Growing Ecommerce – The Retail Growth Podcast
Feed your growth mindset. Ecommerce is growing, and so are the challenges and opportunities for online retailers. In the Growing Ecommerce podcast, Mike Ryan and other smec experts are joined by industry leaders in ecommerce, digital marketing, and data science. By sharing business trends, practical solutions, and best practices, this podcast helps online retailers solve the challenges of tomorrow.
Growing Ecommerce – The Retail Growth Podcast
Instagram's Fake AI Influencers? Plus: A CSAM Scandal Rocks Google & Zuckerberg’s Moderation Cuts
Who’s real anymore?
In this episode of Growing Ecommerce, Mike and Chris break down the chaos unfolding across the digital advertising world:
• Instagram is testing AI-generated user profiles — Are we entering a world of fake influencers, and what does it mean for ad targeting?
• Zuckerberg pulls back on content moderation — We dive into the actual data behind Meta’s shift in strategy.
• Google's ads are showing up next to harmful content — A CSAM scandal rocks Google's Display Network. How many brands were impacted?
• TikTok Shop expands to Europe — Could this reshape the e-commerce landscape or is it just another platform grab?
Whether you're a marketer, brand strategist, or just trying to stay sane in the age of AI illusions and platform power plays, this episode is your no-BS guide to what’s really happening.
🎧 Subscribe for more deep dives into the future of digital marketing, AI, and platform ethics.
💬 Leave a comment: How would YOU feel about running ads next to fake AI profiles?
Grab the State of PMax 2025 report—free on the blog or as a downloadable PDF:
https://shorturl.at/ali5A
—
About smec (Smarter Ecommerce):
At smec - Smarter Ecommerce, we specialize in transforming business goals into optimized ad campaigns. With over 16 years of experience in Google & Microsoft Ads, our intelligent software and expert services help retailers achieve superior results.
We're committed to giving you the tools and insights needed to stay ahead in the ever-evolving world of digital advertising.
—
Make sure to follow smec - Smarter Ecommerce for more performance marketing insights.
smec - Smarter Ecommerce: https://www.smarter-ecommerce.com
LinkedIn: https://linkedin.com/company/smarter-ecommerce-gmbh
Newsletter: https://smarter-ecommerce.com/en/newsletter/
Instagram: https://www.instagram.com/smarterecommerce/
#Instagram #GoogleAds #TikTok #Zuckerberg #DigitalMarketing #AIinAdvertising #BrandSafety #EcommerceTrends #MarketingPodcast
Welcome to Growing E-Commerce. I'm your host, mike Ryan, and with me is Chris. I'm here again. Yes, that's right, our co-host Chris. We're very glad to have you back.
Speaker 2:Chris. Thanks, man, this time with a three-piece suit, because you put a lot of pressure on me right how I dress up and I hope you're happy with it.
Speaker 1:Yeah, I am. I mean, you're making me look like a hobo. I'm like. Last time I wore a t-shirt, I'm putting on a sweater, this time You're getting there, mate, you're getting there.
Speaker 1:For me, that's pulling out all the stops. No, I can clean up well, but I would rather not. So welcome to Growing Ecommerce presented to you by Smarter Ecommerce. We have a lot of exciting topics today and we have actually a huge backlog of content, and some of this stuff is even from late last year, but these are really, to me and to us, really interesting topics Still up to date, like AI-generated users on social media that's a theme that's not going anywhere, I think. Content moderation on meta we uncovered new data about that and TikTok shop coming to Europe, and a lot more. These are things that will be playing out all year long. Yes, so, without any further ado.
Speaker 2:Mike. Let's jump into the first one, the format we call it. What If AI-generated users? Maybe you give us some insights and background stories what's the whole thing about, and then we can start the fruitful discussion.
