Platform = Brief: Creative in the Age of AI and Algorithms

Katie Flynn: Hello. Welcome back to the podcast. Our chief media officer whose role is literally always changing, especially in the world of AI. So that’s what we’re going to focus on today. Yep. Came to my office the other day. We’ve been talking a ton as an agency about a few new, that have happened in the platforms that we work on, um, that we’re paying very close attention to and making sure that we know how to implement as an agency. So kind of got three that we’re going to roll through today. The first being Google’s performance max and AI max campaign types for search. The second metas Andromeda. It’s an AI powered delivery system for large volumes of content basically for personally personalized ads. And then open AI just last week, they announced that they have a new web browser on the scene Atlas. So we’ll chat about that a little bit too, but let’s start with Google. We do a lot of work in Google and I know we’ve been testing out performance max and sort of trying to get a handle on this very new AI max. So what do we, what do we know at this point?

Bob Burch: Yeah. So performance max has, has been in, been around for a number of years now, but it’s evolved pretty considerably, particularly in the last year. And I think one of the biggest updates in the last year, uh, was bringing a lot more transparency and understanding of how it’s working and where it’s running your ads. But at its simplest point, it is a cross channel campaign that you can run within Google. So rather than setting up individual search, uh, display video campaigns, um, shopping campaigns, if you’re e-commerce, you can essentially set up one campaign where you feed it a bunch of different creative assets, image assets, text assets, video assets. Um, and it will essentially, and, and set up a goal for that campaign. So whether that’s a sale or a lead or whatever it might be, and you’re basically giving it a target and saying, go out, find me new customers wherever they might be, and deliver on this cost per that we’re, we’re looking for there. Um, so because of the nature of it, it has always been a little bit of a black box. Um, and there was particularly early on, not a lot of transparency into where you were running, what was performing well, which for us, when we’re always trying to be in the campaigns, looking for opportunities to optimize when they take those levers away of what we can do to, to improve performance. It was always a little bit of a struggle for us. And I will say in the early days of performance max, it tended to work much better in a e-commerce space than it did in, in other types of, of categories. And a lot of that is because, uh, you can’t fake any commerce sale. If you’re tying it into your e-commerce platform, you can see that somebody made an actual purchase there, but things like, uh, form submissions on a website that might be a lead type thing, um, those things can still be faked a little bit. And so there was always concern around the quality. So you do want to make sure that you’re giving it a strong signal. So in our case for lead gen, we want to make sure that if we are running it, that we’re not just looking at a form submission that we’re also connecting into a CRM, whether it be Salesforce or HubSpot, whatever the client might be using. So we know that it is a legit person and then legit new contact that’s, that’s getting created. but the biggest thing that kind of changed this year, they added more channel level reporting, they added more search term reporting. So you can understand what’s running asset group level reporting.

Katie Flynn: Yeah. So there’s a lot of clients too, because they want to know what’s running as well. Like not only do we want to be able to actually make adjustments, but they want to know what the heck is going on too.

Bob Burch: And there were, there were always workarounds to get to that data before you could put custom scripts in and you know, Reddit has a ton of people who have built these custom scripts that you can go into run these reports, but it, it made, it was an extra step to go in. You want to see it in the platform so you can actually act on the data pretty quickly there. so adding a lot of that into it certainly helped. Um, also just a little bit more controls on things like excluding branded search from that. Cause you don’t just want it to go out and just be essentially getting people that you would have already gotten anyway. Um, so there’s just, there’s just been a lot of improvements to that, which has given us the confidence to test it in more and more categories. Uh, it’s still a struggle for, uh, for ones where we have very niche audiences or, you know, different criteria, there’s restrictions in financial service and healthcare in terms of how we can target

Katie Flynn: our two biggest segments. Yes. Yep.

Bob Burch: So, um, there’s always been a little bit more hesitation in those categories versus a standard, you know, standard business or a service related business. but we are, you know, testing it more and more. And for the most part it is, it is pretty successful. I think the biggest thing it can help with is expand the reach of your search campaigns first and foremost. Um, you know, a lot of it is, well, it’s running across all of these different channels that Google owns. We generally see about 80 to 90% of the spends does still stay in search. And so it’s a good way to find new keyword opportunities and test it. And it tends to be very, very low cost traffic for what we get there. So, um, so it’s, it’s definitely an interesting one. Um, you know, we’re excited to see the performance that we’ve gotten from a few clients already. And we’ll see where that continues to take us and continue to test it out in more and more places.

