Growing Ecommerce – The Retail Growth Podcast

Google's A/I Max & The Temu Takeover: Navigating the E-commerce Future

• Smarter Ecommerce • Season 4 • Episode 10

Share episode

Welcome to Episode 10 of Growing ecommerce! Get ready to cut through the noise with your hosts Mike and Chris, as they tackle the real-world implications of three massive forces in the market.

First, they take on Google's newest AI-powered ad product, A/I Max. While Google promises performance, Mike and Chris reveal how it can lead to chaotic bidding on competitor terms and the low-quality Search Partner Network. The hosts also discuss why an "old school" strategy called search query sculpting is more relevant than ever to regain control and maximize your ad spend.

Then, the focus shifts to the critical, yet often overlooked, topic of conversion lag and latency. The hosts explain how this natural consumer behavior creates discrepancies between your ad platform data and your backend sales reports, and why it's a key metric for both marketing managers and business leaders.

Finally, they reveal the latest data on Temu’s staggering growth across Europe. With fresh numbers in hand, the hosts reveal which countries are feeling the most pressure from this new market giant and what this means for local retailers struggling to compete.

This episode is packed with the insights you need to navigate the ever-changing world of ecommerce. Tune in and get ready to level up your ad strategy.

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:

Speaker 1:

Welcome to another episode of Growing E-Commerce. We are your hosts. I'm Mike Ryan.

Speaker 2:

My name is Chris, let's go. Let's go, chris. By the way, I told you that your voice is really, really special. Why thank you, I like it, I really like it, thanks.

Speaker 1:

Chris.

Speaker 2:

Can you say it again? Tell me something with the most beautiful voice there is.

Speaker 1:

Put me on the spot here. I really like it. Fee-fi-fo-fum, let fun, let's go. All right, that's what the giant says in jack and the beanstalk, so all right I, I I, by the way, great movie great movie? Yeah, there's a movie. Yes, yes, I never saw the movie.

Speaker 2:

Yeah, yeah there's a great, great movie I'm gonna have to check it out.

Speaker 1:

Yeah, for sure, definitely talking about. Let's talk about one giant in particular. Yes man what a segue.

Speaker 2:

Yeah, that was beautiful, go ahead.

Speaker 1:

We're going to talk about Google a bit here.

Speaker 2:

Our favorite giant.

Speaker 1:

How unusual for this podcast. We're talking about Google, by the way yeah.

Speaker 2:

No, never mind.

Speaker 1:

I'm not going to derail us already. I'll mention you afterwards, Chris no-transcript search term match types that exist in your reporting and I've been having a look at that lately to try to see what I can make of it.

Speaker 2:

What did you find?

Speaker 1:

Well, thank you for asking, Chris. I thought you would never ask. What did you find, mark? Well, first let me talk about PMAX. Yes, that's basically you know PMAX is covering. It's going to be it's a keywordless technology. You do not target keywords and it's going to automatically match you to shopping ads and text ads on Google search network as well as search partner network on Google search network as well as search partner network.

Speaker 1:

Now there's no, there's no network segmentation, so we can't know how much of that is on Google search network compared to search partner network. Um, that's a bit frustrating, but what we, what I typically see so far, is that it's not the surprising. You know, we see like a median of 90% of costs in PMax campaigns is feed-based and an awful lot of that, of course, are shopping ads. So when you look at your search term report, there's a ton of shopping data in there implicitly and it looks an awful lot like your search term report for standard shopping campaign. And, by the way, standard shopping it is, and always has been, a keywordless technology as well.

Speaker 1:

So this is something that Google has been working on for years and years is this keywordless matching of products and search terms, and I mean basically it looks all fine. But there's one thing that I noticed, which is that you know there's there's all kinds of there are very generic search terms that get matched against. There's also more specific ones and like, as an example, I was looking at a large skincare and beauty brands in the U? S and they show up against stuff like skincare and they show up skincare. One word, skincare has two words. This accounts for a lot of their impressions and it doesn't account for a lot of their conversions.

Speaker 1:

No surprise, but it does. It creates clicks, but it's just. It's the intent is seemingly not there. We have to like we're supposed to trust Google. Yes, they are very familiar with the user's intent.

Speaker 2:

Yes.

