Is ROAS Overrated? – Why the “Revenue Trap” Might Kill Your Google Ads Profit

Episode Transcript

Mike: Welcome, welcome back. Good to see you. Welcome back to all of our listeners. This is another episode of Growing Ecommerce brought to you by smarter ecommerce. I’m one of your hosts, Mike Ryan, and with me today is Christian Scharmüller. My name is, by the way, not as sexy as yours, Chris. I lose right there; it’s hurting.

Chris: Oh, that’s a region bias! You think Scharmüller is sexy for non-German speakers? No, that’s the point, so I will probably introduce myself with only Chris next time, because “Chris” is somehow fine. The struggles I’ve gone through… I’ve lived in Austria for a while now and every time that I have to explain my last name to someone filling out a piece of paperwork, looking me up in a customer database, or whatever…

[0:04:06] Chris: I just end up having to write it down on a piece of paper, and it’s only four letters! Come on, man.

Mike: Yeah, I love it. By the way, are you Mike or Michael Ryan?

Chris: I’m Michael.

Mike: Exactly, and I say “Michelle” so they know how to spell it. Then you’re safe. Exactly. So, I’ll be talking about names today? No, no. Well, you know, I’ve got my transition now.

[0:04:35] Mike: What’s in a name, Chris? “A rose by any other name smells as sweet.” Do you know this Shakespeare line?

Chris: I know it, but I’m not sure what the answer is, man. Okay, well, are you a Rose fan?

Mike: No, I have a long-documented history… I think I’ve talked about it on this podcast before as well. No, I’m not.

[0:04:57] Chris: I’m not sure if I am, but you know the “GOAT” discussions: Federer versus Djokovic versus LeBron.

Mike: That’s not even a discussion! It’s certainly true, but I think ROAS is right up there in the GOAT discussion for the greatest of all time KPI.

Chris: Yeah, for sure. I really think so. ROAS is there, probably, and maybe just between that and ACOS (Advertising Cost of Sale) from Amazon, which is just the inverse of ROAS. We call it cool.

Mike: So, that’s a perfect segue to the discussion we might have. There are two myths left that we didn’t tackle last time. For people just jumping in, last episode we wanted to start off this year—we’re a little late in the year—but for our late-rising start, we wanted to talk about five myths affecting ecommerce growth in Google Ads. We did three of them last time. Let’s speedrun these. The first one is the idea that ROAS is a North Star.

[0:06:07] Mike: Chris just mentioned it’s kind of the greatest of all time. If not the MVP, then certainly one of the most popular, most widely cited.

Chris: That’s for sure. By the way, we have to differentiate between the greatest of all time and the best of all time. I think ROAS is maybe the greatest of all time for a lot of emotional reasons, but I think it’s not the best of all time. I think there are better KPIs out there.

[0:06:29] Chris: For sure. But the myth we were talking about is basically: Is it the North Star of your campaigns? Let’s be precise here—what do we mean by being a North Star?

Mike: Well, I think the reason ROAS is so popular is that it’s just very easy to calculate.

[0:06:48] Mike: At least at face value, it seems quite easy to understand and you can apply it at all different levels. You might call it your Marketing Efficiency Ratio, but you can look at this across all your channels, per channel, per account within a channel, per campaign, and your campaign entities. You can calculate it all the way down to assets inside of Google Ads, and it’s completely independent of the time range you want to look at: hourly, monthly, weekly, yearly.

Chris: Exactly. So, there are advantages to it. And first and foremost, Mike, a lot of clients still use it as a proxy for profit.

[0:07:43] Mike: Exactly. And that’s the next thing. Google has encouraged us; they’ll often talk about your Return on Investment or they’ll talk about your profit, and they’re actually speaking about ROAS, which is crazy in itself. But let’s make an arc for shareholders too. ROAS has been very popular because it looks good in classic attribution. It looks good in bottom-funnel performance marketing channels, and ROAS has been a huge driver of spend and investment in these areas.

Chris: No, I think there’s an argument to be made that there are a lot of advantages coming with this KPI.

[0:08:19] Chris: That’s for sure. But let’s be critical of it now. What are the things you think are connected with ROAS in a negative way in terms of attitudes? It’s used wrongly, it’s interpreted wrongly—what comes to your mind?

Mike: I’d like to attack it constructively from two different angles. We are not out to demonize anything, but just to be critical toward it. How is it understood and for what purposes is it used? It gets used as a proxy for profit. Proxies are really valuable; we talked in the last episode about bringing your margin data in. People want to be more profit-oriented in their campaigns; they want their campaigns to be profitable. That’s fair. A proxy arises as an indirect measure. If you’re having a hard time directly measuring the profitability of a channel, then ROAS becomes the proxy. But something that can happen…

[0:09:36] Mike: …is that the proxy is always going to be imprecise. There’s always going to be some kind of air gap between that indirect measure and what you’re actually trying to measure. I don’t want to say it’s lazy, but it’s human nature that sometimes we forget about that gap and we let the distance collapse mentally. We just let it be a perfect stand-in for profit. That bothers me because it is a revenue-based metric in most implementations, not a profit-based metric.