Katie Flynn: I think the creative aspect of it is really interesting too. I know we chatted about this. We had a funny, funny interaction between the account team and the media team a couple weeks ago when we were going to traffic some assets and Google was like, well, we, we can make all these changes. And so someone on our media team was like, yeah, no, they have what they want or need. And the account person and I were like, but did we give them source files? Like I don’t think we gave them source files. Like we were so confused. She’s like, they said they don’t need it. But I mean, of course they’re Google, like they can figure it out. And of course now in the platforms we can see they have tools like erase tools and all sorts of things to just really quickly update your content.

Bob Burch: Yeah. They’ve added more and more AI optimization on the creative side directly into the platform. Right. So as search marketers, we can go in and make edits to that when we’re going into our responsive search ads, it’s suggesting new headlines, suggesting new descriptions. When we go into the images, to your point, there is an editing tool that we can use to, if we want to take a logo out or move a logo to a different part or, you know, add a new image in or generate images directly. They have image generation capabilities in there. So give us more examples that we can feed into the account that we can get there. And it tends to be very like stock looking. Plus they have a lot of stock that you can use. The tricky part for us is that’s great if we are working with a client and a category that doesn’t require every piece of creative to run through a compliance team. And that’s just not a lot of who we work with.

Katie Flynn: Yeah, exactly. It’s like probably five to 10% of our clients.

Bob Burch: And even ones that aren’t in a heavily, heavily regulated industry, there’s still sensitivity around that that they don’t want.

Katie Flynn: Of course.

Bob Burch: Want us to just, you know, pick up and run whatever.

Katie Flynn: Yeah. Yeah. And I think we’ll talk about this in a minute when we get to met as Andromeda, because there’s some of that there too of this, like this idea of less curated content, but more learnings more quickly. So yeah, I think it’s going to be an interesting shift, but I do want to touch on AI Max for a moment because I know that’s more one that we’re trying to learn as much as we can about, you know, absorb as many case studies as we’re able to now. But that’s also one that just has less, even less control.

Bob Burch: Even less control. So AI Max is specifically an add on for search campaigns. So it’s just a switch that you can turn on. That one just, you know, just rolled out within the last couple of months. It got announced earlier in the year that they were bringing that in there. It’s a little bit similar to performance max again, that it’s going to find opportunities, except it’s very focused on search and the difference is enabling it within an existing search campaign. The difficulty there, it’s going to just basically have the freedom to find and add keywords to your account that it thinks, you know, ultimately would be beneficial and relevant to the search query. It also can do all of the content generation for you. So it will crawl your website and just write new headlines, new descriptions for the RSA, call outs, site links, all of that content gets dynamically created by anything that’s on your website with the idea that, you know, if it’s on your website, it must be okay to say and it must be accurate and it must be approved. And so that again is, there’s one thing where, you know, on the PMAC side, at least we have control over the creative. We have the opportunity to opt out of any sort of dynamically created elements. But with AI Max, it really is leaning into that and saying, I’m going to give you a target. I’m going to say, I want these conversions at this cost per and here’s my website. Google take the wheel. And you know, that is, that is always a difficulty there. We also struggle with, with any of these platforms, right? They’re using their performance signals. They’re grading their own work too. So you’re giving them full control to say, you know, do what needs to be done just to deliver this result. And then also we’re asking them to tell us, is it working or not? And so that’s why we have to validate those things that come in. So it’s, we are seeing leads coming in. If we are seeing sales coming in, what does the quality look like of that? Do we see a dip in quality or, you know,

Katie Flynn: Yeah, those down funnel metrics matter a lot. Exactly. Yeah. No, that makes total sense. And I think, you know, it’s, it’s one we’re going to pay really close attention to and I’m sure we’ll experiment with and try with various clients, but it seems like these new tools and types of campaigns and tools that you can apply on top of campaigns are coming out really, really quickly. And so we’re just going to keep embracing it as best we can. And, you know, in hopes that we’ll probably continue to get better and better and better as things go to.

Bob Burch: Yeah. And there’s, you know, the other big thing with Google right now is the prevalence of AI overviews and search results. And certainly that is impacting where on the page our ads might be appearing. So, you know, understanding what percentage of search queries for a given keyword are showing the AI overview. If it’s above the ad or below the ad, you know, we’re, we’re, there are more tools becoming available that can give us access to that. So if we can decide, Hey, if this is a keyword that is 90% of the time showing an AI overview and not an ad, that needs to be a focus for SEO or GEO for us and less a focus for paid search. Whereas the other ones that they’re going to do, and Google’s always going to balance that and throttle it up and down because Google ads makes them the money. So that will be part of it. But you know, it’ll be curious to see, especially with AI max, eventually at some point we do expect ads to be embedded right in those AI overviews. And the question becomes, if they’re going to do that, are they going to require that to be an element of being opted into AI max? So that’s, that’s what we’re continuing.