Speaker 1:Yeah Well, it's not such a what if? Because it really happened. Late last year on Meta's platforms and I don't know if these were discovered in both Instagram and Facebook. I can't remember, but great coverage here from Financial Times, so look it up. They reported that Meta had silently rolled out exactly what sounds like synthetic profiles. Synthetic users we don't know exactly how many there were or how long this went on. Synthetic users we don't know exactly how many there were or how long this went on. They were taken down shortly after the news broke due to widespread negative feedback.
Speaker 1:Yeah, what a surprise and multiple reasons why, but I think it raises an interesting question what if? Because I mean to be honest, let's think about Elon Musk, his war on bots that he promised when he took over Twitter. Listen, I'm a daily user of Twitter X and I can tell you the bots are still there and those are. You look at those bots and you think this should be so easy to detect and easy to block, and yet nothing happens.
Speaker 1:It's a widespread spread problem. So if platforms can't even block bots, what's to stop them? Or like, how are they going to manage ai users? And what if they start generating their own ai users, which?
Speaker 2:exactly. Yeah, there's this multiplier effect, if you will. What if? If? An ai account on facebook creates hundreds of new ai based accounts, which is which is quite crazy. And and the one thing you mentioned right here, and this is the big what if? Because it changes the whole dynamic of this game. I'm using X as well, and you understand, mostly when there is a bot involved right, because you understand the language, the context. But when we're talking about AI accounts right, ai generated accounts, and they are they are really managed by an ai.
Speaker 2:I think it will get so tough again for the end consumer to understand and differentiate what is real now and what. What, what is what is? What is the? I think? I mean, I read an article and the suck who looks more and more like a checked I don't know surfer boy. I love his transformation, but he was quite open about it. Yeah, they will be flagged as AI accounts and they will. Man, we heard all of that.
Speaker 2:The question is how obvious it will be. I think it's a really, really dangerous and yet potentially powerful thing. I think we briefly talked about that, mike. The big chance would be. I mean, thousands and millions of brands would love to have strong influencers, right? Imagine you have an AI-based generated account who can do precisely that for you, but smarter, with less marginal costs. I mean, there's chances and threats at the same time. My question to you would be now this trade-off, how does it feel with all the information we have right now? Is it rather a big opportunity or is it rather a threat to, yeah, what we have been falling in love with interacting generally with other people on Facebook? What's your take on this?
Speaker 1:I mean, I can sometimes be a bit of a Luddite or a Neo-Luddite, and this is an area where I don't feel too good. I mean there's already, so, like we just talked about influencers to bring this back to e-commerce, like there's already ugc user generated content. At its moment, you can talk if you gc is played out or not, but the idea was that this was really authentic, kind of organic content from real people, and then brands were picking it up and multiplying it with ads and stuff like that. And then brands started really encouraging UGC and it's kind of this interface with micro influencers. And now people are using AI technology to generate deep, faked or artificial UGC. So it's synthetic UGC, which is weird, and I mean that's one level, that's one layer.
Speaker 1:There's also we see in other platforms like Google, ai generated images and backgrounds and so on. But to have an unmanaged like in the end you still sign off, someone signs off on that content but to have AI running around interacting, I just don't know. I mean and there does, by the way to me it doesn't seem that appealing, but for other people it is. There's companies like Character AI where this is what they do. People want to talk with synthetic users.
Speaker 1:They want to, you know, have this relationship and it's going to be, I think, like through the looking glass here.
Speaker 2:Yeah, I mean, honestly, I'm right on your side here. I'm generally, I think, like through the looking glass here. Yeah, I mean, honestly, I'm right on your side here. I'm generally, I think, an opportunity-seeking person, but when this broke the news, especially that it had to break the news the way it broke the news, like why was it hidden? Yes or no, it had this, it was. It was, I don't know, somehow fishy, didn't?
