August 23, 2024 • 54 mins
Mike Ryan, Head of Ecommerce Insights at Smarter Ecommerce, talks to Gordon and Ger about the potentially unholy marriage between data and advertising automation that can erode your control of your ecommerce performance.
Specifically, Mike takes the boys through how to manage ecommerce marketing tools like Google's Pmax, which, left unchecked, can give unintended or unwanted results. How do you tame it and make sure it delivers in line with your strategic business goals?
Inevitably, it comes down to what data you feed it. But that's not always easy when tools like Pmax seem to be shifting to automated, unstructured data inputs and the ability to source data themselves.
Is control slipping from our hands?
If it is, maybe critical analysis of our KPIs will help us get our hands on the levers again? But which KPIs? ROAS? MER? CPC? KFC?
And just when Mike and the boys seem to be heading into calmer waters, the big 800lb Gorilla that is Temu beats its chest into the discussion. But Mike is ready and shows how Temu's seemingly bottomless ad spend is actually benefiting online advertisers in some ways.
Join Mike, Gordon and Ger on an epic rollercoaster of a quest into the realm of ecommerce marketing.
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
(00:01):
This is Functional & Fabulous, the omnichannel podcast where we unbox tales of online retail and digital transformation.
In this episode, Ger is on an epic quest...
I love a quest, I'm on an epic, I'm on an epic quest.
And he's assembled a crack team using a rigorous selection process.
(00:22):
There's Gordon...
Are you good at being responsible?
I'm terrible at being responsible.
And Mike...
The data-driven pushback hipster.
Join them in their bid to find the Holy Grail of Google-driven ecommerce.
How do we enjoy PMax responsibly?
Along the way, there will be traps...
(00:42):
The monkey who sticks the paw into a canister, grabs some nuts and then finds that he can't get his paw out.
There'll also be riddles...
Horizon is just always leaning forward.
And violence...
Get your head cut off waiting for that to come home.
So buckle up for the ecommerce podcast adventure of a lifetime.
(01:04):
It will be very interesting, I think.
This episode of Functional & Fabulous is brought to you with pride by StudioForty9, retail ecommerce experts, omnichannel growth consultants and cut-through performance marketing specialists.
StudioForty9, where your digital retail success is built.
(01:25):
So welcome to another episode of Functional & Fabulous, where we delve into the ever-evolving world of ecommerce and retail.
Today, we have another extremely interesting and experienced guest, Mike Ryan, who is Head of Ecommerce Insights at Smarter Ecommerce, based in Austria.
Mike's dad grew up in Southie, in Boston, and Mike tells us he's Irish at heart.
(01:46):
And with a name like Mike Ryan, he doesn't really have the choice.
But he's with us today because he has an exceptionally deep understanding of ecommerce and digital marketing, which he leverages to provide invaluable insights and create innovative solutions for his clients.
He's a generous sharer of information, so it's well worth following him on LinkedIn, where he's a regular poster.
And also checking out his own podcast, and I'm doing it again, promoting other podcasts, but this one is called Growing Ecommerce.
(02:13):
Very, very interesting, some great leading industry experts.
But on that podcast, Mike explores e-com, digital marketing and data science.
Mike's expertise has been recognised by leading publications such as Adweek, Financial Times and Bloomberg, and he has been a prominent speaker at industry conferences.
We're delighted to have him here with us today.
(02:33):
I'm delighted to make my intros longer than ever, and we're dying to learn more about how to enjoy performance marketing responsibly.
Mike, it's great to have you.
Yeah, thanks so much for having me on.
Thanks for the very warm introduction and also promoting my podcast too.
Don't worry, I don't think it's a zero-sum game.
I think most podcast listeners are podcast junkies, and they'll listen to more than one.
(02:57):
It definitely is, and I got a slap on the wrist not so long back from our producer for promoting other people's podcasts.
Surely not, he said, but anyway.
Yeah, it wasn't a slap on the wrist from me, I'll hasten to add.
Welcome to the podcast, Mike.
I'm very excited to talk to you today about enjoying PMax responsibly, amongst other things.
(03:19):
Amongst other things.
Also, the fact that you're a data-driven pushback hipster.
You're the only one I know, I think.
You were pushing back on data-driven decisions since before it was cool.
That should be my LinkedIn headline kind of thing, you know?
No, I mean, I don't want to say...
Yeah, I just push back a bit on...
(03:39):
We shouldn't be kind of beholden to data.
It should be a tool in the toolbox, but we shouldn't forget other tools in the toolbox.
Absolutely, and as we were talking about, I really enjoyed a talk I heard recently.
Aimee Connolly of Sculpted by Aimee spoke at a conference, and she talked about data gut, which I thought was a really nice kind of concept where, you know, of course, we're looking at data, but we're also checking our gut, we're checking our instinct, you know, checking common sense.
(04:09):
Is the data telling us the right sorts of things and so on?
But one of the things that, you know, we covered quite a bit when we did the pre-recording chat was PMax.
It is kind of one of your areas of expertise, and I love that whole thing about getting to use it responsibly.
Can you tell us where that comes from?
(04:29):
What is a non-responsible use of PMax?
Straight into the good stuff.
Oh, yeah.
Let's go straight in.
Let's start on PMax.
Okay, I just came back from lunch, but not my...
Good, you got my blood running again.
Good.
Super.
We won't do the stand up and wave your hands in the air thing.
We're not going to make you go through this.
(04:50):
We have to rewind a little bit because I've been working with ecommerce ads and Google Ads for a long time, and that in the past was shopping campaigns, standard shopping campaigns.
And now shopping is just kind of one inventory.
Google likes to rather view it as inventory than as a campaign type, although the classic campaign type is still there.
(05:11):
They built this more expansive kind of campaign type called Smart Shopping, and that was already a while back now.
They kind of evolved even further into this thing called Performance Max.
And just in case you're not too familiar with this technology, the basic premise of Performance Max is that it's a buy-in service to all of Google Ads inventory.
(05:33):
So not only the shopping inventory, but also search ads, classic text ads, display, Gmail, YouTube, everything.
It's all kind of possible from there.
And in fact, it's, if you just run it out of the box the way, kind of just turn it on the way Google intends, then you don't really have any control either on that.
(05:57):
Google sort of promises that they're going to kind of handle matching, like finding the audience, matching your products, and just building ads, really a very highly automated campaign type.
So when Smart Shopping came out, there was some pushback from the market and certainly from me too.
There were many things I didn't like about the way that product was designed.
(06:20):
It took away a lot of promotional controls and levers.
It took, like things, like involving bid adjustments and different kinds of adjustments that you could make, scheduling that you could do.
And it also took away a lot of reporting that people were used to, like with a classic search campaign, you are targeting keywords.
(06:46):
So someone out there is searching for something, like, they'll make some kind of commercially related query, like, let's say buying Nike shoes just to make something done.
And imagine that Nike is a keyword that you have, just to use a dirt-simple example.
And shopping now that you're not really targeting keywords, but you are, Google is already matching your products to queries.
(07:10):
And you in the past could see those queries, what they were in search term, of course.
So that's just an example.
People found that really valuable because sometimes Google would be matching you with some strange search terms.
And so that was another type of control you had.
You had the transparency to see what you were matching for.
And you had the control to negative those things, like to say, no, I don't want my products to match for a given search term.
(07:31):
So Google just stripped away a lot of stuff.
Some would say dumbed down the campaign, I pushed back against it a lot.
But you know what I found was that people adopted the technology anyway, with or without me.
They heard what I had to say, but people were using this stuff anyhow. And so that's what I mean by...
How disappointing...
Well, I think Google do an incredible job with their sales teams and with their account managers.
(07:56):
If you ever go to a Google conference, every single one of them will be on message for that conference.
So it's just drilled into you.
For sure.
And I mean, something else that we observed back in the days with Smart Shopping is that often, like, the operational teams, like, the practitioners, also had their reservations or concerns about Smart Shopping.
(08:17):
And you'd see, at least in, it depends on the account size and stuff like that.
But Google's team would go in a top-down approach because they'd reach someone higher up the food chain, let's just say a CMO, for example, who's not into those practitioner level details.
And they could kind of reach that person and then make the change happen from a top-down way.
(08:40):
But the change did occur.
The adoption was pretty massive after a couple of years.
And with PMax, I made the decision, okay, I know, because I've seen this story before.
People are going to use this technology.
In fact, Smart Shopping was migrated over to PMax automatically.
So by default, people would be using this technology.
(09:01):
And I want to help people enjoy it responsibly.
And that means several things.
But I think what's really important is that PMax campaigns are not necessarily going to be incremental in the terms of, like, audience incrementality.
You can think about that.
Like, it tends to reach into remarketing, brand traffic, basically warm or existing traffic.
(09:27):
Which can have lower incrementality and kind of that classic marketing incrementality term.
But also in terms of, like, business incrementality, not every product has equal importance to your business.
Some things are going to have their own natural demand that it's easier for Google to follow.
So PMax will tend to sell into bestsellers.
(09:48):
But you might have a product that doesn't perform as well in platform.
Google's a bit less interested in promoting that because it's going to be harder for them to extract, like, platform value for you.
But it might be higher margin for you, or it might be overstock, and it's behind sell-through rate and you need to clear it.
So these are the kinds of things you have to tell Google what's important about your business.
(10:12):
And you also have to kind of work a bit to get the campaigns to work harder for you.
That's what I mean by enjoying responsibly.
So that was a pretty long answer.
Thanks for that.
No, I think it's okay.
There was one just to summarise a little bit of that and some of the reasons that we're talking about the kind of responsible piece of it.
You know, you use the term, it's like the monkey's paw.
(10:33):
The monkey who sticks the paw into a canister, grabs some nuts and then finds that he can't get his paw out.
And obviously the trick is the monkey has to leave go of the nuts in order to get the paw back out.
And what's happening is the PMax approach, if you just turn it on, it's going to follow the path of least resistance and it's going to try and make whatever sales it can.
(10:55):
And those sales may or may not be good for you.
You may or may not be about to sell anyway those sorts of products, and therefore you need to kind of shape the story for Google to understand what it is that you're trying to sell.
So, like, let's say we agree with all of that.
How do we go about doing it?
I was going to say that will be music to marketers ears of, yeah, okay, we all accept PMax is problematic if we don't do that properly.
(11:24):
So the obvious next question is, let's enjoy it responsibly.