Katie Flynn: Yeah. That’s going to be a rapidly changing place that I think we’re going to have to pay really close attention to. And you know, with GEO too, you know, we have a lot of discussions around the ways that you can optimize your content and your customer experience and all of those things to show up in those summaries. Heather and I actually just did an episode on it a couple of weeks ago. So that’s posted prior to this, if anyone wants to go back and listen. But I think it’s one of those things where in three months time, in six months time, there will probably be so many more monitoring tools, so many more resources out there that are helping you to understand best practices at a really like minuscule level, right? Like really getting into the minutia of what will help. So I think that’s an exciting area too, but it will be interesting to see how it affects SEO and PPC because they’re all tied together in that little ecosystem.

Bob Burch: Sure. Sure. So it’s going to get continually messy and we’ll just, you know, what’s interesting with something like AI Max is Google introduces things. They put it into beta first, then they fully roll it out and it becomes an option. But we know with everything, eventually it becomes a recorder and not an option. And so we know this is going down that path where eventually, and that’s why we need to be testing performance max and AI max and finding ways to make it work. I mean, this goes back a ways, but you know, going back to simple static text ads that we were running versus responsive ones, we’re now we’re feeding them 15 headlines and four descriptions and a lot of different variations and letting them find the right combinations. For a while, it was an option. You can test these types of things, but you could still also have your ETAs, your expanded text ads. And then eventually they just got rid of the one and said, Nope, this is the way going forward. They’re going to force you down this path. But they are, they do tend to, you know, make it a slow transition because they also know they need to work through the bugs and get feedback from the advertisers. Totally. Testing it out.

Katie Flynn: Yeah. And just get people comfortable, right? Because yeah, there’s quite a, quite a lot of buyers on the old Google ecosystem. All right. Moving on. Meta’s Andromeda. So as I mentioned, AI ad delivery system. We had a staff meeting last week where Ashley on your team shared a bunch of information about this. My understanding is it’s basically an evolution of from where we used to like agonize over these two perfect, you know, meta ads. It’s just total opposite end of the spectrum for that. Like maybe we put in 20 images and 20 descriptions and meta is going to do the, you know, optimizing for us and figure out the best pairing for each individual customer or, you know, type of customer or whatever. But, you know, what is it that I’m, what is it that I’m missing on that?

Bob Burch: Well, yeah. So is another one that is just an evolution, right? So one of the things that was introduced the last few years is advantage plus both on the audience side and on the creative side. So the way advantage plus works in, in meta is you can set up a, an audience, but essentially rather than giving them very prescriptive targeting expectations for who you want to reach, you’re giving them suggestions. This is who we want to reach. This is the type of person that we’re looking to go after. Here’s the result that we’re looking to get. And they will go out and find the people that ultimately are going to respond to that. On the other side of that, they had some advantage plus creative optimization where it can take your ads and number one, make it work in all of the different ad formats that meta has. So we don’t have to, you know, create every unique, uh, size and, and, and variation that you need for that. Um, so there’s some of that optimization that’s happening. Andromeda is taking that a step further. For one, it’s flipping how we even set up and target the campaign. So what once was about, you know, finding all of these different audiences and maybe setting up unique ad sets for each audience that we’re going after, what they’re now saying is creative is now you’re targeting. So you’re feeding them, like you said, less about quality, more about quantity, high volume of creative that you’re rotating and pushing into the campaign on a regular basis. And it’s going to use the creative that it has to just go out and find the audience that is more likely to respond to your creative. So again, less control. It can start very, very broad, but just like with performance max, it should be very fast in terms of honing in on who your audience is. Um, hypothetically, if it’s working well, but that is, that is a big, big shift. Yeah. And it’s a big shift. We do everything audience first here. And so we want to make sure that we’re identifying the audience, finding the right ways to target that, whether that’s looking at different interests, different, uh, custom audiences that we can pull into the platform using first party data. So all of that thing that we’re pulling in there, and you can still give it audience signals, uh, as it comes in, but for the most part, you know, you’re moving from having, you know, maybe one campaign with five or six different audiences all set up in different ad sets to one campaign, one ad set, which is all of that, all of that stuff in there and just saying, go get me the result I’m looking for.