Speaker 1:

They kind of handle this, Like for me. I tried searching skincare one word, two words, and I don't get served any ads at all and I don't know. You know, it is such a generic. They're also even showing up for the word skin, Skin, yeah, which is. And I type the word skin and I get a Wikipedia article yeah, but someone else is getting served shopping at it? Yeah, for sure, and it's really interesting.

Speaker 2:

Interesting and, honestly, not that surprising. I would like to talk about some old approach we, as a company, successfully implemented back in the standard shopping news.

Speaker 2:

But before we jump on that, for me it's literally not a big surprise that the search terms are similar to standard shopping because, as we stated out in the state of PMAX report which was globally recognized as hell of a hit piece, by the way well done, mate, that. Yes, there is this 90% shopping share in PMAX, so I think this is quite normal and okay-ish, that the search terms are very similar to standard shopping campaigns. What's also not very surprising is that, of course, google is always always giving this very generic broad search terms a shot. Why? Because it's a hell of a cost driver, and you know what Google is very good at. They are looking at your campaign, you put your ROAS target on it, and they reach the ROAS on average. So Google doesn't give a fuck that skin or skincare drives some clicks, as long as the average ROAS is met.

Speaker 2:

The big question, though, is could you force Google to be more efficient and more selective with the search term? And this is the segue I want to create now. We had a technology which was called search query sculpting back in the days Before we talk about that, and if this might also be some good approach to PMAX. Now, mike, elaborate our listeners to what search query sculpting is in principle, how we implemented it, and maybe then we can discuss is it maybe also a good approach to PMAX?

Speaker 2:

Yeah, sure, I mean, I remember that you were driving this back in the days as product manager. I can remember. Yeah, that's A hell of a feature, by the way.

Speaker 1:

Thanks. I mean it was built on ideas from our friend Martin Hüttgeding Shout out to martin, he's a german agency called blue fusion and yeah, I mean years back he had this genius idea that you could basically combine negative keyword lists and campaign priorities and shopping to funnel the traffic. Because I said there's no, there's no positive keyword targeting in google shopping, but there is negative keyword you can exclude and what you could do in the end is create campaigns where, through carefully building these negative keyword lists, you could create a campaign where these more generic search queries would land.

Speaker 1:

And then more kind of mid-funnel ones and then extremely specific ones, or you could funnel brand and non-brand, or different kinds of builds were possible. And yeah, we had developed technology at the time that automated that. It became obsolete, unfortunately, or for better or worse, when smart shopping came around and then later on PMAX, because there was no search term data in the first place and there was no negative keywords.

Speaker 2:

That's the thing, but now it is available again. That's why PMAX is, from my perspective, the way superior product to Smart Shopping and the question I have. Thanks, Mike, for elaborating on it. For me, the question now is because the idea was great and I think is still great to tell Google that, yes, there are these broad search terms where I, of course, want to be there, but only to a certain degree and especially when you have limited budgets as an online retailer, of course you want to focus this budget on rather commercially intended product specific search term, because what a surprise you have way higher conversion probability with. So this concept was a fucking great idea. We had some tech behind it. Another question what's your take on it regarding PMAX? Because if the search term report is available, you can basically, I think, steer against it through the Google API. Can you do something here? What's your take on it in principle?

Speaker 1:

Yeah, in principle, the idea remains good in my opinion. I mean, google has basically asked us to not worry about that and say that they've got that covered and that's why they didn't provide this data for a long time. It's why they didn't provide negative keywording. They said we've got this. Our audience signals are so strong, we'll handle this for you.

Speaker 2:

That's why you have to show up when someone is looking for skin.

Speaker 1:

And that's the thing. And here's the problem, because it's very all or nothing. You can still add that as a negative keyword in PMAX. Now that's possible again and you can say that's generic crap. I don't want to appear for that, but there's plenty of like you might actually be of two minds about this. Or there might be somewhere it's not so black and white of BF2 minus on this, or there might be somewhere it's not so black and white, it's like well, yeah, I would like. I don't want to not appear for that traffic, but I want to have a different budget for that or I want to bid it differently.