How do we enjoy PMax responsibly?
Yeah, I mean, I think it's definitely a case of be careful what you wish for.
People wanted performance that is basically efficient and scales.
(11:45):
And that's hard to actually do because really incremental marketing is typically not going to look as good in platform.
So what you need to do is just help bring in, there are ways to bring in control and there are ways to tell Google what's important to you.
Let's just look at the topic of telling Google what matters to you because, for example, if we're talking about that thing about a sell-through rate, Google doesn't know that information, that's off-channel information.
(12:14):
They can't know that.
They know your stock level or they basically know your stock status of yes or no.
Or similarly, let's imagine that you've got plenty of availability on product in your core sizing, and another product is kind of thin on the core sizing, and there's just these less common sizes like an extra small and an extra large, or something like that in there.
(12:39):
This is again, this is something that could be important to your business.
It could also impact the campaign and Google doesn't know about that.
That's not their fault.
So what you need to do is bring that data into your campaign.
We also talked about margin, for example.
The challenge that I see, people tend to pick a battle here.
(13:05):
They'll find one or the other thing and try to bring that data in and try to optimise toward that.
It's very simple to just bring the data in that, Google has these things called custom labels in the feed and you can relatively easy bring in a piece of data, but then actually making that work for you is a bit more complicated.
(13:27):
So, like, in the past, these shopping campaigns worked at it.
You could really optimise down to a pretty granular level, but in Performance Max, you really need to have a more consolidated account structure.
You want to have these kind of larger pools of data and campaigns.
You don't want to split it up too much.
(13:48):
And what I like to do, I basically think of these campaigns, typically someone's going to make a product-type campaign or a brand campaign.
Like, they'll have a campaign for ski jackets, for example, or for a given ski jacket brand, let's say.
But what I think is more instructive is to take these campaigns and view them as just like an empty vessel or a vehicle.
(14:15):
And kind of be agnostic to what's in there because the thing about the campaign is that you can set a budget and you can set an efficiency target.
And those are the two main things that you can do to pace yourself on Google Ads.
The budget needs to be there to spend and the efficiency target will determine the aggression level.
(14:35):
So what happens then, if you make some of these buckets with different budget levels, different efficiency targets, you have these different kind of, like, pods or vehicles that are gonna push products more or less aggressively.
And then it's a matter of making sure that products are in that right campaign, that they're getting the right promotional intensity, basically.
(15:00):
So what we like to do is basically create a product score that is going to look at a product on any number of dimensions, however many are important to you.
And products could end up in different of these buckets, these campaigns for different reasons.
For example, one product, it might be important to you to promote because it's behind sales plan.
(15:25):
And another product might be important to promote because the margin is high.
Similarly, you could have kind of a bestseller, something that's flying off the shelf.
And you might actually want to dampen that because Google is going to see that it's performed well and it's going to keep juicing it.
But you might actually want to dampen that because actually you have plenty of organic volume on that and you feel like you're cannibalising that.
(15:48):
Or it's selling fast and well, but it's actually not your highest-margin product.
There could be any number of reasons why.
You know your backorder status or your on-order status.
It's just about thinking about these things and making sure that Google is pacing the right products.
So that's one way.
Yeah, it's quite interesting.
So I suppose just to kind of summarise a little bit, you know, you could obviously send through the exact information about the product margins and all the rest of it, but then you've got the big problem.
(16:17):
So you can, let's say, enrich the data quite substantially, but then you've got the problem of how do you explain to Google what that data means?
So what you're saying is, rather than sending through, let's say the margin on this one is 19.5, the margin on this is 37.5 margin, whatever.
And then having the big problem of trying to explain to Google which ones you're trying to push, you basically would say maybe a bucket for high-margin, low-margin, medium, and then you would indicate to Google, okay, well, it's the high-margin ones I want to push first and foremost.
(16:46):
And it's kind of that taking a guardrails approach, I guess, and ensuring rather than giving the exact instruction saying, okay, well, here's the guardrails when you're driving the car.
Don't go off the side of the guardrails because that's where the cliff is.
And that's the general gist of it, right?
Exactly.
I mean, and Google will never understand if you, in this example of a margin, the margin topic, you can have a campaign for each margin class and Google, it's not like they can read the campaign name and know why that's important or which one.
(17:18):
You have to just speak in its language, which is the budget and the efficiency target because Google's AI is not strategic.
It is not aware of any of this stuff.
It's just looking at it on a per auction basis.
It's got a month's budget in front of it or a day's budget.
And it's going to basically look at the expected click-through rates and the expected conversion rate.
(17:46):
It's very smart in its own way.
I don't mean to dismiss the technology.
It's very effective at that job, but it doesn't have a wider view than that.
That's our job, still.
Yeah, I always think that sometimes we lose sight of the fact that Google's job for itself is to juice enough revenue for Google, or juice the maximum revenue or maximum monetisation of each potential click.
(18:16):
So ultimately, it's still an auction.
Google's still going to be placing results in front of consumers or potential consumers that they are most likely to click on and that's going to earn them the most money.
So we have to have that in the back of our heads when we're planning our campaigns.
Totally.
(18:36):
It's a marketplace of supply-and-demand and it needs to look after the match rates between that.
And for both, because everyone wants a high match rate.
The advertisers want their products to be seen and clicked and potentially purchased.
The consumers want to see relevant products that they might want to buy.
(18:59):
And of course, yeah, there's the part you're talking about.
Google wants to get their clicks.
That's, they build on the clicks.
So it's absolutely the case.
And the thing about this auction is that Google is...
When we're using the smart bidding technology, which almost everyone is using, Google is like, they have the auction environment.
(19:20):
They're the auctioneer and they're the bidder.
It's quite fascinating.
Like they're bidding, you're giving them permission to bid on your behalf, but they're also the auctioneer.
There's not a lot of auctions that work that way where the person collecting the auction is also the person doing the bidding.
It's quite fascinating.
And moving a little bit along the same lines, but moving the story a little bit further along, feed management and exclusions and so on in feed management and that sort of thing.
(19:51):
Where do you see that going?
Well, let's see, which part?
The feed management part?
Well, hold on.
It's funny.
I was just thinking about this and writing about this yesterday.
Something that comes to my mind is that Google is working to try to...
That's another area that they want to automate more.
(20:12):
For someone, in case someone is less familiar, the product feed is just basically a place where you provide structured data so that Google understands your catalog and what you're selling.
So it's explicitly structured data with attributes so they know exactly what they're dealing with.
And a goal of Google for years and years now has been for them to be able to implicitly know.
(20:37):
So without receiving that structured data, for them to, for example, get the URL of your webshop and they can crawl it and they can extract everything that they need to know.
So basically, you could hand them more or less a credit card and a URL and say, let's go.
This is the direction that they're going a little bit further now.
(20:58):
They have a new version of their Merchant Center, where you upload your feeds and your product data, usually.
It's called Merchant Center Next.
And they're doing away with the term feed, interestingly.
That's how they've been gradually downplaying the word keywords for years and years.
Feeds have been absolutely at the core of what the Merchant Center is and how Google shopping is feed-based advertising.
(21:25):
So what are they, where are we going without the feeds?
What are they doing?
They're talking now about data sources.
And the reason why, this allows them a little bit more flexibility.
It's a bit of a broader definition because now they view that structure data as basically data that's provided by you.
(21:45):
But they have another category of data that they call Found By Google.
And this is data that they get from crawling your website.
And so in the end, this has taken a step forward due to large language models and advances in visual AI.
They're able to not only crawl your website in a classic way, and by the way, you'll also have structured product data available on your website.
(22:12):
But they'll also be able to use language models and visual AI to also just understand your website.
And so the feasibility of this idea of kind of hand your credit card and URL has taken a massive step forward.
But it also, it's another thing right now, you can exercise a certain amount of control through your feed, you can bring in data, you can exclude data and so on.
(22:41):
And you will have perhaps a bit less control over that in the future.
It's not quite clear the direction where that's headed.
Some people won't like this, but it's always the case.
If you were a betting man, would you bet the deprecation of that control over time?
I think you'd have to bet that's where we're aiming for.
(23:02):
That's what's gonna happen.
Yeah, it's certainly the direction of travel.
And people have been talking about the death of the keyword since, to the best of my knowledge, since 2014.
So we're now at a decade of the keyword dying.
And I think it's probably something like that with feeds.
I think that it could take quite a while, but it's gonna be a matter of prioritisation, the level of support that's available, how Google will structure features around, they'll gradually steer us a certain way.
(23:32):
But I can't say that it's going to be tomorrow or next year.
Because it feels like, certainly here, that the whole concept of managing the feeds and crafting and sculpting them so that you'll get an outcome that you're looking for, is kind of immature here at the moment.
I don't know, Gordon, what you think?
(23:53):
In a previous life, we spent a long time curating our feeds, trying to manage newness, seasonal product, trying to exclude products that had, like, high likelihood of sell-through in season, all those kinds of things.
That was a big focus, I think.
But more broadly from talking to a lot of retailers, getting the feed in or getting Merchant Center, having all the products in there is probably the step that they're at still.
(24:26):
That's what I mean, yeah.
That's still kind of, like, we're getting the feed in and that idea of controlling the feed is probably a little bit new.
So I think it would be very disappointing for people to get to a stage where they now know how to control the feeds and then they have that carpet whipped from under them.
I think that's constant innovation, though, isn't it?
(24:47):
Innovation in inverted commas to keep all marketers on their toes.
Yeah, I mean, the pace of change is, it's exhausting, to be honest.
Definitely put a few grey hairs on my head.
Not enough yet.
I've got a lot more than you there, Mike.
So we won't talk about that.
(25:07):
So just to move along a little bit, we talked a little bit about KPIs and the importance of KPIs.
Also, we talked about attribution, everybody.
Attribution is such a big subject at the moment and we might, I don't know if you want to touch on it or not, but certainly KPIs, I think you had an interesting comment around ROAS and I suppose, again, we're kind of in a stage where people use ROAS nearly as a, like probably too religiously as a KPI maybe, whereas it's still an important KPI.
(25:35):
But tell us what you think about ROAS as a KPI.
I think there's a saying out there, I've heard this before, you are not your ROAS.
I certainly believe that.
I think that ROAS is a very useful KPI, first off, and I don't want to just, like, knock it or something.