Speaker 2:feel the first yeah in the first place, but in general, I think, if, if, if, if you really can't differentiate anymore and it's happening on a daily basis, we know that for sure but at least with the bot thing right there, there are ways to detect it, to, to, to fight against it. This, this account driven eis, it's, it's crazy and I think there's this opportunity, of course, but I think it's a massive, massive threat To the core. I think that the core value which a lot of people still get from Facebook, which is genuinely interact with people, right. If there is a big question mark to that core value, I don't know, maybe there's a down the line, also a business model risk for Facebook, because I think it's a very the line. Also business model risk for Facebook, because I think it's a very, very risky move because it undermines, potentially undermines, this very core value. Prop, facebook has to talk to each other and, yeah, I'm right on your side, the threat is big here.
Speaker 1:Yeah, and there's even there's these other like alternative platforms, like Be Real, which is it's about authenticity. People are already, I think, a little fed up with that, but right now, it's just that point where there's no rules of engagements and even we see on LinkedIn, like I get people using AI to comment on LinkedIn posts and stuff like that. We see you, we block you and I'm going to use that as a transition. It's awesome. It's awesome speaking of blocking and speaking of meta and speaking of things where your content, like what else, can be in your feed next to your ads or next to your brand. I want to talk about content moderation on meta a little bit yes, I'm looking forward to it.
Speaker 2:I I asked you a couple of times before the podcast what is this whole thing about? You wanted to show me some data.
Speaker 1:Yeah, I'm curious.
Speaker 2:Now I'm really curious, now Definitely. I mean I know what it's going to be about, but the data is somehow telling a quite interesting story. Exactly that's what I expect.
Speaker 1:Hold on, actually, chris, I want to share my screen with you so you can see this and we'll show to people. We're going to tell you what's going on here, describe to you what we're seeing, and we'll also add some visual aids for people who are watching the video podcast. Just for a little context here, mark zuckerberg made some headlines in january this year when he he said that basically, meta is excessively moderated and that there's a political nature to this as well. This is part of this movement that's happening in Silicon Valley right now realignment of priorities to align more closely with the current administration. This is a whole topic. We don't have to get into it too much. We could maybe someday, yeah, but he said that basically, the platform is heavily moderated and he's changing that now. But he said that basically, the platform is heavily moderated and he's changing that now.
Speaker 1:So what's really interesting is that in Europe, these large platforms are required to submit daily moderation data how many moderation events there were and what were the reasons for moderation and they also have to look. They can provide info where people demonetize, that they have their visibility limited. What was the nature of the moderation? It's really interesting data. So I decided to look at that and find out okay, was meta more heavily moderated, or what's going on? And I have a dashboard of the EU's database here and what we see is that if we look back to like late 2023 and early 2024, the volume of moderation happening on meta was pretty low and the two main reasons by far were kind of a general terms of service violation and also data protection and privacy violations. These were the two main reasons.
Speaker 2:So how to read the chart? So, let's say, on September the 3rd, no quarter three, september 30th, there were roughly 144,000 incidents where Facebook had to administrate, and they were basically in the area of scope of platform services, right, exactly.
Speaker 1:That's kind of a baseline. So let's say 150,000 terms of service takedowns per day, and then I'm just going to advance that time series and look at how this developed over time. The number was, the volumes were increasing a bit as we got through into 2024. This is where it got interesting. Around quarter two you see the numbers starting to ramp up a bit.
Speaker 2:So now, just for reference, and again the scope of platform services.
Speaker 1:Yeah, the scope of platform jumped from 150K to like 760K. Yeah, so quite a multiplier there, and we're going to keep going. Then it gets really interesting. The volumes are sometimes jumping up even more, but it's Jesus, yeah, 4 million, almost. Like. Here we're at 2.5 million in September of 2024. Remember a baseline of 150,000. And it really starts taking off around the time of the election. And so we're averaging now yeah, easily averaging like 2 million takedowns per day, and there was even a peak of up to 6.5 million takedowns due to terms of service violations. And what's also interesting is that those other reasons, all kind, of disappear.
