← The Digiday Podcast
The Digiday Podcast
Google’s AI Mode, OpenAI’s io buy, Publicis’s Capitv8 capture + Google’s AI ads era with Dan Taylor
00:38:42
Loading summary
Transcript83 lines
- [00:00]
Dan Taylor
Foreign.
- [00:09]
Kamiko McCoy
Hello. Hello, and welcome to another episode of Digitay Podcast. I'm your co host, Kamika McCoy, senior marketing reporter here at Digiday.
- [00:16]
Tim Peterson
And I'm Tim Peterson, executive editor of video and audio at Digiday Media. What's up, Kimiko?
- [00:21]
Kamiko McCoy
How are you, Tim? We've got no shortage of news happening, whether that be out of the White House, out of AI, out of everything else. How are you holding up through all of this?
- [00:32]
Tim Peterson
Yeah, as usual, we have to timestamp this conversation because who the hell knows what's going to happen between now, when we record it and when it goes out Monday night, Tuesday morning. So we were recording this Friday, May 23rd, at 9:00am, 9:15am Pacific Time, my time in California. Yeah. Woke up this morning to reports that Trump's thinking about tariffs on the European Union starting next month, which starts in a little over a week as we're talking. Unclear if this is going to happen, but it's just like, you know, last week I was in New York for upfront week and some of the conversation was like, okay, there's this pause on the China tariffs. You know, that's good. There's still tariffs that will be in effect, but a pause on that, 145% tariffs. Maybe things are going to be okay. And now European Union deserves.
- [01:34]
Kamiko McCoy
And then you are quickly guided back to reality where the only thing certain is uncertainty. Yeah. Trump made the announcement. Announcement, excuse me, to impose a tariff, I think it's 25% on Apple if the company does not start manufacturing iPhones on this side of the pond.
- [01:51]
Tim Peterson
Yeah. So that's obviously a story, as with everything else, that we'll continue tracking, but we got a, A lot to talk about in this episode because there's been a lot of movement, especially on the AI front this past week. And so for. Our featured guest for this week's episode is Dan Taylor, VP of Global Ads at Google. Google had two big events last week. It had I O, which is its big annual developer conference, and then it had Google Marketing Live, which is kind of like the advertising version of I O. So it makes sense that they happen on subsequent days. And so we talk with Dan about one of, well, some of the bigger announcements there, which were related to how Google's updating its search business for the AI era.
- [02:37]
Kamiko McCoy
Yeah, it was really interesting and great timing given, you know, we're just off the heels of those events and had some questions teed up for Dan. But before we get to the conversation with that, as always, we're here with our Juicy scoops of the week, including, you know, one of those being the rollouts that came out of Google's announcements. OpenAI, which has acquired IO. It was the biggest acquisition to date. That took me a while, but I got it. And then Publicis, which is bought influencer marketing agency Captivate. So we've got some juicy scoops for you and we'll go ahead and break them down.
- [03:11]
Tim Peterson
Yeah, we were so close to having all three have either AI or IO as part of it, but at least the first two AI and I O so as mentioned, Google had its IO developer conference and announced that it's rolling out AI mode for search for people in the US for anyone who hasn't played around with it, it basically takes the traditional search engine results page and search experience on Google and turns it into like a chatbot, which is a huge change for all kinds of reasons, one of which being if you're getting all the information in search, to what extent do you need to click on links anymore?
- [03:52]
Kamiko McCoy
Yeah, I think that is a big. That's been the boogeyman for publishers for a while. Right. And now that fear has come to fruition, as Sarah Guaglioni, senior media reporter here at Digiday wrote in this week's briefing, is that, you know, publishers are concerned about that. And to your point, Tim, they're pretty much satisfied when a user goes to these search engines. They're satisfied with what comes out of the AI optimization. There's what incentive there is there to go and get further information from that link. And it almost seems like the industry's latest or the publishing industry's latest boogeyman, because we saw the same thing with Meta and the pivot to video and whatnot, and AI just kind of seems to be the latest tool there.