Speaker 1:

I want to have a different return on ad spend expectation. For that there's still the strategic element, yeah of course, and that's what something like the old school query sculpting approach. Even if you would, even if you say Google's going to on a raw performance level, they're going to outperform an approach like that there's still strategic and cost-related reasons why you might want to implement a structure like that 1,000%, 1,000%.

Speaker 2:

Sorry, go ahead, Mike.

Speaker 1:

Well, there's one thing we're missing. I mean, do you want to say what do you want to?

Speaker 2:

ask? No, I just want to say what do you want to ask? No, I just want to say that to to our listeners who are operating with performance max on where performance max is a major channel. The cool thing is we're exploring this query sculpting now for pmax and, yeah, maybe one of the next episodes we can shed some light on it if this approach works. Like I said, in principle I just love the idea to tell google which search terms I want to really be aggressively shown and where I'm rather opportunistic, especially when you have limited budget. With open budget, I think you can somewhat trust google. Still there might be some steering to be made, but especially when you are somewhat tight on budget, this career sculpting idea I'm very, very fond of yeah and we will have some data to talk about the next couple of episodes, I think.

Speaker 1:

Yeah, for sure, I mean, I want to. I just want to mention the single largest limitation here. Like we've got the data, we have the negative keyword control, but the other core feature, or rather requirement or Prioritization, it's the prioritization. And there's no campaign prioritization yet in PMax. Google it, please, but.

Speaker 2:

Think about it.

Speaker 1:

Yeah, but then you, it gets complicated.

Speaker 2:

Yeah, it gets complicated, but we are here for the complicated stuff, right, let's have a look at it. But I think it's a great report and let's talk about it in the next couple of episodes. Sure, what can you do with this report? Sure.

Speaker 1:

And the next thing, before we move on from search term reporting, I want to talk about AI Max, that other new search term match type AI is everywhere. Yeah, it sure is. I mean, by the way, just a very short tangent, that technology, because I make it my business to know the history of everything that's going on behind the scenes here and, from what I can tell, that technology has been in various pilots and tests and beta for a couple of years.

Speaker 1:

A couple of years A while, and it was initially called Smart Match and then they called it. They called it originally, yeah, like Search Max, and then, most recently, they called it originally, yeah, like Search Max, and then, most recently, they called it AI Max. And now they're spinning up this story that AI Max is a technology that will help you appear on AI mode and AI overviews when these things get monetized through ads, and there's probably some truth to that statement, but that's not the origin of this technology, that's not the main purpose and it doesn't only apply to those inventory types. It applies to all of Google search, and so I think you know it's a example where you really want to have your eyes open and see through the product marketing or the branding that Google slaps on this. It's very convenient for them and opportunistic, and, to a certain extent, true that this is supportive of these new AI inventory types, but calling it AI Macs, that's not what it's about whatsoever.

Speaker 2:

But it sounds fancy.

Speaker 1:

It sounds fancy as hell.

Speaker 2:

And, by the way, I think the strategy Google applies here. It's a very similar playbook, like with every new channel. Here, the big question for me is and I would really like to get your take on it I'm as a retailer. I would always strive for incrementality, right, everything I implement, what incrementality does it provide? And this is the big question with everything Google implements, what is just like a transformation from one channel to another? Where's the incrementality? And that's the question. Ai max is there. Is there this concept of okay, this drives incremental revenue for me, or is it just like link it to turn for a second? You know what I mean. Like what is it?

Speaker 1:

yeah, from the left pocket left to the right pocket.

Speaker 2:

Do you have a take on that?

Speaker 1:

Yeah, I do and I've looked at. I mean, the thing is that I don't have huge volumes of data yet. I like to, I often really like to look at things in aggregate, but right now I can look at things anecdotally. There's value in that as well. Then the danger the anecdotes might not be representative, but I can also say they're not cherry-picked. I've looked at just a handful of accounts with this technology activated and there were problems in all of them. So it feels somehow, you know, maybe they all had bad luck. But just to explain what happens here, a couple things I've seen. I mean for context, if you haven't heard about AI Max.

Speaker 1:

Google was next. They were looking at migrating search campaigns to PMAX, but they abandoned this strategy of bringing everything to PMAX and the other logical way of solving this as well. What if we'd bring some of the PMAX technology to a search and achieve basically the same result? But it'll be more favorably received and it has been so. Basically, there there's now, you know, just your standard match types, your keyword targeting, and then you can optionally add on some keywordless layers to that and also some, yeah, automated, like the stuff similar to dynamic search ads.

