of stuff into it to several different PMAX campaigns that are all broken out along these lines.

[Zav]

So again that's kind of going to really, really depend on, on a few things. So maybe I'll back up. So to be something called the 3030 rule in Google Ads, right. The 3030 rule was basically to say it kind of still exists, not so much. But 3030 rule is basically to say that for every campaign that exists in your structure, you wanted to have about 30 camp 30 conversions in 30 days in order for you to use an automated bidding strategy. And as we know, I mean automated bidding strategies are probably one of the major wins of any, of any ad platform. So basically. And the 33 rule technically doesn't exist anymore. You don't need that 30 conversions to basically extrapolate the data to successfully use a bid strategy. But it's still a really good place to start when you're trying to define how a campaign should be broken out. So for Benji Sleep, the way that it was broken out initially was we had one campaign with everything in it. There wasn't a really strong segmentation between branded traffic and generic traffic. There wasn't a really strong segmentation between the different products in all of the groups. So everything was all in one group. Which meant that it was difficult for us to not only define specific audiences against specific products, but also to try to define how we're basically positioning each of the specific items in the inventory against the different audiences that we want to position them against. Right. So the way that it looked before, yeah, one big campaign, all of the stuff dumped inside of it with one broad sort of conversion goal set we had. We were using TROIS at the time and so it didn't offer us a lot of control, it didn't offer us a lot of ability to basically understand the results across the campaigns. The way that it looks now is two. Because of their volume, it's basically two campaigns. So there's one for brand, one for generic, and then within those there's a different asset group for every product category that they sell. So you have an asset group for towels, you have an asset group for sheets, you have an asset group for duvets. That allows us to basically make sure that we're using assets that are hyper relevant to each product category against that product category. And it also allows us to make sure that the search terms that we're using, the audiences that we're running against, are as relevant as they can possibly be to each of those product categories. And we can make sure that the signals that we're using to extrapolate on, again, are as relevant as it can be. We can be as specific as saying people who visited the towels page in the past, those are the people that we're targeting or people like that, as opposed to just saying, potentially anyone who visited the Benji website looking for any products in the past 90 days, which can, can sometimes be effective and depending on how, again, depending on the volume of your business can be super useful. But when you're looking at huge audiences against like really specific or really diverse product category, it's really, really helpful to make different asset groups for all of those. And if you have the volume and you really want even control, it can be useful to make different campaigns for each of them. Because budgets are set at the campaign level. Right. So even if you break everything out inside your campaign, you're not necessarily going to be able to always control the spend against each product category. But yeah, if you break it up by campaign, you can. Again, I think like that coming Back to that 3030 rule is a really nice place to start to understand whether you should be breaking things out by campaign or as different asset groups within a specific campaign.

[Eric Dick]

I love stories like this just because it constantly reminds us that AI is not ready to fully take over yet. You need to guide it, you need to put the guardrails on it, the handholds in the right space to make it as effective as it can be.

[Zav]

Yeah, yeah, it's definitely something that you need to push in a direction and keep focused. Again, that's kind of like set it and forget it attitude can be really useful and can be really time saving, but ultimately it usually ends up in kind of a stagnation of results and it kind of gets lazy, to be honest.

[Eric Dick]

Yeah, makes sense. And then so what sort of like improvement were we able to see by like rethinking this PMAX campaign?

[Zav]

Yeah, so basically by rethinking the PMAX campaign after, like post that off period when we made all of those changes. Our ROAS was up 37% almost immediately.

[Eric Dick]

That's wild.

[Zav]

Yeah. I mean, the branding click share was down about 20%, which means that we were able to spend more in prospecting new customers rather than just targeting people who are already aware. And despite being able to spend, despite spending more dollars on prospecting rather than branded, our CPA actually improved slightly and the volume of conversions that we got through those campaigns more than doubled. It was like a 2.2x. And so not only that, not only were we able to actually just improve the results at a very high level, we also were able to do so in a way that helps Benji and ourselves better. Like what was performing, when and how and why, which is really, really important to take back. When we're talking about trying to develop new assets, trying to launch new products, trying to enter new fields, all of those are like, really, really invaluable tools to try to, like, better understand the way that your business is positioned kind of in the Google search landscape.

[Eric Dick]

I love that. And it goes beyond just ads too. Right. You're getting product data about where to expand, how to represent those products to different audiences. By doing this, you get a lot more data that reflects on your whole business.

[Zav]

Yeah. And those are things that you can take to other platforms. Those are things that you can take, like, outside of the Google space and can really help you to kind of like navigate, dictate your next moves, like, potentially far beyond Google.

[Eric Dick]

Super cool. You've got another note here about using scripts to track performance in new ways. What are you talking about there?

[Zav]

Yeah, so basically this is a little bit more advanced and maybe not always necessary, but if you really want to have a good solid understanding of, like, how your campaign is performing across what strata, whether that be, like, how much is being spent on shopping placements versus Gmail, for example, or display. Now, technically, those things aren't always available to you in platform, but there are scripts that you can use that you can add to your campaigns that will basically track that data for you. So you can have a much better understanding of where is my money going, where is it coming from, and what actually is driving results. Inside my PMAX campaign, you can see information per asset, you can see information per placement type, you can see information per ad. And I mean, there's a bunch of them. They all work slightly differently. But I can include some links to some of my favorite, my personal favorites definitely in the notes for this.

[Eric Dick]

Nice. In the show notes. Perfect. Well, Zav, I think this is the perfect example of a all Killer no Filler episode. We're just coming up on 20 minutes here. This was all super valuable to anyone advertising on Google or pmax. Thanks for coming on, man. It was great to have you back.

[Zav]

Thanks, Eric. Good to see you again.

[Eric Dick]

And if you want to work with Pilothouse, you want to work with Zab, just. Just reach out Pilothouse Co. We're ready to break your PMAX campaign so we can rebuild them better. Just. Just give us a shout.

[Zav]

Yeah, thanks.

[Eric Dick]

Thanks for listening to today's episode. If you're not getting the DTC newsletter, you can subscribe for free. And if you want to learn more about Pilothouse's all Killer no Filler services, take off to Pilothouse Co. I'm Eric Dick, and this has been the D to C podcast. We'll see you next time.