I can understand why it's become so popular.
(25:56):
It's available at so many levels in an account, and all the way down to the individual items, and it's so versatile.
It's simple to calculate.
It appears to be simple to understand, but a couple of concerns I have about that, I think that people often kind of conflate it or substitute it with ROI or profit, and I think Google very much welcomes this.
(26:21):
You'll often see in their sales and marketing collateral that they'll talk about ROI when actually they mean ROAS, and these two things are certainly not the same.
There's a risk there.
It's when you've got a proxy metric and you make that mental substitution, that can be quite dangerous because proxies are not perfect, they're inaccurate, and particularly ROAS is available.
(26:47):
You can view it in platform.
You could view it in an analytics tool as well, but that platform only knows what the platform knows and this starts to rapidly move into themes of attribution and so on, but I think it's really important to remember that this is one metric and you better triangulate it with more.
(27:09):
I think you don't want to just lean on that one metric, or let it be kind of your be-all end-all because it is only a proxy, and I actually think if you can, it's funny because it's the primary lever that's available to you in your campaigns anymore is to adjust your ROAS.
That's how you're going to pace your campaigns in large part, but it's also supposed to be a business objective.
(27:36):
If it's really a business objective, then can it be a tactical lever?
You shouldn't just play with your strategic goals, and I actually would rather kind of decouple that and say, yeah, it is actually more of a tactical lever than a business objective.
You need to be actually measuring and setting your strategy on other numbers from other systems, or at least looking at those as well.
(27:58):
What other KPI, like if, you know, obviously, so ROAS for me is, it's a directional, you know, it's a trend kind of a, you know, we're trending upwards, we're going down, you know, it's very, very useful.
It's great to, it's a great measure to use with the, let's say, the C-suite, that kind of a thing.
But what other KPIs do you think are more important or which other ones would you use, as you say, to triangulate the kind of the real story?
(28:25):
I mean, speaking of, like, direction of travel, I think MER is quite popular in that use case these days...
That's my favourite...
Marketing efficiency ratio and it's basically looking at your total sales and your total advertising costs, and it just gives you a much higher-level picture of what's going on.
(28:46):
And it partly solves for things like the fact that you could have Meta Ads and Google Ads both claiming full credit for the same purchase, or stuff like that.
And so this helps to, in a way, to deduplicate that.
So that's certainly a popular one.
I, my only problem with that, I'm always willing to change my mind and stuff.
(29:08):
I was a little anti-MER at the start and it's, I just saw people using it for things I don't think it's appropriate for, like, to make really detailed channel tactical changes or saying like, oh, I made changes to this ad set in Meta.
And then I saw this change in my MER.
(29:30):
And I mean, yeah, if like 90% of your spend is Meta, that might be credible.
But if you've got to, you could just read whatever you want into that.
And then it's too easy for our own confirmation bias to step in and so on.
So we have to use MER responsibly.
Yes, let's use everything responsibly.
We gotta be responsible?
It's, I mean, we're not good at being responsible.
(29:51):
Are you good at being responsible?
I'm terrible at being responsible.
MER for me is a good strategic.
Use MER like a maniac...
Don't use MER like a maniac.
It's a good strategic measurement to track your direction of travel over time.
But I would agree with you there, Mike.
It's not a, it's not a tactical measurement, and any tactical decisions made off the back of MER performance, you should be questioning.
(30:19):
But what about, are there any more detailed metrics, like, I suppose, further down the funnel, more tactical metrics that people should be focusing on?
I mean, well, while we're talking about it, I would say that profits, measurement, profit optimisation are becoming a lot more popular than they were a couple of years back where people were typically very focused on growing revenue, maybe at all costs.
(30:47):
This can, of course, have downsides.
Also, a lot of brands were, you know, I think of Shopify brands, for example, where there were these cool little Shopify tools that helped you understand LTV a bit and it seemed all good, but you can really stick your neck out on LTV and you can really get your head cut off waiting for that to come home.
(31:11):
So now, it's a certainly more conservative financial environment, and we see people wanting to optimise towards near-term profit that's much more fashionable these days, which I think is healthy, and also down to further contribution margin levels that consider, you know, whatever the case might be, different sales tax, different shipping fees, and so on and so forth.
(31:39):
So that's another one that I definitely shout out.
But also, I'm gonna call out one tactical thing that I think is, like, for years and years, people talked a lot about CPC, cost per click, and these are pay-per-click channels, Google Shopping, Meta.
These are, like, depending on the format, these are pay-per-click channels and you get why people are focused on that.
(32:06):
But what I like to talk about more these days is value-per-click or revenue-per-click, typically.
Depends what value you're tracking.
It could also be profit-per-click.
And the reason why is, like, I mean, honestly, Google owns the CPC at this point.
Greater than 80% of the market is using automated bidding.
(32:27):
And so even if you're not, when everyone else is, it doesn't matter because the bid heights are somehow dictated by Google at this point.
And it's become an observation metric.
In the past, when people were much more active in their bidding, this was more important.
But you also have to consider, like, with a campaign type like Performance Max, when you make a change to your ROAS target, it won't only impact your cost-per-click anymore.
(32:57):
It might also impact your channel mix.
They might realise if you set the ROAS a bit lower, Google might realise, oh, I can start expanding into more video at that efficiency level or whatever the case might be.
So it's just more complicated.
And in the end, you know, if you're obsessing over cost-per-click, you are just talking about the channel as a cost centre at that point.
(33:19):
If you're talking about the value-per-click, there's much more that's contained in that.
Your order values are contained in that.
Your conversion rate is contained in that.
Those are the things that produce that.
And those are, like, you can really optimise stuff and see your value-per-click improving.
And it's then focusing on the channel as a value centre.
(33:39):
So I think that's a mindset shift that some people would be well-advised to consider.
It's a challenge.
One of the things I like to look at is that kind of offset between, what's it costing us to acquire a click and then offsetting that against the value-per-click at both a revenue and a profit basis.
(34:00):
So I know that if every click I'm getting onto the website's earning me $1, but it's costing me $1.10 to get every click onto the site, then I have a problem.
Yeah, I mean, and that is, if you hold the value-per-click and the cost-per-click against each other, that's basically return on ad spend at a per-click level.
(34:21):
So it totally makes sense.
But in principle, Google should be caring for this.
And you can't do much to influence the cost-per-click because Google's got that.
You can influence the value-per-click by improving your on-page experience, by improving your ad targeting and all kinds of stuff.
(34:42):
Great.
One of the areas, and a complete kind of shift, I haven't even got a good segue for this.
Pivot, we're gonna pivot.
I've not got a good pivot for this one.
I'm putting on my neck brace.
Since we've hit a couple of trigger points here, one of the areas and one of the businesses in particular that you've been quite vocal about.
(35:04):
Don't say it...
You know what's coming.
It is a brand that seems to be pushing up CPCs and CPMs for everybody.
And it helps a lot of people shop like a billionaire.
It's our friends over at Temu.
And we'd be really interested to hear your take on Temu generally.
(35:28):
Where to begin?
It's a fraught topic for me...
This also, Mike, is the wrapping up topic, by the way.
Yeah, look at the time. Oh, what a lovely note to end on.
Thank you, gentlemen.
Thank you for dumping me with Temu.
No problem.
I couldn't resist.
No, by all means.
(35:49):
I mean, I've been watching Temu for quite some time.
And they've been active here in Europe for about a year now, where they very rapidly came online through advertising.
And we've seen that in Meta Ads.
You know, you can check the Meta Ad library and see how many ads they're running daily and track this over time.
(36:12):
And the number is just insane.
It dwarfs all the other competitors.
And it's just gone up and up and up.
And similarly, something that I've looked at is basically the percentage of Google advertisers that face account-level competition from Temu.
And they just turned on a fire hose of advertising dollars in a way that's, I would argue, unprecedented.
(36:35):
Because I can look back in history and report those same numbers for Amazon, for Wish, when they were around.
And both of those companies had huge and massively aggressive market entries.
But Temu just takes the cake.
Now, I mean, one thing that I would say is that it's a bit different.
(36:55):
When Amazon turned on ad dollars in a huge way like that, this destroyed cost per click and the unit economics overnight.
It was a huge problem.
But now that they're still there and that damage in a way is almost done.
So Temu coming in, it does affect unit costs a bit, but it's almost more about a war for visibility.
(37:25):
And I found some interesting things.
So this company started fluctuating their ad spend, for whatever reason.
Like they were always just going up and to the right, spending more, more aggressive, always, always.
Until quarter one of this year when, a couple of times, we saw some pretty big dips.
(37:45):
We just saw volatility in their spend.
And I love this as an analyst because when there's volatility, there's change.
And then I can start looking for the correlation between that change and other changes.
So I found that, yes, they definitely hurt CPM.
That's true.
But interestingly, cost per acquisition improves when Temu is more visible.
(38:12):
And this is totally weird.
My hypothesis for this, I have a twofold kind of thing.
One of them, part one, I call the shit sponge.
I think that when Temu is very active, they're actually kind of soaking up shitty auctions because they'll actually, they might crowd you out of more generic auctions that you didn't want to be in anyway.
(38:36):
And the other thing is, basically, I think it's called a contrast bias.
If you take someone with a really nice polished appearance and then you take someone with an average appearance and someone with like a very poor appearance, the average person is going to look worse compared to the really polished person.
(38:56):
But the average person will look better compared to the kind of sloppy person.
And I think that consumers are not that trustful of Temu, or many consumers are not.
And when they're crowding out these auctions, it can actually kind of have a halo effect for your brand or for your offer.
(39:18):
So it's very fascinating.
That is very interesting because we are looking at it now in the context of Ireland and we have Temu, let's say, growing share into the Irish market, and we have Amazon.ie opening up in Ireland next year.
And I suppose people are nervous what that's going to do to cost per click and cost of acquisition and all of that side of things.
(39:44):
We've seen them come on in Netherlands, for example, in France in recent years.
It's actually a bit further back now, but we've seen them make those market entries.
And yeah, I mean, it's certainly not a good thing when Amazon or Temu or anyone enters your market, from an advertising perspective.
(40:09):
It'll also be interesting to see how the adoption is among domestic Irish sellers.
Because for example, in Netherlands or France, we saw that Amazon was really getting kind of seeded or bootstrapped by Chinese sellers.
And it took a bit longer for local sellers to kind of adopt the channel.