Speaker 1:It's really just the terms of service violations that scale.
Speaker 2:Do we know what's behind this service thing? We don't, we don't.
Speaker 1:Okay. So this I think I would describe this as a maybe a political litmus test because, without knowing, I think that the way people interpret this largely could depend on their politics. But it can be that the volume of content that needed to be moderated was massively increasing. It could also be that the activity of the moderators was increasing, that they were moderating more. Both things might've increased, but then the numbers starting in January, in mid-January, just like Zuckerberg said he would do, the numbers fell off a cliff. So I think for e-commerce, you have to be aware that it is true just from the numbers. It is true what Mark Zuckerberg said Moderation had dramatically increased. We can get into a debate about the reasons why. And now that moderation has stopped and some people are calling this Zuckerberg's post-fact-check era 2023, it was, rather it had a certain baseline in 2023. It was a lot different in 2024.
Speaker 2:And now it's looking a lot more like 2023.
Speaker 1:Again, impact for advertisers. Is there one? I think brand safety is actually turning into a secret theme of this episode, because we're going to talk about brand safety again in a few minutes and I, you know this. This AI generated users raises brand safety questions. This with meta it's a push channel you can't really control. It's not like Google. You decide sort of what keywords are going to appear against and so on. There's a targeting in there. Meta is push and Twitter even X, these other platforms. These are push platforms where you don't have a lot of control what content you'll appear adjacent to or parallel to, and there is potentially a brand safety thing there, but we don't know more. We just know that there was a very visible movement here in terms of moderation on meta.
Speaker 2:May I put you on the spot here, gladly. No, look, I mean because I I watched, watched the interviews and the podcasts were. Were the big suck? Admitted that moderation was politically motivated? Were they forced? I don't know, but I think it's not even a secret anymore. Knowing that, what do you think was the driver? Because, because we're talking about a factor of, what is it? 50, I don't know 45 to the levels in 2022. And, yes, more reasons to it, like with everything else. But what, what do you think was the main reason here? You can be wrong, but I think you was an answer, or at least me All right.
Speaker 1:Yeah, I mean, I think that Mark Zuckerberg is a political opportunist and so I don't doubt that it was politically motivated and I don't doubt that his new stance is also politically motivated. I mean, if you flashback to 2016, in that era when, like during the first Trump election, one of the single most valuable people in the entire company, palmer Lucky, he was responsible for, yes, palmer, I always say Lucky, palmer, but Palmer Lucky, he was responsible for their whole VR efforts and he got canned because he politically supported Trump. And then it's just rich knowing that. And then to flash forward, and you've got Zuck with his tan and chains and things have changed a lot. And I don't, I don't think that zuckerberg is that ideological. I think he's opportunistic.
Speaker 2:Yeah, I will co-sign that awesome insights. Do you and do we know? So? These are the most current numbers, right or no? 24 and now it? They fell off a cliff again after that right, yes, exactly Exactly.
Speaker 1:They went back down again. Interesting.
Speaker 2:Awesome, mike. Then let's jump on to the next topic. Sure, thing You're in the spotlight now, chris, unfortunately, but I think it's a rather low-key topic, right? Yeah, let's see. Let's see In the news TikTok shop coming to the EU. Yes, I mean low-key thing.
Speaker 2:I personally think it's not a low-key thing because it's, again, I would say, another variable in an already very dense market in Germany. Yeah, to put that whole thing in context, it's not directly related to TikTok, but what we see on a broad set of numbers is that Germany is probably currently, from an e-commerce perspective, the weakest market in whole Europe. It's the problem child of Google. Just the other week I talked to a high exec person at Google. So there's already a lack of demand. The retailers are fighting for every revenue there is and now a major player enters the market. So I think it's not a low key thing from that perspective, but rather less controversial.