[0:10:03] Chris: That’s definitely one part. Plus one; nothing to argue about.

[0:10:08] Chris: By the way, I don’t know if it’s laziness or if it’s sometimes really the lack of capability to have the profit data directly in your Google Ads ecosystem. It’s sometimes fascinating how tough it is for retailers. We have had some projects where we support our retailers to do so. So, maybe sometimes it’s really not laziness but lack of capability and also still the fear: “I don’t want my profit data to be shared with Google.”

Mike: Which I think doesn’t make sense at all. Google is not interested in the profit data, I assume. Not in the grand conspiratorial way that some people say. What should they do with the profit data of a mid-sized online retailer? And if they want to find out, they’ll find out anyway. If I were Amazon, would I submit it? No. Okay, Amazon is different, but for the long tail of online retailers listening to this podcast—let’s do it, there’s no risk.

Chris: It depends, but you already know the range of gross profit margins. You can look at the distribution per category and you can get a pretty close approximation. There are ways, by the way, to modify it.

[0:11:24] Chris: But we’re getting a bit away from the topic. Thank you for moderating this! This is something which I don’t get, man. I don’t get so many things, but this is one of them. Alright, back to us. Another thing I see, which bothers me at least sometimes, is that clients don’t understand the massive impact of the universal law of diminishing returns.

Chris: ROAS, as simple as it is, they look at the campaign and say, “My ROAS is great, it’s at 600%.” And sometimes there’s a lack of know-how that, yes, your ROAS is at 600%, but do you have an understanding of what the incremental ROAS was of the last 2,000 Euros you spent? That’s the basic idea of diminishing returns and the fact that this curve flattens. I’m not kidding, it’s sometimes not there, and I think it’s massively tied to the whole ROAS idea.

Mike: 100%. I think what disturbs me here is that, again, we talked about that air gap or margin of error between how well it models your profit.

[0:12:45] Mike: As ROAS follows a diminishing curve, my concern is that if that margin of error is fixed or constant, you’re getting into this area where there’s way less wiggle room. The margin of error gets bigger proportionately to the increment.

[0:13:18] Mike: Your actual profit and your ROAS will both be on diminishing return curves. The question is: are they on different curves? How is the shape of those curves next to each other? When a campaign is early scaling, there’s probably not too much of a problem, but it’s in that mature campaign where you’re trying to optimize and spend more that this becomes really scary.

[0:13:44] Chris: It can become scary. This understanding of incremental performance in general is something which I don’t see discussed enough in the market. I would certainly encourage everyone to have a close look at that. You can run with ROAS as a proxy for your profit because, to a certain degree, it works, but be aware of this flattening return curve because the margin can become scary if profit stays constant.

[0:14:18] Mike: Definitely. I think we have a visual aid for this. One thing is this idea of ROAS as a North Star—are you measuring it and allowing it to overtake what you’re actually trying to measure? But another topic is that, due to changes in campaign technology, ROAS is in an awkward spot because it can’t only be your business objective that you’re trying to reach.

[0:14:56] Mike: Do you want to finish this off for me, Chris? Because this is not even a knock on ROAS—I love ROAS as a KPI—but actually, no one asked ROAS to even be in this position.

Chris: No one talked to him! But ROAS—and by the way, this was a big part of my talk last week at the Ecommerce Expo—ROAS is probably the greatest of all time, but for me, it is currently also one of the most misunderstood KPIs because people don’t understand that ROAS is not only your North Star in terms of having a goal.

[0:15:23] Chris: Next to the campaign structure, it is the most important communication vessel to signal to Google what you want to do, due to the new technologies, especially PMax. I tell you, Mike—and maybe we can also show a slide here I talked about last week—when we do audits or account analysis across the board, from small to big retailers, we analyze accounts where the client is super proud of their very granular structure.

[0:16:22] Chris: They do it based on margin classes. Purely doing it based on margin classes is not the best idea either, but let’s say there are 10 campaigns in the account based on different margin classes.

[0:16:33] Chris: And then you look at the ROAS target setting.

[0:16:36] Chris: And they are all the same! I’m not kidding. This is—pardon my French—but this is almost stupid.

[0:16:46] Chris: Almost stupid. You treat your 10 campaigns, which you actually want to treat differently, with the one major signal you can give to Google, which is the ROAS, and you have a similar ROAS across all these tanks. Sorry, man. Then your only differentiator is your budget.