That's quite an interesting point, actually, of what other reactions have you seen?
(40:34):
Like, if we take The Netherlands, what were the other key learnings that retailers would have taken from that market entry?
Yeah, I mean, in The Netherlands, for example, there is a very dominant marketplace, homegrown marketplace called bol.com.
(40:56):
And it just was kind of a war between these two titans.
And I think a lot of companies had probably frenemy relationships with bol.com because they're probably selling there.
They're probably also competing with bol.com to a certain extent.
And yeah, Amazon comes in.
(41:18):
And I don't think, I think people would rather compete against bol.com than Amazon.
Similarly, like in France, who has their seed discount and these big dominant local retailers.
And France is a very, a nationalist market.
(41:40):
And so I think Amazon had a bit of a tougher time there, but yeah, it depends a lot on consumer preference.
Like, I don't know the extent to which Irish consumers are already shopping from Amazon.co.uk or to what extent they, I don't know how committed they are to purchasing locally.
(42:05):
You'd know more about that than me.
It will be very interesting, I think, particularly the Amazon.co.uk. I think Irish shops have been using that in anger for a number of years.
Yeah.
I mean, in my market, I'm in Austria and I'm using Amazon.de in anger.
(42:30):
Yeah.
I think it would be similar to if you were going to get an Amazon.at at some point.
I don't even know if that's on the horizon.
I don't think so.
I don't think so.
But at some point.
Okay.
So unfortunately, we've come up against the time.
It's an interesting, I mean, that whole area, obviously it's an interesting question, how the Irish market is going to react to these two giants coming into the business.
(43:02):
It's going to be a transformative shift, I feel, in Irish ecommerce.
And definitely it will be a challenge.
So we'll be keeping quite a close eye on that, Mike.
Thanks so much for your insight today.
We got a couple...
We didn't get to the data-driven decision, hipster pushback piece.
(43:25):
But I'm looking forward to seeing that on your LinkedIn profile soon.
And thanks so much for joining us.
Yeah.
Thanks very much for having me.
Thanks for sharing with us.
We'll be enjoying our PMax responsibly.
Cheers, Mike.
Well, let's enjoy PMax responsibly and all other media mechanisms.
(43:47):
We never got onto Advantage Plus, actually, which I'm a little bit disappointed about.
We ran out of time.
I don't know about you, Gordon, but I find this time management business, I mean, maybe we need to just focus on how people need to listen to long-form podcasts and not try and make it, you know, nicely bite-sized for the commute.
(44:07):
We're trying to get it commute-sized, but people have got interesting stuff to say.
Yeah, but maybe people have longer commutes now, post-Covid, you know, it's getting, it's taking longer to get to work.
Or people have actual commutes now.
Maybe they have actual commutes, exactly.
I think this is such an interesting area because he has this crossover between the data science and obviously then the commercial aspect, you know.
(44:31):
And it's always interesting to hear a little bit about the data science.
There are certain areas that I would have loved to get into in a bit more detail, but you just don't cover everything.
I think we talked a bit about the, you know, the path of least resistance that PMax provides us with.
And obviously, like as he mentioned, that challenge of scaling efficiency because they're kind of, they're competing.
(44:54):
They're competing things.
You're looking for, I want more, I want more, but I want it kind of cheaper and cheaper, more efficient.
It's almost like you can't say, I want to go and search for new customers, identify them, attract them and then interest them.
But I'd like to do that at the same efficiency level as brand search, please, is a very unrealistic ask.
(45:19):
Yeah.
So you have to manage that.
Yeah.
And it's very difficult to understand how to do that with PMax.
So Mike mentioned confirmation bias here, and I was delighted to talk to him because it feeds into my confirmation bias that PMax is too much of a black box.
Yeah.
And to be able to listen to someone who has actual strategies of how you manage PMax in order to balance those competing objectives, because of course you want your marketing to be as efficient as possible.
(45:53):
Yeah.
Of course you want to service those customers that are already partway through the buying funnel or about to enter the buying funnel.
And then you have to go and look for that growth.
And it's a tool, it's a fabulously powerful tool to be able to do that if you can drive it.
(46:14):
But I think just to use it as a, almost plug and go...
Yeah.
...isn't going to get you the results that you need.
And to get some practical advice on that, I hope that people listening will be able to really take something from that because it must be confusing people.
I think it is.
I think what we typically see is you turn it on and then kind of walk away whistling.
(46:37):
Panicking.
Definitely the idea of setting up those buckets or guardrails, whatever you want to call it.
But, and using that information in the feed to inform Google or to, I suppose not inform, but like influence, I guess really, or direct what actually happens.
You know, it's a really important thing.
The other area that was interesting, and I guess within that is, and we didn't get to speak about it, but it was maybe the importance of testing.
(47:04):
And in particular, because what I noticed when we have situations where we're trying to optimise all the time, you get to a point where you reach the maximum optimisation you can do in a particular area.
So I'm going to explain this.
It's kind of like a mountain range, and you've got a series of hills and then you've got a series of mountains, and then you've got, let's say, Everest and a peak and whatever.
(47:26):
Do you love a quest?
I love a quest.
I'm on an epic quest.
So I'm on an epic quest to find the ultimate summit and I'm going up a hill and my instruction is optimise my, you know, how high I travel.
So I'm up and going up the hill.
I'm going up the hill.
I get to the top of the hill.
That's the optimum that I can do on that particular hill.
(47:48):
So this is my local maximum.
My instruction doesn't tell me, alright Ger, now go back down the hill because maybe that's not the highest hill you can actually get on to, or the highest peak.
And that's where we run into these problems because if our instruction is always do the best and climb the highest, we're not going to go to make the effort or take the risk of going back down the hill to see if there's a higher peak elsewhere.
(48:10):
And in order, what you have to do in order to do that is, you've got to look at, let's say, testing and testing budget and risk budget where you say, righteo, I'm going to jump off of this hill.
I'm going to drop myself some other part of the valley and then I'm going to start maximising and optimising, and then I'll see if I reach a higher hill.
So that idea of having a budget to play around with, to take risks, risks that are not going to be looked upon favourably by the CFO or some other people on the team, it's also still an important thing because you don't want to optimise yourself into, let's say, Carrigtwohill, or sorry, Carrauntoohil.
(48:47):
You don't want to optimise yourself onto the top of Carrauntoohil.
You want to optimise yourself onto the top of the Everest, the Alps or somewhere else.
What do you think, Roger?
Do you agree with me?
Roger agrees with me.
Look, he's totally...
He wants to optimise himself onto Carrigtwohill.
Ha, ha, ha, ha...
Whereas you want to optimise yourself onto Everest...
Oh, absolutely.
Along with everybody else.
There's shocking pictures from Everest, by the way...
(49:09):
Oh, the queues...
But we're not even going to go there.
Okay, let's go somewhere else.
Okay, we'll go in a slightly different direction here.
I think that's great.
Don't bet the entire shop on it.
Yeah.
And to have your optimisation activities and your testing activities really well structured and broken out so that, yes, you've got your core activity that you're going to continue to optimise for that first hill.
(49:36):
But actually set aside a meaningful level of budget so that you can test new channels, creative optimisation.
I was going to say bid optimisation, but as we discussed with Mike there, Google's setting the cost of those bids anyway.
But then some feed optimisation or Merchant Center optimisation, because obviously we've also discussed issues with the feed today.
(50:02):
Yeah.
And that structured approach to testing is really important, but just don't take huge risks and bet the entire shop.
Yeah, of course, you know, I suppose that it's...
You're back to that whole thing about optimising for scale and then optimising for efficiency, and how the two balance off against each other and...
Like, you can be ruthlessly efficient with no growth.
(50:24):
Yeah.
And maybe that's a valid objective for some people, but I can't see how being ruthlessly efficient and missing opportunities is what would fire up the excitement of ecommerce marketers listening.
(50:45):
It's never going to be.
And like, what I've seen recently with retailers, obviously this year, it's a year of lean.
It's a year of, let's sharpen the pencils, let's look at the P&Ls, let's start figuring out where our cost centres are, let's start cutting them out.
But Mike spoke about this idea of not looking at it as a cost centre, but trying to change the perspective and looking at it as a value centre.
(51:11):
And so that means that you kind of have to look at it with slightly different metrics.
You don't look at it as cost-per-click, but a value-per-click or a revenue-per-click and that sort of thing.
And I think that shift in mindset is very important, but also very challenging.
It's very difficult.
And we have definitely seen situations where we're looking at, let's say you've got a budget to spend and you've got a growth target and you will often see a celebration of reducing the budget.
(51:40):
Well, it's an expectation to increase growth?
Yeah, the expectation is to increase growth, but, like, the obsession or the focus is nearly on, hey, I reduced my budget.
But then you're looking at your growth and saying, well, I didn't quite achieve the growth I was looking for.
So, you know, I reduced my budget by 10% or I spent less by, let's say, 10% and I got like, you know, 10, 5%, 12% growth.
(52:02):
And then forgetting the fact that, well, if you spent your full budget, maybe you'd have gotten your 30% growth.
Well, the efficiency question then comes back.
So how efficient are you being with your spend?
And have you got the majority of your spend at an efficiency level that you are comfortable with so that you can then start spending on top and testing?
(52:24):
The problem with that for many CFOs is the size of the overall advertising budget and the hole that that tends to make in the overall P&L.
Yeah.
And particularly if you're at a point where your efficiency level isn't where you would want it to generally be.
(52:48):
So if you're at, in some cases, 30% plus of total marketing or 30% of your total revenue is being invested in advertising, that's going to make anybody really, really uncomfortable.
And you're going to be under pressure to reduce that.
Probably rightly so.
(53:09):
But in order to do that, you're then on this, well, efficiency, discovery, optimisation testing.
And that's why you need to have structure around it.
And you can see as well, I suppose from Mike's position, as he mentioned himself, like the thing is moving so fast.
It is, the pace of change is incredible.
And as we're getting used to things like, hey, I can manage my feed, the feed is being pulled from under you.
(53:34):
That horizon is just always leaning forward.
Yeah, totally.
And you've got to chase it.
So what my learning is, we've got to get to the top of the mountain and then we lean forward over the horizon.
Something along those lines.
Ah perfect, I have it.
I think that's poetic in your usual style.
Thank you very much.
Thank you, Gordon.