Speaker 2:What's the story behind it? So TikTok basically opens kind of a marketplace in Germany. Now, as you know, mike, they have been running international marketplaces in Spain, italy and the UK, whereas UK was failing in the beginning, which is kind of crazy because you could think what reason can there be that TikTok is failing, but it was not a big success in the UK. They are now trying it again and it seems like they are now successful and found the right code and there is one strong incident indication for that we talk about in a minute. But for Germany, they will, as I understood correctly, have a fulfillment by TikTok, so they will create warehouses, they hire stuff on the ground in Germany and will become, I think, a major, major player. And life will not get easier for I think for the marketplace in general, and my take is it certainly will mean heated up competition for Meta, maybe Amazon, otto, a huge marketplace in Germany. So they're already, I would say, dense market. We'll get more denser, which is kind of a good and bad thing I'm going to come at it.
Speaker 1:Let's just talk for one second about TikTok shops. So some of the options here. This allows you to, for example, do live shopping. It allows you to have basically the ability to check out from posts, like there can be a post that features a certain product and people can then buy that, and then there's also a kind of a proper store or marketplace in there, and I mean for a brand.
Speaker 1:To see it from another angle from you, chris, maybe this is a new distribution channel for them that they wouldn't have before. I just don't know. If I don't know, I don't know if this is can create new demand, can it locate new demand, or is this the same demand and it's being fragmented across another channel and you know what will the unit economics of that channel look like? I also think with TikTok shop, something that makes it special or different from what I've heard brands that make it on TikTok shop. There's an unpredictable kind of virality element there. If you really want to scale, some of that is out of your control element there. If you really want to scale, some of that is out of your control depends if you achieve some viral effects on the content and then that content has shopping attached to it. It's really positive.
Speaker 2:But point, point taking. So you, the argument is that it could be incremental. So so there's an incrementality in in tiktok as a new or somehow differently shaped distribution channel. Yeah, that's what, okay, could make sense. I mean one thing for sure as you know, lidl did some test flight with TikTok. It was I think a live sales promotion.
Speaker 1:Yeah, they had some. What was it? Protein bundles, protein bundles.
Speaker 2:I should order some.
Speaker 1:I'm not in shape, right now You'll bust out of your suit, Chris, the suit. It's a tailored suit.
Speaker 2:Yeah, exactly, I have to be careful.
Speaker 1:It's fit to how big you wear when you.
Speaker 2:I have to be careful. But what's crazy about the Lidl numbers, ladies and gents? I think they were sold out in 18 minutes. But again, the question is is this incremental to Lidl or was it somehow just a fragmented demand, which is there anyway but, point taken, could be an incrementality there Probably. We'll have to find out. Yeah.
Speaker 1:I mean, I think the beauty of something like that for Lidl is they probably introduce their protein line to some new people, yes, and then those people might return in store and buy it. Yes.
Speaker 2:And I think TikTok has a chance here to be maybe push and capturing at the same time, especially with this live sales promotion functionality. Because one thing is for sure there's always a level of virality with TikTok which other platforms don't have. And for Lidl, this protein line. I think you convinced me that there might be some positives about it. In general, I would assume that it's still if demand is not there, which is the case right now in Germany. The question is how much incrementality there is right In a marketplace where the sentiment is just not good. But this will change, that's for sure. Yeah.
Speaker 1:We should move on, but it's a very interesting. We can come back. I'm sure we'll come back to TikTok again. There's just this trade-off. Like everyone says, I'm too dependent on meta Google For sure but on the other hand, things become more fragmented, which means increased complexity.
Speaker 2:Life is a trade-off, but we can tackle this. Maybe next time.
Speaker 1:Let's stick a pin in it for now what's next on our agenda for us? Let's stick a pin in it for now. What's next on?
Speaker 2:our agenda for us. Yeah, this one, I mean this is certainly not low-key, because it's a highly, highly, highly sensitive topic and I would like to get your reaction here and maybe again provide some background information. Google, amazon, csam scandal. Enlighten us. What is CSAM about, what does it stand for and why is it such a big scandal?