Speaker 1:

This is another part of it, but we're talking about the keywordless technology right now and what I look like, what I see consistently when it's been activated, is these accounts on an impressions basis and sometimes also on a conversion basis. They do take off. You can really see a pretty strong scaling effect as soon as AI Max is turned on, and Google's numbers here I think their figure they like to throw around is a 14% conversion uplift, which is it's a hell of a number. It's a hell of a promise to make or an anchor to set in people's minds. Well, never sure, but I would say maybe, if you are, if you do not have mature, built-out search campaigns, this could be more beneficial.

Speaker 1:

If you have a pretty built out mature search account structure right now, the odds are to me not that likely that this is going to deliver good incremental traffic and what it will find. It might find something, but it might not be what you want. That's what I saw in a couple of cases here. Looking at the search term report, I saw one case. There's a large multi-brand retailer in Germany and they, their impressions went way up. Their conversions also went up, and then when you look at where that's coming, from they suddenly started bidding on competitor terms like crazy.

Speaker 1:

All kinds of competitors, but one of them in particular. The algorithm went insane and just started really bidding hardcore against the competitor, really bidding hardcore against the competitor. And this is an example. Like that's not really. I mean, that could be good or bad, but this advertiser it's not. Like they never had the idea before. Hey, I can bid against. Yeah, yeah sure. They chose not to so they're good reasons yeah, they were not. They were not positively targeting this some other match types served.

Speaker 1:

they didn't negative the traffic either. They didn't exclude the traffic, so that's how this was still possible, but they weren't targeting it. The other match types were more or less leaving it alone. But then they turn on this and the thing goes insane, and I'm serious. If I remember right, it was 69% of impressions in the whole campaign were now on this one competitor and if you look at all the competitor terms in there, it was basically about 85%, 90% of all the impressions. The thing went insane. That's not acceptable.

Speaker 2:

One last question Was it justified so in terms of incremental revenue and return net spend?

Speaker 1:

Yeah, so these, like I said, it was bringing in conversions. And the question is is it smart to do that?

Speaker 2:

Is that on strategy for you it?

Speaker 1:

conversions. And the question is is it smart to do that? Is that on strategy for you? Okay, it was. And another case I saw just quick and we'll wrap this up this advertiser. They also had a big spike in impressions not in conversions, not really in costs either. And that's because, for whatever reason, the algorithm, like, decided that it wanted to go crazy on search partner network and it went insane on search partner network. We're talking just huge volumes, and the problem with search partner network is that it's too big. Yeah, there are some. There are some great. Youtube is a search partner, ebay is a search partner. There are lots of great search partners and then there's a lot of everything else. We talked about this in the past. Then there's everything else.

Speaker 1:

We talked about this in the past. Then there's everything else and, in general, what I find is that there's very little intent in the search partner network. So in a way it's harmless, but it's also just not necessary. And it's not what you expect from a campaign that says it's going to like that, like that campaign was already in pretty good shape and it's just desperately looking and find wow there's all this extra demand and search, let's go for it, let's go for it.

Speaker 2:

Interesting but, like I said, pretty classic playbook of Google. Yeah. I think that the core idea is a good one, but, like with every other campaign, especially now PMAX, aimax you have to really really steer and manage the algorithm. You have to really really steer and manage the algorithm. It's more important than ever, that's for sure.

Speaker 1:

If you have an immature, let's say or less mature search build, then this could probably find some good stuff for you. And then you might get good ideas of what you want to target. If you have a mature one, you might find that it's looking in odd places and you know. Then you need to put guardrails in place. Like then you're going to have to start excluding those, those competitor kids. You might have to opt out of search partner network, but then the question is what will be left?

Speaker 2:

Will it still find anything? What's then the increment?

Speaker 1:

Yeah, and so this, it's a. It's a very new technology. They've been testing it for years, but it's also I, so I have to say I don't know in aggregate what's going on yet. I just look at anecdotes and it's very new. So maybe in a year from now I'll have a revised opinion.