(53:55):
Thanks everyone for listening.
It's been a pleasure.
See you soon.
You've been listening to Functional & Fabulous with Ger Keohane and Gordon Newman.
If you'd like to know more about the podcast, or about StudioForty9 and Omnichannel Stories, please go to functionalandfabulous.ie. Our sound engineer was Elaine Smith and the show was produced by Roger Overall.
This is Functional & Fabulous, the omnichannel podcast where we unbox tales of online retail and digital transformation.
In this episode, Ger is on an epic quest...
I love a quest, I'm on an epic, I'm on an epic quest.
And he's assembled a crack team using a rigorous selection process.
(00:22):
There's Gordon...
Are you good at being responsible?
I'm terrible at being responsible.
And Mike...
The data-driven pushback hipster.
Join them in their bid to find the Holy Grail of Google-driven ecommerce.
How do we enjoy PMax responsibly?
Along the way, there will be traps...
(00:42):
The monkey who sticks the paw into a canister, grabs some nuts and then finds that he can't get his paw out.
There'll also be riddles...
Horizon is just always leaning forward.
And violence...
Get your head cut off waiting for that to come home.
So buckle up for the ecommerce podcast adventure of a lifetime.
(01:04):
It will be very interesting, I think.
This episode of Functional & Fabulous is brought to you with pride by StudioForty9, retail ecommerce experts, omnichannel growth consultants and cut-through performance marketing specialists.
StudioForty9, where your digital retail success is built.
(01:25):
So welcome to another episode of Functional & Fabulous, where we delve into the ever-evolving world of ecommerce and retail.
Today, we have another extremely interesting and experienced guest, Mike Ryan, who is Head of Ecommerce Insights at Smarter Ecommerce, based in Austria.
Mike's dad grew up in Southie, in Boston, and Mike tells us he's Irish at heart.
(01:46):
And with a name like Mike Ryan, he doesn't really have the choice.
But he's with us today because he has an exceptionally deep understanding of ecommerce and digital marketing, which he leverages to provide invaluable insights and create innovative solutions for his clients.
He's a generous sharer of information, so it's well worth following him on LinkedIn, where he's a regular poster.
And also checking out his own podcast, and I'm doing it again, promoting other podcasts, but this one is called Growing Ecommerce.
(02:13):
Very, very interesting, some great leading industry experts.
But on that podcast, Mike explores e-com, digital marketing and data science.
Mike's expertise has been recognised by leading publications such as Adweek, Financial Times and Bloomberg, and he has been a prominent speaker at industry conferences.
We're delighted to have him here with us today.
(02:33):
I'm delighted to make my intros longer than ever, and we're dying to learn more about how to enjoy performance marketing responsibly.
Mike, it's great to have you.
Yeah, thanks so much for having me on.
Thanks for the very warm introduction and also promoting my podcast too.
Don't worry, I don't think it's a zero-sum game.
I think most podcast listeners are podcast junkies, and they'll listen to more than one.
(02:57):
It definitely is, and I got a slap on the wrist not so long back from our producer for promoting other people's podcasts.
Surely not, he said, but anyway.
Yeah, it wasn't a slap on the wrist from me, I'll hasten to add.
Welcome to the podcast, Mike.
I'm very excited to talk to you today about enjoying PMax responsibly, amongst other things.
(03:19):
Amongst other things.
Also, the fact that you're a data-driven pushback hipster.
You're the only one I know, I think.
You were pushing back on data-driven decisions since before it was cool.
That should be my LinkedIn headline kind of thing, you know?
No, I mean, I don't want to say...
Yeah, I just push back a bit on...
(03:39):
We shouldn't be kind of beholden to data.
It should be a tool in the toolbox, but we shouldn't forget other tools in the toolbox.
Absolutely, and as we were talking about, I really enjoyed a talk I heard recently.
Aimee Connolly of Sculpted by Aimee spoke at a conference, and she talked about data gut, which I thought was a really nice kind of concept where, you know, of course, we're looking at data, but we're also checking our gut, we're checking our instinct, you know, checking common sense.
(04:09):
Is the data telling us the right sorts of things and so on?
But one of the things that, you know, we covered quite a bit when we did the pre-recording chat was PMax.
It is kind of one of your areas of expertise, and I love that whole thing about getting to use it responsibly.
Can you tell us where that comes from?
(04:29):
What is a non-responsible use of PMax?
Straight into the good stuff.
Oh, yeah.
Let's go straight in.
Let's start on PMax.
Okay, I just came back from lunch, but not my...
Good, you got my blood running again.
Good.
Super.
We won't do the stand up and wave your hands in the air thing.
We're not going to make you go through this.
(04:50):
We have to rewind a little bit because I've been working with ecommerce ads and Google Ads for a long time, and that in the past was shopping campaigns, standard shopping campaigns.
And now shopping is just kind of one inventory.
Google likes to rather view it as inventory than as a campaign type, although the classic campaign type is still there.
(05:11):
They built this more expansive kind of campaign type called Smart Shopping, and that was already a while back now.
They kind of evolved even further into this thing called Performance Max.
And just in case you're not too familiar with this technology, the basic premise of Performance Max is that it's a buy-in service to all of Google Ads inventory.
(05:33):
So not only the shopping inventory, but also search ads, classic text ads, display, Gmail, YouTube, everything.
It's all kind of possible from there.
And in fact, it's, if you just run it out of the box the way, kind of just turn it on the way Google intends, then you don't really have any control either on that.
(05:57):
Google sort of promises that they're going to kind of handle matching, like finding the audience, matching your products, and just building ads, really a very highly automated campaign type.
So when Smart Shopping came out, there was some pushback from the market and certainly from me too.
There were many things I didn't like about the way that product was designed.
(06:20):
It took away a lot of promotional controls and levers.
It took, like things, like involving bid adjustments and different kinds of adjustments that you could make, scheduling that you could do.
And it also took away a lot of reporting that people were used to, like with a classic search campaign, you are targeting keywords.
(06:46):
So someone out there is searching for something, like, they'll make some kind of commercially related query, like, let's say buying Nike shoes just to make something done.
And imagine that Nike is a keyword that you have, just to use a dirt-simple example.
And shopping now that you're not really targeting keywords, but you are, Google is already matching your products to queries.
(07:10):
And you in the past could see those queries, what they were in search term, of course.
So that's just an example.
People found that really valuable because sometimes Google would be matching you with some strange search terms.
And so that was another type of control you had.
You had the transparency to see what you were matching for.
And you had the control to negative those things, like to say, no, I don't want my products to match for a given search term.
(07:31):
So Google just stripped away a lot of stuff.
Some would say dumbed down the campaign, I pushed back against it a lot.
But you know what I found was that people adopted the technology anyway, with or without me.
They heard what I had to say, but people were using this stuff anyhow. And so that's what I mean by...
How disappointing...
Well, I think Google do an incredible job with their sales teams and with their account managers.
(07:56):
If you ever go to a Google conference, every single one of them will be on message for that conference.
So it's just drilled into you.
For sure.
And I mean, something else that we observed back in the days with Smart Shopping is that often, like, the operational teams, like, the practitioners, also had their reservations or concerns about Smart Shopping.
(08:17):
And you'd see, at least in, it depends on the account size and stuff like that.
But Google's team would go in a top-down approach because they'd reach someone higher up the food chain, let's just say a CMO, for example, who's not into those practitioner level details.
And they could kind of reach that person and then make the change happen from a top-down way.
(08:40):
But the change did occur.
The adoption was pretty massive after a couple of years.
And with PMax, I made the decision, okay, I know, because I've seen this story before.
People are going to use this technology.
In fact, Smart Shopping was migrated over to PMax automatically.
So by default, people would be using this technology.
(09:01):
And I want to help people enjoy it responsibly.
And that means several things.
But I think what's really important is that PMax campaigns are not necessarily going to be incremental in the terms of, like, audience incrementality.
You can think about that.
Like, it tends to reach into remarketing, brand traffic, basically warm or existing traffic.
(09:27):
Which can have lower incrementality and kind of that classic marketing incrementality term.
But also in terms of, like, business incrementality, not every product has equal importance to your business.
Some things are going to have their own natural demand that it's easier for Google to follow.
So PMax will tend to sell into bestsellers.
(09:48):
But you might have a product that doesn't perform as well in platform.
Google's a bit less interested in promoting that because it's going to be harder for them to extract, like, platform value for you.
But it might be higher margin for you, or it might be overstock, and it's behind sell-through rate and you need to clear it.
So these are the kinds of things you have to tell Google what's important about your business.
(10:12):
And you also have to kind of work a bit to get the campaigns to work harder for you.
That's what I mean by enjoying responsibly.
So that was a pretty long answer.
Thanks for that.
No, I think it's okay.
There was one just to summarise a little bit of that and some of the reasons that we're talking about the kind of responsible piece of it.
You know, you use the term, it's like the monkey's paw.
(10:33):
The monkey who sticks the paw into a canister, grabs some nuts and then finds that he can't get his paw out.
And obviously the trick is the monkey has to leave go of the nuts in order to get the paw back out.
And what's happening is the PMax approach, if you just turn it on, it's going to follow the path of least resistance and it's going to try and make whatever sales it can.
(10:55):
And those sales may or may not be good for you.
You may or may not be about to sell anyway those sorts of products, and therefore you need to kind of shape the story for Google to understand what it is that you're trying to sell.
So, like, let's say we agree with all of that.
How do we go about doing it?
I was going to say that will be music to marketers ears of, yeah, okay, we all accept PMax is problematic if we don't do that properly.
(11:24):
So the obvious next question is, let's enjoy it responsibly.
How do we enjoy PMax responsibly?
Yeah, I mean, I think it's definitely a case of be careful what you wish for.
People wanted performance that is basically efficient and scales.
(11:45):
And that's hard to actually do because really incremental marketing is typically not going to look as good in platform.
So what you need to do is just help bring in, there are ways to bring in control and there are ways to tell Google what's important to you.
Let's just look at the topic of telling Google what matters to you because, for example, if we're talking about that thing about a sell-through rate, Google doesn't know that information, that's off-channel information.
(12:14):
They can't know that.
They know your stock level or they basically know your stock status of yes or no.
Or similarly, let's imagine that you've got plenty of availability on product in your core sizing, and another product is kind of thin on the core sizing, and there's just these less common sizes like an extra small and an extra large, or something like that in there.