- [04:35]
Tim Peterson
Yeah. And Dan makes the point and the interview that Google's actually seen AI. I forget if he was talking specifically about AI overviews or AI mode, but that it has led to an increase in traffic to sites. I'm curious about that because I know in my own experience when it comes to AI overviews and I've been playing around with AI mode this week, I'm definitely less inclined to click through to links or I start entirely new searches to then find authoritative sources for the information I get in AI overviews or AI AI mode. So yeah, a lot of the publishers are very concerned about that. Another thing that's interesting. So Search Engine Land published a story saying that AI mode traffic isn't trackable, that if you are a website and you're getting traffic through AI mode, it's going to look like it's just Google search traffic. You're not going to be able to see like, oh, this is specifically coming from AI mode, which I don't know, seems like a questionable move because if, if Google is seeing AI mode drive traffic to publishers, I would think Google would want to communicate that. Unless Google is concerned about maybe that not being the case eventually and so not wanting to have that be so transparent to website owners.
- [05:55]
Kamiko McCoy
I was just about to say, to Dan's point, if Google's grading their own homework here, then that kind of raises the question of how much can I rely on what Google is saying here if they're saying search is coming, but to your point, you know, there's no separation or breakout of what that search, those results and those traffic numbers look like. Then again, there's questions and concerns.
- [06:17]
Tim Peterson
Yeah. So hopefully it's something that Google eventually does provide because, you know, just as comes up in the conversation with Dan with Performance Max, it's AI based, you know, ad optimization product. Google historically hadn't broken out channel reporting. It is now providing that transparency to advertisers. Unclear if that transparency is going to go so far as to breaking out when ads are clicked within AI mode or AI overviews, but at least there's some precedent for Google not providing the information and then eventually at making the information available. Speaking of information being made available, OpenAI has made its biggest acquisition to date which seems primed to create a new way for people to get information out of ChatGPT. So they've acquired IO, which is Jony I've's company and it seems like the plan is to create some sort of hardware device that would have chatgpt built into it.
- [07:22]
Kamiko McCoy
Pal. I'm starting to think this AI hype is no longer just hype.
- [07:28]
Tim Peterson
Just starting to think that I'm starting.
- [07:30]
Kamiko McCoy
To think we might be for real here. So the Wall Street Journal did a reporting on it and kind of describes this. What could be right? The plans for this device as being capable of fully being aware of a user's surroundings in life and able to rest in one's pocket and whatnot. Which there's a movie that seems similar to this. No Her.
- [07:51]
Tim Peterson
I love that movie. My better half hates that movie. Unfortunately, it seems like she's going to be living that movie in late next year potentially.
- [08:01]
Kamiko McCoy
Yeah, I think, I think it's really interesting. But also to the point about movies like we've kind of seen this play before. There was is his name? Marcus Brown. Marquis Brown.
- [08:11]
Tim Peterson
Marques Brownlee.
- [08:12]
Kamiko McCoy
Yes. Who? I can't remember. You know, it was within recent years where there was humane, which was a similar faction, whereas this little wearable pin and it got just shredded in her review and there was a big hoopla about it. And the interesting there thing there is that Sam Altman had an investment in that as well. So maybe the hope is that with John I backing this and it being Apple, that this got a better standing chance, better fighting chance, but we'll see.
- [08:42]
Tim Peterson
Yeah, yeah. Because I mean, I think the issue with the humane AI pin was that it just didn't work well. It sounded like there was a lot of latency and Marques Brownlee was, I think, the lightning rod for a lot of that. But it got panned pretty much across the board. If I remember right, I think this is only like last spring. And so I don't know how much of that is a referendum on people don't want another device or an AI device as much as it is. Just like people want devices that are good and are going to be worthwhile. Because according to that Wall Street Journal report, they describe this OpenAI hardware device as capable of being fully aware of a user's surroundings in life, will be unobtrusive, able to rest in one's pocket or one's desk, and will be a third core device a person would put on a desk after a MacBook Pro and an iPhone. So shots at iPads and watches and glasses. But it's hard to think of what it's going to be if not a pin or some sort of earbud thing. But then if it's capable of being fully aware of a user's surroundings in life, that makes me think it's going to have some sort of vision capability.