Speaker 1:Yeah, csam was an acronym that, thankfully, I had never even heard of until some weeks back when this news first broke, but it stands for Child Sexual Abuse Material, and so what happened with Google and Amazon? It's pretty concerning. You know, there's, they've both, they're both building these display empires, and the idea with these display empires is that more is more. They want to monetize as many possible websites as possible so that they can offer the biggest reach possible and with that become. There are these huge scaling challenges where you know what kind of quality control is in there, what kind of brand safety is in there and also, just from an ethical standpoint, what sites are they monetizing.
Speaker 1:It's not the first time these networks have come under fire. Google search partners came under fire over a year ago for helping to monetize theoretically at least, helping to monetize sanctioned Iranian websites, for example. These controversies pop up and they're usually, in recent memory, driven by a company called Analytics, which is another interesting company to talk about one time. But yeah, so to bring it to the point, there was these completely unmoderated free photo sharing websites and these websites. There's valid users of these, I guess, but they were also used to host exploitative material. Okay, and you could have display ads showing directly adjacent to this material.
Speaker 2:so they were.
Speaker 1:These websites were part of the display google display net network exactly exactly these websites were in there, okay, and google and amazon have had a pretty weak kind of lame defense of that, which is that this didn't happen a lot but now you get the reaction this, this was, this was the argument here. Yeah, okay and to me it's something that should be so easy to agree on, zero tolerance, and that one ad impression helping monetize this content is one too many.
Speaker 2:It's like a broken window theory, right? I mean, even if it's just one impression, and how can you control it that there won't be another one, a third one, a fourth one? Yeah, this is crazy, do you have some numbers here.
Speaker 1:I do. Yeah, I had a special look at this because one thing you know, we have access to hundreds of advertising accounts and so we can just I can check this stuff in minutes. And so I did that and what I found? It's true that these impressions, in the grand scheme of things, were vanishingly rare, okay, but what's disturbing to me is how widespread they are. That might sound a bit contradictory, but what I mean is that any given account might only have a couple of impressions served on this property.
Speaker 1:But a lot of accounts had. Okay, a lot of them did Just to show you impressions on these domains are one of the problem. Domains I analyzed were detected in 23% of a sample of 650 accounts, so significant one in four. And what's worse than that, a lot of them also had the domain in their placement data. They didn't have impressions. So that might sound a little weird, but basically it means that the ad loaded and was ready to fold. Yeah, below the fold. So as soon as you would scroll down on the page would have been an impression, you'd see the fold. So as soon as you would scroll down on the page you'd see the impression. And when you consider that data too, then 40% of accounts had this data in there. I mean, I don't blame the advertisers here at all, because you're showing against so many domains and advertisers try to use brand safety tools. They're coming under fire as well, verification tools coming under fire here, but I actually think the responsibility belongs at the platform when they onboard these no doubt these partners.
Speaker 2:No, they're accepting them as a business partner, exactly yeah it's not like that you can't join a google display network like I mean. There has to be a certain registration process in place.
Speaker 1:Yeah, I mean, probably part of the problem here is that then google signs agreements with other like aggregators who resell like it's it's.
Speaker 2:It's just not moderated, it's not regulated at all and what's, what's the best, best way to to to get this, this, this fixed? Because I the question. I read the statements partially. I read the statements. I think they were completely bollocks, honestly speaking, because even if I just get three impressions on a website I know sharing CSAM content, I think it's a massively bad thing For Google. What is a realistic way to get this? And I'm not just Google, by the way, it was Amazon as well, right?
Speaker 1:Yeah, Amazon as well, yeah.
Speaker 2:What is the best way to get this under control? Such sensitive topics, yeah, yeah. What's your take on it?
Speaker 1:I mean, I think that they need to work closer with authorities who monitor this type of thing, because this domain was on a problem list for years. So Google did take it down this domain in July last year, for example, but that domain had been on a watch list for years, Years before yeah.