Katie Flynn: Yeah. It’s going to be very interesting. And I think not only is it a shift for the media team, of course, and for just all of us who are going to be watching for those results to, you know, have the faith that it will figure it out. Um, but a big shift for the creative teams, as I mentioned, it’s going from this like super curated, like, let me find the perfect picture to no, I need to find a bunch of pretty good pictures that represent the audiences that we’re going after. Um, and we’ll do a great job of, you know, optimizing this system. And I think it’s really exciting because I think we’re going to learn a lot. My hope is that we’ll learn a lot and we’ll get enough data back in the platform that we’re able to say, Oh, this was a really interesting insight. We tried this one image of a rocket ship and for some reason it did amazing. And why do we think that is? Obviously that’s a terrible example, but, the hope is that it’s going to help our teams get smarter too, uh, on all things that we do.

Bob Burch: Yeah. And so, you know, this is another area too, where there are great AI, AI tools that can help us create a higher volume of the creative or even create different variations of it. So there’s, there’s tools that you can use to, you know, take a static image and turn it, give it a little bit of emotion and, you know, kind of bring it to life a little bit. So if we are using, you know, some stock images that are out there, we’re creating those images with, with AI, um, we can add some movement to it. Cause we do know that video tends to, to perform a little bit better. Um, you know, we want to make sure that we have means of getting user generated content, whether that be customer testimonials or things like that, or, you know, people talking about it, your brand, but it can also be employees. It can be, you know, we can get content from those people, but you, you want more of that talking head type creative too. We know that performs well. So you want that in there with a mix of static images with the texts that we need. So as much variety as we can push in there, I think where we’re still trying to understand, you know, certain things is, you know, we have businesses that have several different product lines. Well, you know, meta might say that the best way to do that is just to push all the different products in there and let them find people that are responding to each of them. But a lot of times different product lines get different budget allocations. So when we set up unique ad sets more often than not, it’s because we are given a budget allocation of acts for this product. And you know, in some cases, again, our clients are reporting to different product marketing teams within their organization or product teams. So they’re also required to spend certain amounts of money against certain products. Whereas this is more of a, like some creative is going to win. There’s going to be winners and losers in, in, in certain aspects of this. And so it would also require clients to give us a little bit more of, of fluid budget with those types of things to say, you know, these are the types of products. I want to be promoting, but you know, our total spend for the month should be X and we know it’s going to, you know, shift between all of those things. But you know, we’ll be curious to see, I mean, we see even when we have a tight ad set in an audience in there and we maybe have five, six, seven different pieces of creative in there. When it does find that one piece of creative that’s working well, it becomes 90% of the spend very, very quick. And so, you know, is it going to just again, push everything towards the one that is responding? So there’s, there’s a lot of, there’s a lot more to be learned about that approach. I think what we can do as an agency and what we can do with our clients is at least start getting better at rapidly developing creative. Because even in the existing format, meta wants you to feed the beast no matter what. So we can still be feeding more creative into these campaigns on an ongoing basis. So that’s, that’s something that we can, we can do now without even shifting our entire targeting mindset.

Katie Flynn: Yeah. Well, it’s interesting to see where it’ll go. I mean, I think from a, like a trust perspective and for the clients that we work with and how their financial teams work, you made some really good points there. We’ll see, we’ll see how quick the adoption happens.

Bob Burch: Yeah. It was exciting too. When we, when, when Ashley was talking through it at staff, I loved how just how much dialogue there was from everybody within the agency, the account teams, the creative teams, everybody was asking questions. You know, we had three, four more things on the agenda and we didn’t get to any of it because everybody was just talking through all of this. So, um, you know, it’s, it’s an exciting thing. Uh, but yeah, that’s, there’s still so many unknowns with it. Yeah.

Katie Flynn: I hear. Yeah. All right. Our last topic. We have a browser called Atlas. Yeah. What the heck do we think this means for Google?

Bob Burch: Well, you know, it’s definitely a threat to them for sure. are a lot of different browsers out there that have competed with Chrome. Chrome has the dominant market share. Um, but obviously you have Safari that’s native on the iPhone. You have to change your settings. If you want to use Chrome. Um, you’ve got other browsers like brave, which are again, more focused on the privacy aspect of things. You still have some legacy things like Firefox. There’s some smaller ones out there. So you know, for the most part, again, Google has had that dominant share for, for a while in that space. as with anything, there will be a lot of early adopters or will be a lot of buzz. We’ll have to see how it functions and how it sticks. I think with Atlas, there’s a lot of questions around and really for open AI as a whole, when are they going to start monetizing within the, whether that’s chat GPT or within the browser in terms of ads. So, um, we’ve seen fragmentation be the trend in virtually every other tactic we’re running. This is probably coming to search now too, where, you know, we are seeing that and we see all some of that happen already. A lot more search does now start in Tik TOK. A lot more search starts in Amazon or in the retail platforms that you’re going to shop for versus going to Google first. And now you’ve added already added chat GPT to that layer. Now you bring a new browser into the mix that has AI just in embedded throughout the entire experience of what you’re looking to do. And you can ask it to go book you flights and you can do all of these things. It’s, it’s exciting. I think there’s a lot of questions about again, what will, what will it look like from a, you know, opportunity standpoint for advertisers? Yeah. How will analytics work? Are we going to be able to trust the data? If it’s now bots that are going out and completing these exercises on your website, how are we going to track that? Are they real people that are doing it? so there’s still, there’s still a lot to be said about that. I think it will, it will be interesting to test out. There’s also a cost associated with it. So that’s another consideration.