- [10:02]
Kamiko McCoy
Yes. I'd also one I'm curious, like, I don't know. I feel like Meta has had similar ambitions when it comes to like Meta glasses and Ray Ban and whatnot and AI powered things. I don't know how far those things have gotten. I feel like they've been pushing them for years.
- [10:17]
Tim Peterson
I feel like it's actually starting to work with the Ray Ban Meta glasses.
- [10:21]
Kamiko McCoy
We've arrived. We've arrived.
- [10:22]
Tim Peterson
It's like it's become, it's a more affordable price point. They don't look stupid like Google Glass. And I know like, at least, you know, there are a lot of like YouTube creators that seem to talk about them and actually be using them. Like I don't know how much of that Is because like you do a video about meta. Well, not even getting paid by meta, but just if people are really interested in the technology and so you get views and therefore like indirectly get paid off. That. That said, if those videos are getting views, that means there is an interest in this kind of thing.
- [10:55]
Kamiko McCoy
Yeah, that's fair. I'd also be curious kind of how this all pans out, because if it's, if it's setting itself up in to be a competitor to Apple, like will it fit into Apple's ecosystem so that people even adopt the tech so that you're so like my MacBook and my watch and my phone all kind of work together. If you're adding another device to what I am doing right. It needs to kind of flow into that. That workflow. And then also. Yeah, I'd be curious if we've abandoned. You said this on stage at the Programmatic Marketing Summit, if we've abandoned data privacy at this point as to that being so close, the technology being so close, you know, and fully aware of a user's surroundings, which is given smart house to me.
- [11:34]
Tim Peterson
But yeah, I guess that's where now we're getting a little bit the deep end of things. But like, if we're just talking about on device computing because like, so this would be not an Apple device. Jony, I've is not at Apple, has not been at Apple for years at this point, but obviously was very integral in all of the Apple devices that we kind of like know and have, you know, iPhone, iPad, you know, watch MacBook Pro, MacBook Air, all of that. So. But Apple with its approach to AI, which hasn't worked super well, um, for anyone who's used Apple intelligence would know that. But they're relying on, on device computing, which is basically like keeping as much of the data limited to the device itself as can be, which mitigates some of those privacy concerns. I don't know. I think I would actually be interested in a device like this, but I think I lean like more into the AI. Curious than most people. Kimiko, would you want an AI, a ChatGPT device?
- [12:48]
Kamiko McCoy
Absolutely not. I'm running to unplug my Google home right now.
- [12:53]
Tim Peterson
Okay, so in that case, let's move on from this conversation to our third piece of news. So Publicis has acquired Captivate, which is an influencer marketing company. Publicist is paying $150 million for captivate, according to the Wall Street Journal. And this is Publicis second big influencer marketing acquisition in the past year. Last year, they acquired influential now they're acquiring Captivate.
- [13:20]
Kamiko McCoy
There's a lot of movement in this space and influencer marketing at this point obviously has become like a pretty standard line when it comes to the marketing strategy and media plan. I think what's interesting is that eMarketer and a handful of others have kind of reported that there's going to be consolidation happening in this space. And I think we're starting to see that at this point. But, you know, all of these, they're competing with one another to differentiate themselves and that's getting harder and harder to do.