[0:17:06] Chris: So, we can try to pace it with budget. Again, I talked about this and I looked into the faces of the crowd—a great crowd, and hopefully some of them are listening now—and it felt like they understood it. But in conversations afterward, it’s a quite common thing: ROAS is primarily understood as a goal.

Mike: Let’s spell it out here. On the one hand, ROAS is a stand-in for a business objective, like your break-even or profitability. So, it’s a sufficiency target that should be rather stable. That’s kind of your North Star. But on the other hand…

[0:18:15] Mike: …it’s also one of your primary pacing tools. If you set a more conservative ROAS target or a more aggressive ROAS target, this is going to change how aggressively your campaign behaves. In the good old-fashioned days of manual bidding, we could influence bids directly. We can no longer do that. Our way of doing that is by changing the ROAS target. ROAS became basically a proxy for a bid. Now, it’s supposed to be both a proxy for your business model and a proxy for the bid, and this creates a discrepancy. People fiddle with the ROAS as a thermostat to pace their campaigns, but on the other hand, they want the campaign to hit a certain goal, and this is the challenge.

Chris: That’s why I said no one talked to poor ROAS! This thing is transforming into a very complex identity. People have to deal with the fact that if you treat ROAS as the proxy for your profit, that itself is an issue, but if you treat it as your North Star and you don’t see it as this communication vessel with Google, you are missing out on big leverage.

[0:19:35] Chris: One thing I would like to add here: because this next discussion is not directly tied to the ROAS KPI itself but how you work with ROAS. I still see, across the board from small to big retailers, that 7 out of 10 online retailers, if they even try to be dynamic with the ROAS setting, do it based on gut feeling.

Mike: Yeah.

[0:20:01] Chris: Which is beyond my thinking. Why don’t you have a scientific, data-based approach? Like what we do with our platform—no search here, but everything is at least based on a certain set of horizons. A lot of people just say, “Okay, my campaign is performing well, I’ll change it.” It’s really something which drives me crazy.

Mike: I hear you. Google is also working on their own target ROAS recommendations and budget recommendations. We’ve myth-busted them a bit in the past as well. But it’s definitely a challenge. In the end, what you want to be able to do is make sure that your campaigns, taken together, are achieving an account-level goal so that you can have that directional guidance of ROAS, which should be more stable. But then at that sub-account level, you need to be able to have flexibility with your budgets and your ROAS targets.

[0:21:26] Chris: Maybe we can show this slide. Basically, ROAS should be the North Star on the account level, but once you try to steer your campaigns—where you put some thought into how to separate them for specific reasons—let’s work with the ROAS. Again, it’s a proxy for the CPC; you want to treat those campaigns differently. So yeah, ROAS, man. I’m a big fan, but it’s become complex.

[0:21:33] Mike: Absolutely. Let’s move on to our next topic. I think we can keep it pretty concise because I’m working on some data for the next episode.

Chris: I know what you’re working on; I’m really curious about it. Anyway, let’s talk about the move. PMax is a black box.

Mike: The black hole! By the way, back in 2020 or 2019, I called Smart Shopping not a black box campaign type, but a black hole.

[0:22:22] Mike: But you’re right; back then this was absolutely the case. No technology was more opaque or intransparent than Smart Shopping. PMax has actually been improving on that for a long time now. It still comes up all the time. I’m working on this PPC survey together with a bunch of partners and people still complain that there’s no transparency in PMax and that it’s a black box. It’s a bit of first-impression bias because it was a black box in the beginning. I think we do, at a certain point, need to give credit to Google where credit is due.

[0:23:13] Chris: Carla, if you’re listening—our good friend Carla, our Google Partner Manager—really a great job. Credit where credit is due. Fun stuff: it was rather late and I was already at home, and I had a phone call with a client who was complaining about whatever. My son was listening, and my son is very much into astrophysics. He’s five years old but knows a lot about stars. I don’t know why, but we talked about a “black box” like a “black hole,” and he heard it and said, “Black hole? Why? Your job is way cooler than I thought it was!” He was asking why I was talking about TON 618, which is the biggest black hole. No, no, I was talking about the campaign type.

[0:24:15] Chris: But hold on, we’ll get back on topic. Do you know about Planet X? It’s a theorized additional planet they think might be hidden in the solar system. You have to check out TON 618; it’s a supermassive black hole that is off the charts. But yeah, PMax is not a black hole. PMax has become a great campaign.

Mike: Exactly. So, she took on astrophysics? We’ll do one episode on that; that’ll be this year’s Christmas special. I have enough knowledge to make a fool out of myself. I’m just saying my son was obsessed with Planet X at that age. All right, back on topic.