Katie Flynn: Adoption could be a little slower.

Bob Burch: Yep. Yeah. There’s, there’s going to be people who are willing to spend the money for that type of thing, but safari’s free, chrome’s free. So, um, you know, there are others who are just going to be happy using what they’re doing. It’s also just like anything else. There’s, there’s transition that needs to happen. Right. So, you know, when you are using something like Chrome, again, you might be using their password tools. You’ve got your payment stuff saved in there. You’ve got all your bookmarks. You probably have 40,000 tabs open right now that you would have to close to go to this other browser.

Katie Flynn: It’s a fair point. And I also have 40,000 bookmarks.

Bob Burch: Yeah.

Katie Flynn: So many.

Bob Burch: Yeah. So those are all things. And I mean, look, they’ll, they’ll create tools to just port these things over note. You know, the same way that it has happened when people moved from one into Chrome. Uh, so that will all, that will all sing.

Katie Flynn: No, that’s a good point though. It’s like getting a new phone when you have to sign into all your apps again, like special kind of torture. That’s exciting for a new phone. Yeah. Don’t love that part.

Bob Burch: Yep.

Katie Flynn: Um, no, I mean, it’s, you know, it’s a wild time we’re living in. There’s so many developments happening so quickly, but I think it’s also really exciting.

Bob Burch: the other thing I would add about that, I think there will be a lot of questions around security. There, there is definitely some sensitivity around what you want to feed into even chat GPT today and, and AI models. And we see it with companies having different AI policies about feeding information. They don’t want proprietary information going in there. Certainly on things that aren’t releasing. Sure. now you’re asking people to be comfortable also just putting their, payment information in there. Is it going to be secure? What happens if it gets hacked? Yeah. Um, you know, now there are still all of those same questions about Google and what they’re collecting and what met is collecting and what every, you know, there’s six companies that know everything about all of us right now.

Katie Flynn: I think probably a lot more than six.

Bob Burch: Yeah.

Katie Flynn: Well, yeah, but like the big ones, but like the dominant,

Bob Burch: you know, there’s these dominant ones that are essentially just, you know, own all of the data in the world. And then sell it to everyone and then sell it to everybody else. So, it’ll be interesting, um, monitoring for sure. I think what I would say as it relates to search, because we see conversations around like, should I be abandoning paid search? And the answer to that is even if they lose market share, there’s still a lot of opportunity to capture customers there. So for a lot of our clients, we, you know, we look at impression share metrics and for many of our clients, our impression share hovers in the 20 to 25% range. So even if they lost 30% of the people using there, there’s still people that are going to be using that tool and there’s still low hanging fruit there to go out in there. So again, we might have to expand, you know, whereas right now it’s like 90% Google and 10% Bing and, you know,

Katie Flynn: well, I was going to say Bing is a good example. Like we still run on Bing cause there are still people on Bing. I don’t know them, but they’re there, but they’re there.

Bob Burch: And it’s, yeah, it’s a significantly lower investment than what we put into Google. But there are people using it. It will be the same idea. So we’ll just have to be a little bit more diversified with what we’re doing. Um, but I think the other side of that too is again, the, you know, performance max there. We need to be expanding to other channels and testing other things beyond just straight search. So we’ll continue to move that switch, but you I don’t think it’s right for people to ever make the searches dead type of comparison with this search is not dead. It is evolving. Like yellow pages became Google search. Like this is an evolution. It will continue to evolve and it will continue to get more and more fragmented and will shift as, as that behavior changes, but it’s not going to happen overnight. And, um, you know, there have been some big changes this year that the AI overviews, like that is the last six months or so thing that has really taken off. So, um, you know, where we’re seeing it on a higher percentage of searches. So all things that we’ll monitor, but you know, search should still be a core part of your marketing mix, even if you are maybe pushing and expanding to more of a full funnel play.

Katie Flynn: Yeah. And we’ll just keep up with it all and we’ll keep letting people know. I’m sure in three weeks, we’re going to be talking about something