- [13:46]
Tim Peterson
Yeah. And it seems like there is some differentiation in this case between Captivate and Influential. Both are influencer marketing companies. But it sounds like Influential lean more towards being able to do the matchmaking between brands and influencers. And it was more of a like client services was its skill set or its differentiator. Whereas Captivate is more about having the tech to manage influencer marketing. It seems like there's a big emphasis on social commerce. So it seems like the two kind of come together in that respect where Influential brings kind of that brand influencer matchmaking skill set, Captivate brings the tech to then manage those matches.
- [14:31]
Kamiko McCoy
How do you think that this sets publicis up like influencers? It seems like influencers at this point kind of become like media properties in their own right.
- [14:41]
Tim Peterson
Yeah, no, that's. I mean, ever since the nano influencer trend of like, what was that, like 2017, 2018, there's been this trend of marketers treating influencers more like media properties. Because, I mean, if you're working with whether we call them nano influencers, micro influencers, but when you start to work with influencers at scale, you're not working with them on A1, on one basis. You probably as a marketer, let alone an agency, are probably never actually talking with these influencers, meeting with them in any respect. It's sort of, here's the brief, go make the content, show us the content, show us that the content worked. Or give us some way of tracking the content. Here's your money.
- [15:29]
Kamiko McCoy
Yeah.
- [15:30]
Tim Peterson
Kind of thing. Which is media buying.
- [15:33]
Kamiko McCoy
Exactly. I'd be curious kind of what concerns that influencers have as that trend kind of continues. Because already I think influencers and reporting that I've done and a handful of others at Digiday have done is like creators and influencers have already been fighting for their voice to get paid on time and things like that. You add another layer on top of this. Now there comes questions about the pipelines and whatnot.
- [16:00]
Tim Peterson
Yeah. Because it's really about the economics of these companies. It's always about the economics. It's a stupid thing for me to say. I'm being insightful here, but these companies are being. These influencer marketing companies largely are being paid by the advertisers or agencies. So those are their customers. That creators are just a means to an end with that money. Now you do have talent managers, talent agencies who they work with the creators, with the influencers, and they make their money based on taking a percentage of the money that the influencers or the creators make. It just feels like so many of the companies in the influencer marketing space skew towards the buy side and now are in fact becoming part of the buy side that I get to wondering how much agency, I guess, if you will, will influencers have or is this going to be similar to what we've seen in digital publishing where the publishers get disintermediated from the buy side? Like to what extent is there a supply side platform for influencers?
- [17:08]
Kamiko McCoy
Yeah, yeah. I think the roadmap there is still shaking out, similar to how Dan shook out the roadmap for us. Did you like that transition? Shook out the roadmap for us for Google's latest announcement. So Tim, remind us one more time of the conversation that we're going to be having with hand here.
- [17:26]
Tim Peterson
Yeah, so we got Dan Taylor, who's the VP of Global ads at Google, on the podcast. We talked the day after Google Marketing Live during which Google announced that it's testing ads in AI mode and that it's also expanding AI Overviews, which also has ads. So we talked with Dan all about just how Google is transitioning its traditional search advertising business into the AI era, specifically when it comes to AI overview use in AI mode.
- [17:54]
Kamiko McCoy
Perfect. Well, without further ado, let's get to it.
- [18:04]
Tim Peterson
Hey Dan, welcome to the show. Welcome back to the show.
- [18:06]
Dan Taylor
Yes, thank you. Good to be here.
- [18:08]
Tim Peterson
Yeah, thanks for joining us. So we are talking the day after Google hosted its Google Marketing Live event. A lot of announcements that were made, a lot of AI related search announcements which we definitely want to dig into. I watched the full keynote. I think it was like close to an hour and a half if I'm not mistaken. A lot in there. I didn't see anything in there about third party cookies or privacy sandbox though.
- [18:33]
Dan Taylor
Well, yesterday's announcements were really focused on our core ads business and how AI is transforming every phase of the marketing journey. And so we really wanted to lean into the major ways that we're seeing marketers take advantage of that technology today and how we're making it easier for customers and businesses of all sizes to start putting AI to work for them.