Speaker 1:So why did it take so long? And, by the way, google is, they're quite active If you look at their organic search. They have a database there, a transparent database of data, like the number, the volume of websites that they have restricted visibility, taken out of, like the listing, like de-indexed to prevent them. Because of exactly this reason, and my question is why, like it's proof that they can do it, why don't they do that for their paid network as well?
Speaker 2:I mean, it's, it's, it's crazy to bring it up, but I think it's it's. It's crazy to bring it up, but I think it's an interesting angle to this whole story. From a performance perspective, I mean, I would be shocked that these sites were performing right, because there are traffic on these websites. Yeah, I mean, what about that? Wouldn't that be a way for Google to detect these websites which are just not performing at all, on top to the highly critical content?
Speaker 1:Yeah, I mean, I think it's a good point. This is a general problem with the Google Display Network. It has just become too big. You know you hear about too big to fail. Google Display Network is too big to succeed. It's just so massive. And I'm certain that there is a strong correlation between brand safety and performance, and it sucks to talk about performance in this context. But if we view the issue more broadly than that and look at the whole network, if you would solve for one problem, you would solve for the other. I'm convinced about that. And Google has a kind of a financial responsibility toward advertisers on the performance side and they also have an ethical responsibility which is heavily overlapping with the brand safety topic.
Speaker 2:Both responsibilities were probably let down in this case Absolutely.
Speaker 1:These placements were there for the sake of them being there. They were there honestly for the chance to make Google a buck, and then the sad part is that these terrible websites get monetized in the process.
Speaker 2:Crazy thing. Crazy and disturbing to a certain degree. I mean, we will keep an eye on that. For sure. The sites are down now.
Speaker 1:That's what we know right, I don't know the sites. I mean, I think the sites are up but they're not being monetized, they're not being part of the app, exactly, yeah.
Speaker 2:Okay, so this, at least, is a good news here. The next topic and the last one for today's episode is for sure a more positive one. At least I think so. We have the section by the numbers where a state of PMAX report which unpacks, I think, some super interesting insights. Mike, do you just want to kick it off?
Speaker 1:and I'll jump in. Yeah, definitely it's. Oh man, I'm feeling a lot of whiplash here. It's a bit like you know, they talk about on the news, they talk about some tragedy and then they're like and now back to sports.
Speaker 1:Now back to sports, but yes, let's talk a little bit here about some other research that I've been working on.
Speaker 1:As I mentioned, we have access to a lot of account data, and so I just want to here I'm going to share my screen with you again, chris, so you can see this stuff I'm looking at performance max data from thousands of PMACs, campaigns across hundreds of accounts and trying to derive new insights at kind of a market level, an aggregated level, and also figure out what we can learn from those things.
Speaker 1:And just to show you a couple of things that are going to be appearing in this report, I have a time series of the count of PMax campaigns per account over time, how this has developed, because people joke a lot about PMax that Google wants you to just turn on a single PMax campaign and let it run, and we could get into a whole episode about why that's a bad idea, but what we see is that this number is developing strongly over time. So, indeed, like if we go back to September 2022, when PMAX really started rolling out at a very large scale, the average was like 2.5 campaigns per account, and now we're up much closer to five per account. So we've seen that steadily trending upward over time. Fact of the matter is that the mode, the most common value, is still one.
Speaker 2:Okay, but it's something that I'm certainly watching closely may I just jump in here because I think it's important to discuss just briefly. So we think it's rather a question now, yeah, do we think that this development is a good sign by default, so the more granular the better, or are there some some things to to have a look at, because it's not just a good thing to have five campaigns instead of what? What is your take on it?