[0:24:55] Mike: This has been ongoing for quite some time. At this point, you know, some of the big criticisms of PMax were: “It is this cross-channel campaign type, but we have no visibility into those channels or ad networks.” “No search terms.” “At least let us know!” It’s fully handling the audience, the website placements… “At least let us see that.” And with basically all of these topics, Google has done a great job.

[0:25:42] Chris: Let’s preface here—the biggest tools without going too much into detail: you have the channel distribution performance report. That’s the newest of these, and it also has its flaws. Should we show a slide here for the channel distribution script? We have some scripts and tools that we’ve built on some of these new reports to help; it’s for free. It’s this channel insights report on steroids. Google makes the data available, but they’re not always good at turning that data into information. It’s our job to derive information and actions, but Google clearly signaled that it’s important to listen to the market.

Mike: Yeah, and the data is available. Then there’s the search term report, also now connectable directly by API. You can do a lot of crazy things with that.

[0:26:58] Mike: Search terms can generate lots and lots of rows of data. It’s a very heavy report, so the user interface is not a very good place to work with your search terms; it’s better to use a tool for that. But again, the data is there. This is the cool thing: I think we’re moving into a place again where—and we had an episode where we compared Smart Shopping versus PMax—Smart Shopping was a garbage product. You were literally forced to play a standardized game and there were hardly any chances to differentiate yourself against your competition by being smarter. PMax enables you now to really be the smart kid on the block again.

Chris: Yeah, I love it. I really love it.

Mike: I had to really swallow my pride because Smart Shopping was just such a poorly designed product, and it became very popular anyway, which was not encouraging. But with PMax, not just me but we as a company were like, “Alright, well, we see people are going to work with this technology anyway, so let’s help them do it.” Now PMax is at the point where it’s not a dirty word anymore. We openly recommend it as a future-proof campaign type. And again, if you do it correctly, you can be significantly better than your competition.

[0:28:02] Chris: That’s what I love. That’s why we exist as a company. The experts listening to us, you want to be smart, and Google enables you to do crazy intelligent shit again.

Mike: A placement report, exactly. This is something where, emotionally, the placements are horrible.

Chris: You’re right.

Mike: We’ll talk about this more in the future because it’s not all sunshine and roses, but at least we can see it. We can see the websites, apps, YouTube channels, and videos where our ads are being placed. That’s important because you should just know where your ads are showing up. We’ve also created a script here. We should show this slide. We opened the floodgates now!

[0:29:27] Mike: If you’re watching us on YouTube or Spotify—video should be coming to Apple soon too—very exciting. But if you’re only listening, it’s really worth it. You can Google “smec” or “Smarter Ecommerce script” to find them. They are really helpful. But back to the topic: overall, we can certainly constitute that Google has listened to the market.

[0:29:41] Mike: I think we can also say they have listened to us to a certain degree, because we are an important partner for Google. We appreciate the open conversations we have with them, and they did some significant changes.

[0:29:49] Chris: That’s it, and I love it. On a final note, Mike—because I think we discussed this as well—the more possibilities you get, the more complex this whole thing will be.

Mike: That’s true. I guess that is… yes.

[0:30:22] Mike: This is probably the only downside: if you want to untap these features or these new data points, the complexity increases. But I think for a lot of people, that’s actually welcome news. I think there’s a paradox with PMax where at a purely functional level it was simpler to use, but it was harder to succeed. Now, with the additional insights and complexity, it opens up some more differentiation space. If you put in that work, you can differentiate.

[0:30:43] Mike: The transparency is a bit ahead of the controls in some cases. For search term reporting, there’s negative keywording; for placements, there are exclusions. But the elephant in the room is the channel reporting—there’s actually nothing you could really do there. They offer some demographic exclusions, but they don’t give you demographic reporting. So, there’s still some work to be done. But we are in such a better place than we were three years ago.

[0:31:14] Chris: That’s appreciated. We could say some criticism, but I know some Google PMs listen to this podcast too. Listen, it’s not often we say it, but here’s to you: thank you for listening and thank you for making these changes.

Mike: Thank you very much. Let’s end this episode on this awesome high note. We usually somehow find our way down to a low point, but we’re ending on a rare high note and a tip of the hat to Google. Today is October 19th? Let’s mark this down in the calendar: an utterly positive episode, nothing to complain about. The stage is yours.

Chris: I think we’re at our time box. Thank you so much for giving us your time and attention. This has been another episode of Growing Ecommerce brought to you by Smarter Ecommerce, also known as smec. You can learn more, as always, at smarter-ecommerce.com. And again, if you like this podcast, please give us a shout-out on social media, or give us a review or rating on your podcast platform. We really appreciate it.

Mike: Thank you, Chris.

Chris: Thank you, sir. Alrighty.