- [18:53]
Tim Peterson
So AI doesn't figure into third party cookies or Privacy Sandbox then?
- [18:58]
Dan Taylor
Not directly. So Privacy Sandbox is an initiative that is being run by the Chrome team and is focused on creating a more private web, but not immediately related to any of the announcements we made yesterday.
- [19:09]
Tim Peterson
Okay, so let's talk about, I mean one of the bigger announcements that was made yesterday was the rollout of or the testing of ads within AI mode, which is the version of Google search that more closely resembles a chatbot experience. I've been playing around with AI mode since it rolled out yesterday. How are ads going to figure into this? Because it is very much like a chat experience as opposed to the typical we're past the days of 10 blue links in the traditional search engine results.
- [19:38]
Dan Taylor
Page, but we certainly are far beyond those days. I think I've been at Google for 19 years and as the ways that people search has evolved, Google has evolved search along with it. And we think that AI overviews and AI mode and these generative AI driven experiences are pretty transformational change, but they're in line with changes that we've made over many years to adapt to mobile and high speed broadband and more. And so pretty excited about that. But to your specific question, we're starting to test ads in AI mode and where they will appear is either below or integrated into AI mode responses within the US on both mobile and desktop. I think an important addition to that is we only show ads today on any search engine results page when we feel an ad would be relevant to what's being searched for. That will hold true in AI mode experiences as well.
- [20:36]
Kamiko McCoy
Follow up question to that. How. What's the kind of plan for it to be determined how an ad will appear? Right? How is that content separated from what people are searching for?
- [20:47]
Dan Taylor
So generally speaking, when people come to Google, I'll put them in one of general two categories. The first would be they have a pretty good idea what they want, right? So let's say you're searching for a business traveler and you're searching for an expandable suitcase. So got a pretty good idea what you're looking for. You maybe just want to compare brands, you want to compare prices, you want to look at reviews and you want to make a purchase pretty quickly. But then there are these other more exploratory searches which may be an appropriate place for an AI overview or AI mode to be the right experience. To get the best answer to that. So that would be essential. Business travelers packing list for Europe in July. That's a difficult query for a search engine answer with one thing. There's no single right answer. It could be, well, are you traveling light? Are you looking a backpack? If it's in the summer, you probably need lighter weight clothes. It's going to be hot in a lot of parts of Europe. You might need an international ESIM card to go with your travels and more in that kind of experience. With an AI overview or an ad that would appear in an AI mode, that search that is longer and more complex could generate multiple different queries underneath that are related to what's the best answer for this? What are all the different things that would go into what a business traveler would be looking to do or to have on their list before to go to Europe in the summer? And so that could create more than one ad opportunity. It could create more than one different type of ad that could be eligible to show in that experience. And so hopefully that answers your question.
- [22:27]
Kamiko McCoy
Will we be seeing any more of our Google's famous blue hyperlinks in there for the ads?
- [22:33]
Dan Taylor
Well, you can see a little bit of how the ads and AI overviews appear today, especially if you're in the US and so we're in pretty early days for how those format experiences show up alongside and above and below. But they're clearly identified as ads. And obviously as those experiences evolve, we'll expand the formats with them as we have over time. And so one of the things that I'm excited about as someone who works in the ads business is what we're seeing with ads and AI overviews is that people are engaging with those links, including the ones that are ads. They're spending more time on the publisher and advertiser sites when they click through, which means it's more quality and engaged traffic. The monetization is happening at the same rate that we would see in kind of the non AI overview monetization model. And so that gives us a wonderful place to build on from an innovation and formats perspective. And so to your point on formats and what that could look like, we're certainly looking to build and innovate in line with user experience as well as advertiser results.
- [23:42]
Kamiko McCoy
Got it, Got it.
- [23:43]
Tim Peterson
Will they always be separated from the actual content of the conversation, or will there be hyperlinks within the response that I get that are effectively like affiliate links?