Speaker 2:

Yes, that's what I wanted to say. We keep an eye on that for sure. We talk about that topic also at the D-Maxco right. Yeah, that's right, and yeah, the more data flows in, the more educated or the more information we can share with you guys. But quite interesting, one thing is for sure aimx is here to stay, that's what we can say. So for for better or worse. That's the thing what's next, mate?

Speaker 1:

let's talk about conversion lag and conversion latency.

Speaker 2:

Interesting topic sounds nerdy right, sounds nerdy but has a very, very real impact on metrics which are relevant not just for the operational online marketing manager, but also for decision makers. Yeah, I agree. Tell me what it is about.

Speaker 1:

Yeah, I mean, I think that this is something that operational teams will be more or less familiar with, and then it's a question of, I think, at that decision maker level as well, it's worth understanding these. I mean conversion lag, also known as delayed conversions. I think there's a decent level of awareness about this. But let's recap it this is just naturally occurring consumer behavior because Google measures by default. It measures things from the interaction date, usually the click date, and then someone interacts with your ad and some time might pass till they actually complete a purchase. And that's this attribution thing. And that's all fine.

Speaker 1:

It can take, depending on how long your cookie window's set or whatever can take a while. And then what you particularly see, this will typically correlate with average order value, so that more expensive items often have a longer consideration phase. It's kind of self-explanatory. Also, it can have a seasonal characteristic. For example, what I've seen time and again is that in the days, slash, weeks leading up to Black Friday, we see that that lag increases because people are clicking on ads but they're waiting for sales to begin, they're comparing options and then in the peak holiday season people are under a lot more time pressure, so they're clicking an ad and buying same day and stuff like that.

Speaker 1:

So these are, you see, these very consumer and demand driven effects on this. Where it's important is that there'll be a discrepancy between this and your backend data. Yes, because your backend data is just going to report purchases.

Speaker 2:

The purchase date yeah, when they happen.

Speaker 1:

And Google is reporting it. It's attributing it all back to the date of the clicks, so it's a very googly view on the top and there is a solution for that. What is it? It's a metric or a set of metrics. They're conversions by conversion time, and this basically means that that data is going to be reported by the purchase date instead of the click or interaction date, and so that's automatically going to line up with your backend data better. It's also going to help you make if you want to look at last week and or even today or yesterday if you want to compare that to last year or whatever period. That's hard to do because normally that data is not finished attributing yet and you're comparing it against data that is fully attributed, so it's not a fair comparison. So using this helps you get better like-for-like comparisons on recent time periods, so there's advantages of using it.

Speaker 2:

Yeah, for sure. By the way, this leads me to another. So very interesting. This leads me to another topic we can't discuss today because we are pretty much into the time box. Maybe next time, because you were talking about the backend right In general. So, yes, you have basically a revenue reported based on the purchase date, which is fine. You have the discrepancy based on this latency effect with Google, but, in general, what I find more and more happening to big retailers, to small retailers, is that, in general, they're looking at backend data, which mostly is not representing the Google data. Yeah, and the crazy thing is that a lot of retailers take decisions on the backend data and try to optimize the Google campaigns based on the backend data, which I don't know. I mean, there might be correlation, there might not be. Maybe we should discuss this topic next time. A, what a big thing this is and why these data have to be harmonized and which ways maybe they are available to harmonize their data. It's crazy Feeds into that partially, but it could be a topic on its own.

Speaker 2:

I will take a note here. We can talk about this next time.

Speaker 1:

Yeah, I like that. Let's do it All right. Do we have time for Timo or are we out of time?

Speaker 2:

You are the Mr China, right? You are called the MrT because every big media outlet wants to talk with you, Timo. Timo itself wanted to talk with you, right?

Speaker 1:

Yeah, no, yes, did it. Did we never have a follow-up about that, chris? No, no, why we should?

Speaker 2:

Also put it in the backlog. Let's put it in the backlog, chris. How was the meeting with Timo? Were they friendly? They were cordial Cordial. Okay, let's go. Yes, but for we should. We should Three minutes.

Speaker 1:

Real quick.

Speaker 2:

Real quick.

Speaker 1:

I just want to mention that Tmoo is required to share lots of data with the EU because, yeah, because of regulation, and one of those things they have to share they're monthly active users. So there's fresh monthly active user data from Tmoo for the last six month period and we can compare that to the last time they reported this. Let me guess they are growing. They compare that to the last time they reported this. Let me guess they are growing. They're growing, chris, they're growing. What would you consider a good growth rate?