(12:39):
This is again, this is something that could be important to your business.
It could also impact the campaign and Google doesn't know about that.
That's not their fault.
So what you need to do is bring that data into your campaign.
We also talked about margin, for example.
The challenge that I see, people tend to pick a battle here.
(13:05):
They'll find one or the other thing and try to bring that data in and try to optimise toward that.
It's very simple to just bring the data in that, Google has these things called custom labels in the feed and you can relatively easy bring in a piece of data, but then actually making that work for you is a bit more complicated.
(13:27):
So, like, in the past, these shopping campaigns worked at it.
You could really optimise down to a pretty granular level, but in Performance Max, you really need to have a more consolidated account structure.
You want to have these kind of larger pools of data and campaigns.
You don't want to split it up too much.
(13:48):
And what I like to do, I basically think of these campaigns, typically someone's going to make a product-type campaign or a brand campaign.
Like, they'll have a campaign for ski jackets, for example, or for a given ski jacket brand, let's say.
But what I think is more instructive is to take these campaigns and view them as just like an empty vessel or a vehicle.
(14:15):
And kind of be agnostic to what's in there because the thing about the campaign is that you can set a budget and you can set an efficiency target.
And those are the two main things that you can do to pace yourself on Google Ads.
The budget needs to be there to spend and the efficiency target will determine the aggression level.
(14:35):
So what happens then, if you make some of these buckets with different budget levels, different efficiency targets, you have these different kind of, like, pods or vehicles that are gonna push products more or less aggressively.
And then it's a matter of making sure that products are in that right campaign, that they're getting the right promotional intensity, basically.
(15:00):
So what we like to do is basically create a product score that is going to look at a product on any number of dimensions, however many are important to you.
And products could end up in different of these buckets, these campaigns for different reasons.
For example, one product, it might be important to you to promote because it's behind sales plan.
(15:25):
And another product might be important to promote because the margin is high.
Similarly, you could have kind of a bestseller, something that's flying off the shelf.
And you might actually want to dampen that because Google is going to see that it's performed well and it's going to keep juicing it.
But you might actually want to dampen that because actually you have plenty of organic volume on that and you feel like you're cannibalising that.
(15:48):
Or it's selling fast and well, but it's actually not your highest-margin product.
There could be any number of reasons why.
You know your backorder status or your on-order status.
It's just about thinking about these things and making sure that Google is pacing the right products.
So that's one way.
Yeah, it's quite interesting.
So I suppose just to kind of summarise a little bit, you know, you could obviously send through the exact information about the product margins and all the rest of it, but then you've got the big problem.
(16:17):
So you can, let's say, enrich the data quite substantially, but then you've got the problem of how do you explain to Google what that data means?
So what you're saying is, rather than sending through, let's say the margin on this one is 19.5, the margin on this is 37.5 margin, whatever.
And then having the big problem of trying to explain to Google which ones you're trying to push, you basically would say maybe a bucket for high-margin, low-margin, medium, and then you would indicate to Google, okay, well, it's the high-margin ones I want to push first and foremost.
(16:46):
And it's kind of that taking a guardrails approach, I guess, and ensuring rather than giving the exact instruction saying, okay, well, here's the guardrails when you're driving the car.
Don't go off the side of the guardrails because that's where the cliff is.
And that's the general gist of it, right?
Exactly.
I mean, and Google will never understand if you, in this example of a margin, the margin topic, you can have a campaign for each margin class and Google, it's not like they can read the campaign name and know why that's important or which one.
(17:18):
You have to just speak in its language, which is the budget and the efficiency target because Google's AI is not strategic.
It is not aware of any of this stuff.
It's just looking at it on a per auction basis.
It's got a month's budget in front of it or a day's budget.
And it's going to basically look at the expected click-through rates and the expected conversion rate.
(17:46):
It's very smart in its own way.
I don't mean to dismiss the technology.
It's very effective at that job, but it doesn't have a wider view than that.
That's our job, still.
Yeah, I always think that sometimes we lose sight of the fact that Google's job for itself is to juice enough revenue for Google, or juice the maximum revenue or maximum monetisation of each potential click.
(18:16):
So ultimately, it's still an auction.
Google's still going to be placing results in front of consumers or potential consumers that they are most likely to click on and that's going to earn them the most money.
So we have to have that in the back of our heads when we're planning our campaigns.
Totally.
(18:36):
It's a marketplace of supply-and-demand and it needs to look after the match rates between that.
And for both, because everyone wants a high match rate.
The advertisers want their products to be seen and clicked and potentially purchased.
The consumers want to see relevant products that they might want to buy.
(18:59):
And of course, yeah, there's the part you're talking about.
Google wants to get their clicks.
That's, they build on the clicks.
So it's absolutely the case.
And the thing about this auction is that Google is...
When we're using the smart bidding technology, which almost everyone is using, Google is like, they have the auction environment.
(19:20):
They're the auctioneer and they're the bidder.
It's quite fascinating.
Like they're bidding, you're giving them permission to bid on your behalf, but they're also the auctioneer.
There's not a lot of auctions that work that way where the person collecting the auction is also the person doing the bidding.
It's quite fascinating.
And moving a little bit along the same lines, but moving the story a little bit further along, feed management and exclusions and so on in feed management and that sort of thing.
(19:51):
Where do you see that going?
Well, let's see, which part?
The feed management part?
Well, hold on.
It's funny.
I was just thinking about this and writing about this yesterday.
Something that comes to my mind is that Google is working to try to...
That's another area that they want to automate more.
(20:12):
For someone, in case someone is less familiar, the product feed is just basically a place where you provide structured data so that Google understands your catalog and what you're selling.
So it's explicitly structured data with attributes so they know exactly what they're dealing with.
And a goal of Google for years and years now has been for them to be able to implicitly know.
(20:37):
So without receiving that structured data, for them to, for example, get the URL of your webshop and they can crawl it and they can extract everything that they need to know.
So basically, you could hand them more or less a credit card and a URL and say, let's go.
This is the direction that they're going a little bit further now.
(20:58):
They have a new version of their Merchant Center, where you upload your feeds and your product data, usually.
It's called Merchant Center Next.
And they're doing away with the term feed, interestingly.
That's how they've been gradually downplaying the word keywords for years and years.
Feeds have been absolutely at the core of what the Merchant Center is and how Google shopping is feed-based advertising.
(21:25):
So what are they, where are we going without the feeds?
What are they doing?
They're talking now about data sources.
And the reason why, this allows them a little bit more flexibility.
It's a bit of a broader definition because now they view that structure data as basically data that's provided by you.
(21:45):
But they have another category of data that they call Found By Google.
And this is data that they get from crawling your website.
And so in the end, this has taken a step forward due to large language models and advances in visual AI.
They're able to not only crawl your website in a classic way, and by the way, you'll also have structured product data available on your website.
(22:12):
But they'll also be able to use language models and visual AI to also just understand your website.
And so the feasibility of this idea of kind of hand your credit card and URL has taken a massive step forward.
But it also, it's another thing right now, you can exercise a certain amount of control through your feed, you can bring in data, you can exclude data and so on.
(22:41):
And you will have perhaps a bit less control over that in the future.
It's not quite clear the direction where that's headed.
Some people won't like this, but it's always the case.
If you were a betting man, would you bet the deprecation of that control over time?
I think you'd have to bet that's where we're aiming for.
(23:02):
That's what's gonna happen.
Yeah, it's certainly the direction of travel.
And people have been talking about the death of the keyword since, to the best of my knowledge, since 2014.
So we're now at a decade of the keyword dying.
And I think it's probably something like that with feeds.
I think that it could take quite a while, but it's gonna be a matter of prioritisation, the level of support that's available, how Google will structure features around, they'll gradually steer us a certain way.
(23:32):
But I can't say that it's going to be tomorrow or next year.
Because it feels like, certainly here, that the whole concept of managing the feeds and crafting and sculpting them so that you'll get an outcome that you're looking for, is kind of immature here at the moment.
I don't know, Gordon, what you think?
(23:53):
In a previous life, we spent a long time curating our feeds, trying to manage newness, seasonal product, trying to exclude products that had, like, high likelihood of sell-through in season, all those kinds of things.
That was a big focus, I think.
But more broadly from talking to a lot of retailers, getting the feed in or getting Merchant Center, having all the products in there is probably the step that they're at still.
(24:26):
That's what I mean, yeah.
That's still kind of, like, we're getting the feed in and that idea of controlling the feed is probably a little bit new.
So I think it would be very disappointing for people to get to a stage where they now know how to control the feeds and then they have that carpet whipped from under them.
I think that's constant innovation, though, isn't it?
(24:47):
Innovation in inverted commas to keep all marketers on their toes.
Yeah, I mean, the pace of change is, it's exhausting, to be honest.
Definitely put a few grey hairs on my head.
Not enough yet.
I've got a lot more than you there, Mike.
So we won't talk about that.
(25:07):
So just to move along a little bit, we talked a little bit about KPIs and the importance of KPIs.
Also, we talked about attribution, everybody.
Attribution is such a big subject at the moment and we might, I don't know if you want to touch on it or not, but certainly KPIs, I think you had an interesting comment around ROAS and I suppose, again, we're kind of in a stage where people use ROAS nearly as a, like probably too religiously as a KPI maybe, whereas it's still an important KPI.
(25:35):
But tell us what you think about ROAS as a KPI.
I think there's a saying out there, I've heard this before, you are not your ROAS.
I certainly believe that.
I think that ROAS is a very useful KPI, first off, and I don't want to just, like, knock it or something.
I can understand why it's become so popular.
(25:56):
It's available at so many levels in an account, and all the way down to the individual items, and it's so versatile.
It's simple to calculate.
It appears to be simple to understand, but a couple of concerns I have about that, I think that people often kind of conflate it or substitute it with ROI or profit, and I think Google very much welcomes this.
(26:21):
You'll often see in their sales and marketing collateral that they'll talk about ROI when actually they mean ROAS, and these two things are certainly not the same.
There's a risk there.
It's when you've got a proxy metric and you make that mental substitution, that can be quite dangerous because proxies are not perfect, they're inaccurate, and particularly ROAS is available.
(26:47):
You can view it in platform.
You could view it in an analytics tool as well, but that platform only knows what the platform knows and this starts to rapidly move into themes of attribution and so on, but I think it's really important to remember that this is one metric and you better triangulate it with more.