- [23:57]
Dan Taylor
As long as I've been working in the advertising business with Google, the ads team and the search team have been very separate. Now, obviously, we work together to make sure that we create ads experiences that integrate well into the overall search experience. But ads will continue to be clearly identified, clearly delineated in terms of how that looks in terms of actual visual representation. I have no idea how that will evolve over time, but our core principle will be to make sure that they're clearly identified and separate from what we call organic content. And AI overviews is a form of organic content as well.
- [24:30]
Tim Peterson
How are these ads being priced, especially the ones that are being tested in AI mode? Are these on a cost per click?
- [24:37]
Dan Taylor
So it really depends on how the advertiser has their campaign set up. But the easiest way to think about it is ads that can appear in AI overviews or ads that could appear in AI mode, although we're still just testing there, they are an extension of your existing search and shopping campaigns. So advertisers don't need to specifically create a new campaign. They don't opt in or opt out of that experimentation. It's all part of the same campaign. It's similar to when we expand in other surfaces. We want to make it as seamless as possible for advertisers to participate, and so we're taking that same approach here.
- [25:14]
Tim Peterson
So what are the pricing models then?
- [25:16]
Dan Taylor
Well, they're the same pricing models that we have historically had in Google Ads. There's cost per click. We also offer the opportunity to buy on a cost per acquisition basis, focus on a target per return on ad spend and multiple other models. I can send you the full list, but variety of other ways that advertisers can choose to optimize their spend. But we're not introducing new ways as part of a rollout of AI overviews or ads within them.
- [25:40]
Kamiko McCoy
Talk to me a little bit about kind of striking a balance here. Right? How many ads are going to be shown per response or per session to make sure you balance monetization with the user experience?
- [25:51]
Dan Taylor
I don't have an exact number to give you on that because we're still experimenting with the overall approach and how we roll out AI overviews and how we roll out AI mode. But it will be very consistent in the approach that we've taken in every other search experience, which is we want to make sure that we're providing ads that are helpful and ads that make sense and are relevant to the query. But in terms of how many ads above, how many ads integrated, how many below, we're going to be focused very much on making sure that the user experience is Right. And then the monetization opportunities will follow.
- [26:26]
Tim Peterson
Kamiko raises a really interesting question specifically with AI mode, because traditional search, I input one query into the Google search bar and I get all the results and the ads that follow. And if I want to do a new query that's effectively a reload of the page, with AI mode I can be asking follow up questions and it's all just one query that can have multiple subqueries. Will ads be refreshed and evolve as the query goes on? Because again, I was playing with AI mode, now I can go back into AI mode, see all the different, I don't know, are we calling these searches, are we calling these conversations? But there's a history there. And so conceivably a month from now I could go back into one of these searches, continue that conversation, would I see different ads then?
- [27:19]
Dan Taylor
So as part of the overall search experience today, as well as with these evolving AI powered experiences, we want to be able to add context for what a person is searching for. Not just the words on the page, but the intent and what they're trying to accomplish. And so to the degree to which you saw some of the demos in AI mode, specifically, yes. Being able to have an ongoing conversation and understand the context of the queries that came before it will be part of the user experience and thus can also inform the advertising experience that comes with it. You do raise a good point. That I think is what I think the most exciting part as an advertising professional is that in today's landscape, traditional media planning tactics for search engine marketers just aren't going to be able to keep up with the evolving way that people are searching searches. 15% of the ones that we see every day today are new. And not just new. They're getting longer, they're getting more conversational to your point, Tim. And keyword based planning just is not a durable way to approach that. Which is why we really leaned into launching AI maps for search campaigns, which gives advertisers the opportunity to power their search campaigns with AI as these search experiences evolve, enabling them to show up against queries that are relevant to their business, not just the queries that are on their keyword list.
- [28:38]
Tim Peterson
So you have AI max for search, now you have performance max. What's the difference between, why not just call it performance max for search?