Speaker 1:yeah, it's a good question. I mean, in general, we're seeing that campaigns or, excuse me, accounts are a lot more consolidated than they were a few years back, and that this is generally a good thing. It's supportive of the technology, and so I don't want that chart to just keep going up forever. Actually, I think there's probably a comfortable ceiling in there somewhere, but it does show me that, in general, advertisers are getting a little bit more sophisticated with their builds. I think it's very healthy to see that number developing. I do think there's a logical extreme, though, because you need to find the balance of granularity that offers you control and strategic steering, and then consolidation, which is the optimal environment for the algorithm.
Speaker 2:And I think one number I would like to throw out here to the audience is that the limits of granularity are somehow tied to the number of conversions per campaign, exactly Because the more granular you are, the less conversions you have, the less data that Google algorithm has to really perform. And I think that number we think is the absolute minimum per campaign should be around 30 conversions roughly. Everything below I think we both agree on is over granular or too granular, so to speak. Yeah, and a more consolidated approach would be more useful here.
Speaker 1:Yeah, Even ideally, you'll have like 60 or more conversions, and I'll talk about that in the report too. I also want to show you quickly something that's developing with bid strategies. This one's really interesting to me. You know, I first ran this report back in 2023 and there was like an 80% adoption of TROAS in PMAX, and now we're seeing that that number has been diminishing a bit over time and I think it's I don't know why exactly, but it's an interesting trend. It's down more like around 70% or so now. Potential reasons Well, one theory that I have I noticed that the number started dropping around peak season in 2023. Okay, so I think in 2022, when PMACs rolled out, people were just trying to handle the new technology, handle the migration. That's also why they probably had fewer campaigns back then. They were trying to keep things simple. I think in Q4 of 2023, maybe some people started trying to lift their rose targets for aggression, for the seasonality.
Speaker 1:Makes sense and they decided not to turn it back on. It's just a hypothesis, I don't know. Okay, interesting. But in the spirit of that, to show you one last thing before we wrap it up for today, I also had a look at how performance is with or without rose target. Like what can you expect from that? And so it's really interesting. Like right now I've got a chart on screen and I'll describe it to people who are listening.
Speaker 1:You can also check out the report when it's published. We can see what happens with conversion or, excuse me, with ROAS, with or without a target. And basically, when there's a target in place, it's quite stable. It fluctuates a bit. When there's not a target in place, it's fluctuating a lot per day and there isn't necessarily a huge penalty on the final, on the rows over the course of the whole month, but the penalty is more in terms of like the volatility and the stability of that, in terms of like the volatility and the stability of that. And I mean Google has a kind of a counter argument or counter narrative to that. They call that demand-led budgeting, because they say that if you don't put a ROAS target in place, then you're going to follow demand more naturally.
Speaker 2:Yes, Look. I mean, I love Google, but there is probably another argument. Why do you want to do it?
Speaker 1:More spent means more revenue for them. Exactly.
Speaker 2:I still believe that the return on investment, the target for us to set on campaign level, is probably next to the campaign structure itself, the one big steering wheel you still have under control, and you shouldn't this is my clear message to everyone you shouldn't leave that to Google blindly. I guess it's as great as Google is, but you have to run your own business and you have to steer your own business. Yeah, I agree. I'm very, very confident in stating this.
Speaker 1:Totally. It's something that you have to signal to Google what's important to you or not. There are not so many things left, so that's a big one. Yeah, that's for sure. What's important to you or not? There are not so many things left, so that's a big one. Yeah, that's for sure. But I think that brings us through the agenda today, chris. Time flies, mate.
Speaker 2:Time flies. Anything you would like to add here or no? Thanks for coming back. You can count on me Next time again with a three piece.
Speaker 1:Mike, next time again with a black, plain black t-shirt. I'm going to go for it, I think. Looking forward, sir. It was a pleasure again. Thanks, chris, thank you everyone and, by the way, this podcast is produced by Smarter eCommerce. You can learn more about Smarter eCommerce at smarter-ecommercecom.
Speaker 2:Where's your smile here? It is Right here. Thanks, mate.