(27:09):
I think you don't want to just lean on that one metric, or let it be kind of your be-all end-all because it is only a proxy, and I actually think if you can, it's funny because it's the primary lever that's available to you in your campaigns anymore is to adjust your ROAS.
That's how you're going to pace your campaigns in large part, but it's also supposed to be a business objective.
(27:36):
If it's really a business objective, then can it be a tactical lever?
You shouldn't just play with your strategic goals, and I actually would rather kind of decouple that and say, yeah, it is actually more of a tactical lever than a business objective.
You need to be actually measuring and setting your strategy on other numbers from other systems, or at least looking at those as well.
(27:58):
What other KPI, like if, you know, obviously, so ROAS for me is, it's a directional, you know, it's a trend kind of a, you know, we're trending upwards, we're going down, you know, it's very, very useful.
It's great to, it's a great measure to use with the, let's say, the C-suite, that kind of a thing.
But what other KPIs do you think are more important or which other ones would you use, as you say, to triangulate the kind of the real story?
(28:25):
I mean, speaking of, like, direction of travel, I think MER is quite popular in that use case these days...
That's my favourite...
Marketing efficiency ratio and it's basically looking at your total sales and your total advertising costs, and it just gives you a much higher-level picture of what's going on.
(28:46):
And it partly solves for things like the fact that you could have Meta Ads and Google Ads both claiming full credit for the same purchase, or stuff like that.
And so this helps to, in a way, to deduplicate that.
So that's certainly a popular one.
I, my only problem with that, I'm always willing to change my mind and stuff.
(29:08):
I was a little anti-MER at the start and it's, I just saw people using it for things I don't think it's appropriate for, like, to make really detailed channel tactical changes or saying like, oh, I made changes to this ad set in Meta.
And then I saw this change in my MER.
(29:30):
And I mean, yeah, if like 90% of your spend is Meta, that might be credible.
But if you've got to, you could just read whatever you want into that.
And then it's too easy for our own confirmation bias to step in and so on.
So we have to use MER responsibly.
Yes, let's use everything responsibly.
We gotta be responsible?
It's, I mean, we're not good at being responsible.
(29:51):
Are you good at being responsible?
I'm terrible at being responsible.
MER for me is a good strategic.
Use MER like a maniac...
Don't use MER like a maniac.
It's a good strategic measurement to track your direction of travel over time.
But I would agree with you there, Mike.
It's not a, it's not a tactical measurement, and any tactical decisions made off the back of MER performance, you should be questioning.
(30:19):
But what about, are there any more detailed metrics, like, I suppose, further down the funnel, more tactical metrics that people should be focusing on?
I mean, well, while we're talking about it, I would say that profits, measurement, profit optimisation are becoming a lot more popular than they were a couple of years back where people were typically very focused on growing revenue, maybe at all costs.
(30:47):
This can, of course, have downsides.
Also, a lot of brands were, you know, I think of Shopify brands, for example, where there were these cool little Shopify tools that helped you understand LTV a bit and it seemed all good, but you can really stick your neck out on LTV and you can really get your head cut off waiting for that to come home.
(31:11):
So now, it's a certainly more conservative financial environment, and we see people wanting to optimise towards near-term profit that's much more fashionable these days, which I think is healthy, and also down to further contribution margin levels that consider, you know, whatever the case might be, different sales tax, different shipping fees, and so on and so forth.
(31:39):
So that's another one that I definitely shout out.
But also, I'm gonna call out one tactical thing that I think is, like, for years and years, people talked a lot about CPC, cost per click, and these are pay-per-click channels, Google Shopping, Meta.
These are, like, depending on the format, these are pay-per-click channels and you get why people are focused on that.
(32:06):
But what I like to talk about more these days is value-per-click or revenue-per-click, typically.
Depends what value you're tracking.
It could also be profit-per-click.
And the reason why is, like, I mean, honestly, Google owns the CPC at this point.
Greater than 80% of the market is using automated bidding.
(32:27):
And so even if you're not, when everyone else is, it doesn't matter because the bid heights are somehow dictated by Google at this point.
And it's become an observation metric.
In the past, when people were much more active in their bidding, this was more important.
But you also have to consider, like, with a campaign type like Performance Max, when you make a change to your ROAS target, it won't only impact your cost-per-click anymore.
(32:57):
It might also impact your channel mix.
They might realise if you set the ROAS a bit lower, Google might realise, oh, I can start expanding into more video at that efficiency level or whatever the case might be.
So it's just more complicated.
And in the end, you know, if you're obsessing over cost-per-click, you are just talking about the channel as a cost centre at that point.
(33:19):
If you're talking about the value-per-click, there's much more that's contained in that.
Your order values are contained in that.
Your conversion rate is contained in that.
Those are the things that produce that.
And those are, like, you can really optimise stuff and see your value-per-click improving.
And it's then focusing on the channel as a value centre.
(33:39):
So I think that's a mindset shift that some people would be well-advised to consider.
It's a challenge.
One of the things I like to look at is that kind of offset between, what's it costing us to acquire a click and then offsetting that against the value-per-click at both a revenue and a profit basis.
(34:00):
So I know that if every click I'm getting onto the website's earning me $1, but it's costing me $1.10 to get every click onto the site, then I have a problem.
Yeah, I mean, and that is, if you hold the value-per-click and the cost-per-click against each other, that's basically return on ad spend at a per-click level.
(34:21):
So it totally makes sense.
But in principle, Google should be caring for this.
And you can't do much to influence the cost-per-click because Google's got that.
You can influence the value-per-click by improving your on-page experience, by improving your ad targeting and all kinds of stuff.
(34:42):
Great.
One of the areas, and a complete kind of shift, I haven't even got a good segue for this.
Pivot, we're gonna pivot.
I've not got a good pivot for this one.
I'm putting on my neck brace.
Since we've hit a couple of trigger points here, one of the areas and one of the businesses in particular that you've been quite vocal about.
(35:04):
Don't say it...
You know what's coming.
It is a brand that seems to be pushing up CPCs and CPMs for everybody.
And it helps a lot of people shop like a billionaire.
It's our friends over at Temu.
And we'd be really interested to hear your take on Temu generally.
(35:28):
Where to begin?
It's a fraught topic for me...
This also, Mike, is the wrapping up topic, by the way.
Yeah, look at the time. Oh, what a lovely note to end on.
Thank you, gentlemen.
Thank you for dumping me with Temu.
No problem.
I couldn't resist.
No, by all means.
(35:49):
I mean, I've been watching Temu for quite some time.
And they've been active here in Europe for about a year now, where they very rapidly came online through advertising.
And we've seen that in Meta Ads.
You know, you can check the Meta Ad library and see how many ads they're running daily and track this over time.
(36:12):
And the number is just insane.
It dwarfs all the other competitors.
And it's just gone up and up and up.
And similarly, something that I've looked at is basically the percentage of Google advertisers that face account-level competition from Temu.
And they just turned on a fire hose of advertising dollars in a way that's, I would argue, unprecedented.
(36:35):
Because I can look back in history and report those same numbers for Amazon, for Wish, when they were around.
And both of those companies had huge and massively aggressive market entries.
But Temu just takes the cake.
Now, I mean, one thing that I would say is that it's a bit different.
(36:55):
When Amazon turned on ad dollars in a huge way like that, this destroyed cost per click and the unit economics overnight.
It was a huge problem.
But now that they're still there and that damage in a way is almost done.
So Temu coming in, it does affect unit costs a bit, but it's almost more about a war for visibility.
(37:25):
And I found some interesting things.
So this company started fluctuating their ad spend, for whatever reason.
Like they were always just going up and to the right, spending more, more aggressive, always, always.
Until quarter one of this year when, a couple of times, we saw some pretty big dips.
(37:45):
We just saw volatility in their spend.
And I love this as an analyst because when there's volatility, there's change.
And then I can start looking for the correlation between that change and other changes.
So I found that, yes, they definitely hurt CPM.
That's true.
But interestingly, cost per acquisition improves when Temu is more visible.
(38:12):
And this is totally weird.
My hypothesis for this, I have a twofold kind of thing.
One of them, part one, I call the shit sponge.
I think that when Temu is very active, they're actually kind of soaking up shitty auctions because they'll actually, they might crowd you out of more generic auctions that you didn't want to be in anyway.
(38:36):
And the other thing is, basically, I think it's called a contrast bias.
If you take someone with a really nice polished appearance and then you take someone with an average appearance and someone with like a very poor appearance, the average person is going to look worse compared to the really polished person.
(38:56):
But the average person will look better compared to the kind of sloppy person.
And I think that consumers are not that trustful of Temu, or many consumers are not.
And when they're crowding out these auctions, it can actually kind of have a halo effect for your brand or for your offer.
(39:18):
So it's very fascinating.
That is very interesting because we are looking at it now in the context of Ireland and we have Temu, let's say, growing share into the Irish market, and we have Amazon.ie opening up in Ireland next year.
And I suppose people are nervous what that's going to do to cost per click and cost of acquisition and all of that side of things.
(39:44):
We've seen them come on in Netherlands, for example, in France in recent years.
It's actually a bit further back now, but we've seen them make those market entries.
And yeah, I mean, it's certainly not a good thing when Amazon or Temu or anyone enters your market, from an advertising perspective.
(40:09):
It'll also be interesting to see how the adoption is among domestic Irish sellers.
Because for example, in Netherlands or France, we saw that Amazon was really getting kind of seeded or bootstrapped by Chinese sellers.
And it took a bit longer for local sellers to kind of adopt the channel.
That's quite an interesting point, actually, of what other reactions have you seen?
(40:34):
Like, if we take The Netherlands, what were the other key learnings that retailers would have taken from that market entry?
Yeah, I mean, in The Netherlands, for example, there is a very dominant marketplace, homegrown marketplace called bol.com.
(40:56):
And it just was kind of a war between these two titans.
And I think a lot of companies had probably frenemy relationships with bol.com because they're probably selling there.
They're probably also competing with bol.com to a certain extent.
And yeah, Amazon comes in.
(41:18):
And I don't think, I think people would rather compete against bol.com than Amazon.
Similarly, like in France, who has their seed discount and these big dominant local retailers.
And France is a very, a nationalist market.
(41:40):
And so I think Amazon had a bit of a tougher time there, but yeah, it depends a lot on consumer preference.