- [28:46]
Dan Taylor
You have to talk to the marketing team about that. But if you want a job in marketing, so performance max, we utilize as all of Google's best AI capabilities rolled into a single campaign across all of the surfaces at which Google can address ads, right? Search, display, video, owned and operated publisher partners, and so on. And so AI Max for search campaigns allows you to take a lot of the same technology that you can achieve on the search part of performance max and apply that specifically to your search campaigns.
- [29:21]
Tim Peterson
So I'm trying to understand like aimax for search and like what all is different about it compared to traditional search based bidding, keyword based bidding. Like historically an advertiser would say, okay, Google, here are my keywords, here's my budget, go get me sales or leads or whatever it is. My metric is Aimax for search seems to be, I imagine it's Gemini coming in and saying like, okay, cool, here's the advertiser's keywords. But I know context. The whole value of large language models is understanding context. I actually think in addition to these keywords, these are other keywords or combination of words that could apply for this campaign.
- [30:02]
Dan Taylor
Traditional search engine marketing would be, I have a keyword list and maybe I've applied phrase match or exact match or even broad match to say, within boundaries. I want my ad to appear when someone searches something in the neighborhood of this list of keywords. What AI Max research does is it does three things. Number one, it moves beyond the actual queries and tries to understand the intent of the searches that are happening. So moves from a semantic understanding to really digging into a large language model understanding of the intent of the search. That's number one, expanding the number of queries that you can appear for that are relevant. Number two, it also can customize the ad text in real time so that it's relevant to the user's query. And number three, it can also determine the best landing page within an advertiser's set of assets that will match that query. So it does all three of those things at once for this kind of expanding universe of searches that we see people come into Google for today that are new and again, people are coming to Google to ask different types of questions and we want to be able to give them answers for those questions, including ones that are commercially relevant for businesses.
- [31:18]
Kamiko McCoy
Here's a question that I think oftentimes marketers and advertisers have is kind of how the sausage gets made, right? Especially when it comes to AI, I think pmax has often been critiqued as a little bit of a black box. So how do you guys kind of relay the transparency and how the sausage is made to marketers and advertisers so that they understand?
- [31:38]
Dan Taylor
Yeah, thank you for the question, Kiko. So Yesterday at Google Marketing Live we announced channel transparency and reporting within performance max. And so now you can see where your traffic is coming from, where your impressions are showing, where your conversions are coming from in your performance max campaign. So by search, by display, by video and YouTube and more. And so that level of transparency helps advertisers understand how their working media is actually operating under the hood from an AI max for search campaigns. Going back to Tim's earlier question, all of the AI powered enhancements that I mentioned to you in terms of what queries are your ads now eligible to show on that weren't necessarily on a keyword list, what ad text showed up for that user in the query, and what landing page was ultimately delivered for users that clicked. All of that shows up in your reporting. So the advertisers can see what exactly happened with their campaign, including what are these queries that I'm showing up for, what ad text is starting to resonate and what landing pages are getting the most attention. And all of that again feeds into a better understanding for the marketer of their working media and powered by AI.
- [32:49]
Tim Peterson
Will Google be breaking out AI mode and AI overviews within the channel reporting?
- [32:55]
Dan Taylor
I don't know the answer to that question, but we can look into that, I think, because we're just now testing ads in AI mode. I don't have an answer for that.
- [33:05]
Tim Peterson
Going back to AI max for search. So you clarified the difference between AI max for search and traditional search based bidding, but now you've thrown in smart bidding exploration. And I've been trying to understand what that. I even had a conversation with Gemini. I brought it into NotebookLM yesterday, tried to get a conversation there. It seems like smart bidding exploration is just taking aimax for search and say now let's get even looser with this.