Speaker 2:

You know like for normal retailer or for Tim.

Speaker 1:

Let's hear both.

Speaker 2:

No, I mean the reality is. And when we jump into the data now we see some rather interesting discrepancies between certain markets. Yeah, quite interesting. Rather interesting discrepancies between certain markets yeah, quite interesting. But in general, I mean the reality is retail in today's economically environment, I think, double-digit growth. If you hit that you should celebrate the hell out of it. Yeah, that's great Well for TEMO, I think different standards apply. Sure.

Speaker 1:

Let me get into it. Let's start with the worst market, the worst market, the worst market, their worst market. Actually, their worst market is Denmark. Okay, at about like minus 6%, in fact.

Speaker 2:

Minus 6% yeah.

Speaker 1:

Netherlands also right. Netherlands was the only other market that had negative growth, and that was my Shout out to bullcom right, they have this market in.

Speaker 2:

I love both.

Speaker 1:

You have to assume, you have to assume I also know some of these Eastern European markets with like EMAC, EMAC yes. I think when you have a market that has a dominant homegrown marketplace, that's going to be a tough market for Teemu Shout out to EMAC Great clients.

Speaker 2:

I love what they do. A hell of a player, awesome. Okay, so there are some markets that shrink.

Speaker 1:

Yeah, shrink or flat or slower growth. But overall Timu in Europe grew 23%, which?

Speaker 2:

is solid 0.5. Which is solid.

Speaker 1:

That's solid, and you know what that means, Because that's definitely faster than the market at large. So that means that they're unfortunately they're concentrating market share still.

Speaker 2:

Yes, what a surprise. They eat our lunch. Yeah, and I don't like it In Germany.

Speaker 1:

that's their number one. They were a bit slower at 18%, but I'm Still crazy. But Germany is particularly flat right now, so I'm sure they're outperforming the market anyway.

Speaker 2:

And if I were the team OCO God forbid Germany would be somewhat, a little bit surprising, because Germany is the biggest market and that's, out of the biggest markets, that's the market which is growing the slowest right.

Speaker 1:

Yeah, for the rest of the top five they're at about a third. So I mean that's a lot.

Speaker 2:

A reason for the rather disappointing Amazon journey. I mean disappointing, yeah.

Speaker 1:

I don't know. I mean, I think that Timu has a particularly poor reputation among consumers and in the media in Germany. I could be wrong, but that's the feeling that I have. That's it, and it can also be that there is this slower overall growth in Germany, perhaps.

Speaker 2:

But you know how it is. I mean, look at your network. How many guys are bashing Timu seven days a week and twice on Sunday? Yeah, and they still buy on it. Or seven days a week and twice on Sunday yeah, and they still buy on it yeah, but I'm with you. I think Germany, the German consumer, is probably the most critical about these Chinese mega cap companies or going into European markets. Let's see. So bottom line is they're growing.

Speaker 1:

Yeah.

Speaker 2:

They're growing rapidly. Yeah, unfortunately, and faster than the market. That means the European retailers have to buckle up. Exactly, yeah, they're growing rapidly. Yeah, unfortunately, and faster than the market.

Speaker 1:

That means the European retailers have to buckle up, exactly, yeah Well, on that very cheerful note, that was another episode of Growing E-Commerce.

Speaker 2:

I enjoyed it so much, mike, looking forward to the next one.

Speaker 1:

Likewise.

Speaker 2:

Henchik, thank you, thank you very much. The voice is happening again.

Speaker 1:

Should I drop it down?

Speaker 2:

low, yes, this happening again. Should I drop it down low, yes, please oh hey, now it gets cheesy.

Speaker 1:

Thank you for listening to another episode of Growing E-Commerce and for tolerating us. We enjoyed it. This podcast is brought to you by Smarter E-Commerce, also known as SMAC. As always, you can learn more at smarter-ecommercecom, and we really appreciate it if you support us through ratings, reviews or word of mouth. That's going to do it.

Speaker 2:

That's going to do it.

Speaker 1:

See you next time.

Speaker 2:

Bye guys.