Like, I don't know the extent to which Irish consumers are already shopping from Amazon.co.uk or to what extent they, I don't know how committed they are to purchasing locally.
(42:05):
You'd know more about that than me.
It will be very interesting, I think, particularly the Amazon.co.uk. I think Irish shops have been using that in anger for a number of years.
Yeah.
I mean, in my market, I'm in Austria and I'm using Amazon.de in anger.
(42:30):
Yeah.
I think it would be similar to if you were going to get an Amazon.at at some point.
I don't even know if that's on the horizon.
I don't think so.
I don't think so.
But at some point.
Okay.
So unfortunately, we've come up against the time.
It's an interesting, I mean, that whole area, obviously it's an interesting question, how the Irish market is going to react to these two giants coming into the business.
(43:02):
It's going to be a transformative shift, I feel, in Irish ecommerce.
And definitely it will be a challenge.
So we'll be keeping quite a close eye on that, Mike.
Thanks so much for your insight today.
We got a couple...
We didn't get to the data-driven decision, hipster pushback piece.
(43:25):
But I'm looking forward to seeing that on your LinkedIn profile soon.
And thanks so much for joining us.
Yeah.
Thanks very much for having me.
Thanks for sharing with us.
We'll be enjoying our PMax responsibly.
Cheers, Mike.
Well, let's enjoy PMax responsibly and all other media mechanisms.
(43:47):
We never got onto Advantage Plus, actually, which I'm a little bit disappointed about.
We ran out of time.
I don't know about you, Gordon, but I find this time management business, I mean, maybe we need to just focus on how people need to listen to long-form podcasts and not try and make it, you know, nicely bite-sized for the commute.
(44:07):
We're trying to get it commute-sized, but people have got interesting stuff to say.
Yeah, but maybe people have longer commutes now, post-Covid, you know, it's getting, it's taking longer to get to work.
Or people have actual commutes now.
Maybe they have actual commutes, exactly.
I think this is such an interesting area because he has this crossover between the data science and obviously then the commercial aspect, you know.
(44:31):
And it's always interesting to hear a little bit about the data science.
There are certain areas that I would have loved to get into in a bit more detail, but you just don't cover everything.
I think we talked a bit about the, you know, the path of least resistance that PMax provides us with.
And obviously, like as he mentioned, that challenge of scaling efficiency because they're kind of, they're competing.
(44:54):
They're competing things.
You're looking for, I want more, I want more, but I want it kind of cheaper and cheaper, more efficient.
It's almost like you can't say, I want to go and search for new customers, identify them, attract them and then interest them.
But I'd like to do that at the same efficiency level as brand search, please, is a very unrealistic ask.
(45:19):
Yeah.
So you have to manage that.
Yeah.
And it's very difficult to understand how to do that with PMax.
So Mike mentioned confirmation bias here, and I was delighted to talk to him because it feeds into my confirmation bias that PMax is too much of a black box.
Yeah.
And to be able to listen to someone who has actual strategies of how you manage PMax in order to balance those competing objectives, because of course you want your marketing to be as efficient as possible.
(45:53):
Yeah.
Of course you want to service those customers that are already partway through the buying funnel or about to enter the buying funnel.
And then you have to go and look for that growth.
And it's a tool, it's a fabulously powerful tool to be able to do that if you can drive it.
(46:14):
But I think just to use it as a, almost plug and go...
Yeah.
...isn't going to get you the results that you need.
And to get some practical advice on that, I hope that people listening will be able to really take something from that because it must be confusing people.
I think it is.
I think what we typically see is you turn it on and then kind of walk away whistling.
(46:37):
Panicking.
Definitely the idea of setting up those buckets or guardrails, whatever you want to call it.
But, and using that information in the feed to inform Google or to, I suppose not inform, but like influence, I guess really, or direct what actually happens.
You know, it's a really important thing.
The other area that was interesting, and I guess within that is, and we didn't get to speak about it, but it was maybe the importance of testing.
(47:04):
And in particular, because what I noticed when we have situations where we're trying to optimise all the time, you get to a point where you reach the maximum optimisation you can do in a particular area.
So I'm going to explain this.
It's kind of like a mountain range, and you've got a series of hills and then you've got a series of mountains, and then you've got, let's say, Everest and a peak and whatever.
(47:26):
Do you love a quest?
I love a quest.
I'm on an epic quest.
So I'm on an epic quest to find the ultimate summit and I'm going up a hill and my instruction is optimise my, you know, how high I travel.
So I'm up and going up the hill.
I'm going up the hill.
I get to the top of the hill.
That's the optimum that I can do on that particular hill.
(47:48):
So this is my local maximum.
My instruction doesn't tell me, alright Ger, now go back down the hill because maybe that's not the highest hill you can actually get on to, or the highest peak.
And that's where we run into these problems because if our instruction is always do the best and climb the highest, we're not going to go to make the effort or take the risk of going back down the hill to see if there's a higher peak elsewhere.
(48:10):
And in order, what you have to do in order to do that is, you've got to look at, let's say, testing and testing budget and risk budget where you say, righteo, I'm going to jump off of this hill.
I'm going to drop myself some other part of the valley and then I'm going to start maximising and optimising, and then I'll see if I reach a higher hill.
So that idea of having a budget to play around with, to take risks, risks that are not going to be looked upon favourably by the CFO or some other people on the team, it's also still an important thing because you don't want to optimise yourself into, let's say, Carrigtwohill, or sorry, Carrauntoohil.
(48:47):
You don't want to optimise yourself onto the top of Carrauntoohil.
You want to optimise yourself onto the top of the Everest, the Alps or somewhere else.
What do you think, Roger?
Do you agree with me?
Roger agrees with me.
Look, he's totally...
He wants to optimise himself onto Carrigtwohill.
Ha, ha, ha, ha...
Whereas you want to optimise yourself onto Everest...
Oh, absolutely.
Along with everybody else.
There's shocking pictures from Everest, by the way...
(49:09):
Oh, the queues...
But we're not even going to go there.
Okay, let's go somewhere else.
Okay, we'll go in a slightly different direction here.
I think that's great.
Don't bet the entire shop on it.
Yeah.
And to have your optimisation activities and your testing activities really well structured and broken out so that, yes, you've got your core activity that you're going to continue to optimise for that first hill.
(49:36):
But actually set aside a meaningful level of budget so that you can test new channels, creative optimisation.
I was going to say bid optimisation, but as we discussed with Mike there, Google's setting the cost of those bids anyway.
But then some feed optimisation or Merchant Center optimisation, because obviously we've also discussed issues with the feed today.
(50:02):
Yeah.
And that structured approach to testing is really important, but just don't take huge risks and bet the entire shop.
Yeah, of course, you know, I suppose that it's...
You're back to that whole thing about optimising for scale and then optimising for efficiency, and how the two balance off against each other and...
Like, you can be ruthlessly efficient with no growth.
(50:24):
Yeah.
And maybe that's a valid objective for some people, but I can't see how being ruthlessly efficient and missing opportunities is what would fire up the excitement of ecommerce marketers listening.
(50:45):
It's never going to be.
And like, what I've seen recently with retailers, obviously this year, it's a year of lean.
It's a year of, let's sharpen the pencils, let's look at the P&Ls, let's start figuring out where our cost centres are, let's start cutting them out.
But Mike spoke about this idea of not looking at it as a cost centre, but trying to change the perspective and looking at it as a value centre.
(51:11):
And so that means that you kind of have to look at it with slightly different metrics.
You don't look at it as cost-per-click, but a value-per-click or a revenue-per-click and that sort of thing.
And I think that shift in mindset is very important, but also very challenging.
It's very difficult.
And we have definitely seen situations where we're looking at, let's say you've got a budget to spend and you've got a growth target and you will often see a celebration of reducing the budget.
(51:40):
Well, it's an expectation to increase growth?
Yeah, the expectation is to increase growth, but, like, the obsession or the focus is nearly on, hey, I reduced my budget.
But then you're looking at your growth and saying, well, I didn't quite achieve the growth I was looking for.
So, you know, I reduced my budget by 10% or I spent less by, let's say, 10% and I got like, you know, 10, 5%, 12% growth.
(52:02):
And then forgetting the fact that, well, if you spent your full budget, maybe you'd have gotten your 30% growth.
Well, the efficiency question then comes back.
So how efficient are you being with your spend?
And have you got the majority of your spend at an efficiency level that you are comfortable with so that you can then start spending on top and testing?
(52:24):
The problem with that for many CFOs is the size of the overall advertising budget and the hole that that tends to make in the overall P&L.
Yeah.
And particularly if you're at a point where your efficiency level isn't where you would want it to generally be.
(52:48):
So if you're at, in some cases, 30% plus of total marketing or 30% of your total revenue is being invested in advertising, that's going to make anybody really, really uncomfortable.
And you're going to be under pressure to reduce that.
Probably rightly so.
(53:09):
But in order to do that, you're then on this, well, efficiency, discovery, optimisation testing.
And that's why you need to have structure around it.
And you can see as well, I suppose from Mike's position, as he mentioned himself, like the thing is moving so fast.
It is, the pace of change is incredible.
And as we're getting used to things like, hey, I can manage my feed, the feed is being pulled from under you.
(53:34):
That horizon is just always leaning forward.
Yeah, totally.
And you've got to chase it.
So what my learning is, we've got to get to the top of the mountain and then we lean forward over the horizon.
Something along those lines.
Ah perfect, I have it.
I think that's poetic in your usual style.
Thank you very much.
Thank you, Gordon.
(53:55):
Thanks everyone for listening.
It's been a pleasure.
See you soon.
You've been listening to Functional & Fabulous with Ger Keohane and Gordon Newman.
If you'd like to know more about the podcast, or about StudioForty9 and Omnichannel Stories, please go to functionalandfabulous.ie. Our sound engineer was Elaine Smith and the show was produced by Roger Overall.
Popular Podcasts
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com
24/7 News: The Latest
The latest news in 4 minutes updated every hour, every day.
Therapy Gecko
An unlicensed lizard psychologist travels the universe talking to strangers about absolutely nothing. TO CALL THE GECKO: follow me on https://www.twitch.tv/lyleforever to get a notification for when I am taking calls. I am usually live Mondays, Wednesdays, and Fridays but lately a lot of other times too. I am a gecko.