- [33:34]
Dan Taylor
It's a slightly different approach than you're describing. And so smart bidding exploration is an enhancement to our bidding technology specifically. And so advertisers are still going to want to specify what keywords and what categories they're interested in bidding on, including whether or not they want to be more expansive in the keywords and user queries that they show up against. What smart bidding exploration does is it combines an AI powered predictive model with a lucent return on ad spend target. And so what that means is historically, maybe going back to our prior analogy where we talked about before versus after before, if you were to set a target return on ad spend, for example in your account or in your campaign in Google Ads, it would apply that return on Ad spend requirement to every ad in your campaign. With smart bidding exploration, it gives the algorithm the freedom to bid opportunistically or in some terms higher on queries that might be a little bit more difficult to win, but also might result in the return on ad spend that you're requiring. Quite frankly, it's exactly what we named it, which is it's an exploration of bidding to help uncover new categories of queries and new types of queries. And so the most basic example that I can give is if you're a financial institution and you do home loans, looking for someone searching for a mortgage, that's a pretty clear area that I'd like to bid on. I have a pretty clear return on ad spend target, and it's pretty tight, right? It's pretty aggressive. But places to buy homes in Hoboken, New Jersey, that's a bit more exploratory. Maybe someone's looking for a real estate agent, maybe someone's looking for a school district, maybe someone's looking for a home loan. So I might want to bid on that, but I have a different return on ad spend expectation in order to take advantage of both of those opportunities within the same campaign. Smart bidding exploration allows you to tap into that.
- [35:37]
Tim Peterson
Okay, so yeah, I think that's where I got hung up because that was the example in the keynote and the blog post. And it was really just focused on the queries, not so much the flexible return on ad spend. So can an advertiser set their plus minus on like, here's my return on ad spend. I'm willing to go x percent above, x percent below on this. And how do they also then manage the budget? So the AI isn't going going so far over the ROAS that the advertiser is like, crap, I wasn't planning to spend 10, $100,000 more than.
- [36:11]
Dan Taylor
Yeah, certainly there's going to be controls for an advertiser to operate within their daily spend limits and their budget limits and ROAS expectations within parameters. And we can share more of the details on those specifics. But yeah, it's not target roas gone wild or anything like that.
- [36:30]
Tim Peterson
Okay, but those details seem hard to come by at the moment.
- [36:34]
Dan Taylor
Okay, well, we can get those here.
- [36:35]
Tim Peterson
Okay.
- [36:36]
Kamiko McCoy
I guess the last thing that I want to ask, and this is a more zoomed out, broad question here, is like, you know, Google really built its business on its traditional search engine, right? So how do you kind of, I guess, balance the push forward toward testing AI and investing in that while also maintaining the traditional business?
- [36:55]
Dan Taylor
Well, I would say we built our business on making information available and making our products helpful to users. And so as consumers are searching in different ways and expect different things from their engagements with technologies, we want to meet them in that moment. And AI, particularly the Gemini models that we built within Google, are really enabling us to do that. And so in many ways I see this as an evolution, although a really big evolution of how our search solves that core need for all of our consumers. And so we think about, we said a couple of times here, the initial 10 blue links on a desktop browser to the massive shift of consumers towards mobile and the reinvention that happened there with search. And I think this AI transformation is just as big, but it all is still in service of what can we do to be helpful to users and get them to the information they're actually looking for. One of the things that Larry Page said early on in my time at Google, some large group meeting was the perfect search engine would understand exactly what you want and then give you the exact information that you need. And AI powered technologies, including AI overviews, AI mode, I think get us even closer to that vision.
- [38:07]
Kamiko McCoy
Well, thanks so much.
- [38:09]
Dan Taylor
Yeah, this is a lot of fun. It's great to see you again.
- [38:16]
Tim Peterson
Thanks for listening to this episode of the JJ Podcast. If you enjoyed it, please leave us a rating and a review on Apple Podcasts, Spotify or wherever you're listening. Get more from Digiday with our daily newsletter sent out each weekday morning. Visit digiday.comnewsletters to sign up.