Marketing lessons from 20 years at Google - Dan Taylor

19:43Dan Taylor:It's funny how you're connecting the two things brilliantly in terms of people are searching for such deeper level of detail than they were. And this helps to simplify that process. I was adjusting. So I like to go cycling. And on the bottom of your shoes, you have this thing called cleats, which basically attach you to the pedals, right? And it's very important the cleats are in exactly the right position. I was trying to just adjust the cleats. but i realized yesterday it was a non-standard screw that had been used to put the cleats in i'd never seen the type of uh fitting that these these shoes had so i literally was like what on earth what what what you know what uh tool do i need to try and adjust my cleats so i did the same thing like right what is this circle it what am i what am i what am i looking up i didn't know what to order i'm like what is this thing you know it's a star shape rather than the cross shape which most of them most of them are so you know it just it overcomes that barrier of ignorance that so many of us have and then connects it to being able to buy it as well, which is so handy.

20:38I think the other piece that's gotten a lot of marketers to move to an aha moment is with these longer searches and with experiences like AI overviews, it's creating new opportunities for brands to get discovered perhaps earlier in their journey. And so maybe a personal example. My wife has a cat. She thinks the cat gets lonely when we travel. We don't really know if she's lonely or not, the cat, that is. And so we're like, oh, maybe we should bring the cat with us to London. And so what does it take to bring a cat, you know, on a plane and overseas, right? That's a, in the traditional days of search engine marketing, that's not something that a brand would necessarily bid on.

21:23But it does trigger an AI overview and it gives me tips on, well, you know, maybe you need to give your cat a sedative and perhaps you need a cat carrier. In the UK, do you know that you need like a medical certificate?

21:35Dan Taylor:No, I didn't know that. You do. And if you don't have one, you have to quarantine your cat for four months. Wow. And so we're not doing that. As part of that AI overview, the mention that you might need a cat carrier creates a commercial opportunity. And so there's an opportunity for a brand to put a product that is relevant right in the moment that maybe didn't exist before. And so these are the kind of opportunities where brands are like, okay, these longer, more complex queries actually do create commercial opportunities at the same time. And actually, one question I was going to ask you about, which I think this helps to answer, is the opportunity for smaller, more nimble brands to create demand.

22:12Dan Taylor:So in that situation, presumably a smaller brand, by being much more targeted about the circumstances in which they show up, like a cat-carrying business, could probably take advantage of that, rather than previously the bigger brands would have dominated the search outcome. I'll give you another example, actually, a real one from yesterday. So I wanted to get a new pair of jeans to wear to today's podcast. Not that it'll matter. Imagine, ladies and gentlemen. Yeah, exactly. And I typically wear Levi's. And so I did a search for Levi's, you know, waist size 30 near me, couldn't find any. And that's sort of like a typical search engine marketing way of searching, right?

22:48Like very specific, the brand location, right? Search like a search engine marketer. And I said, well, okay, let me think more expansively here. And I said, I want a pair of dark wash jeans, slim fit. I want to be able to try them on in store this weekend. I'm going to be in London. And I see an AI overview and I see an opportunity for Rag & Bone. So I end up at Rag & Bone. Brilliant. Yeah, and so that's a brand that I wouldn't have thought of before.

23:19Dan Taylor:Love that. I've got a funny story actually on coming on to the impact that AI has had on search. so about um uh was it maybe two or three years ago uh a youtuber uh did a seven minute kind of takedown video of this podcast right of this podcast he did a ranty video for seven minutes uh i got called tyrannically british bullshitter and this kind of thing he gave me a two and a half two and a half star review which i thought was quite you know generous of him given he spent seven minutes you know telling me how bad this show was and uh anyway i i thought what do you do in this situation, right? You can either get upset or you can kind of engage with it and see the funny side.

23:56Dan Taylor:And I reached out to him on LinkedIn and just said, oh, you're up for a chat. And he said, yeah, yeah, of course. And I phoned him. Oh, so there's a really important bit of information I forgot to tell you. When I first launched the podcast seven years ago, I'd read this book called It's Not How Good You Are, It's How Good You Want to Be by a guy called Paul Arden. And the subtitle in the book is the world's bestselling book by Paul Arden. And I thought, what a genius subtitle. So when I launched the podcast, I just, as a joke, I just said, the world's number one marketing podcast by John Evans, you see.

24:27Dan Taylor:So I'd said this like seven years ago. Anyway, so I phoned him and I said, dude, why'd you pick on me? And he said, well, I went to AI. I asked AI what the world's number one marketing podcast was. So I was like, oh, right. And they said yours. Yeah, yeah, yeah. Now I'm kind of getting to, I'm getting to make the claim real, you know, gradually. But it was just a bit of a, but I'd obviously, you know, train the AI. AI had searched out there and looked for the world's number one marketing podcast, and there it was. So talking about AI then, how much is AI changing search? I think everyone listening is going to have a perception that everything's changed, we're all going to AI.

25:03Dan Taylor:So how much are search volumes impacted, or are they still going up, or how many people are now using AI to search for things? So the volume of searches that we're seeing on Google is actually going up year over year, including the number of commercial queries. and a lot of that is driven by the innovations that we have made in AI. Not only the new AI-powered experiences that we have, but the expectations that consumers have about what they can ask technologies. As I mentioned a little bit earlier, people are realizing they can ask Google more and different types of questions, and we're also able to improve upon the types of answers that we give with things like AI overviews and AI mode.

25:46And so that's created new opportunities for better discovery, including of brands. And so we're really excited about this, what I would call a really expansionary moment for the entire industry. And so that's been great.

26:01Dan Taylor:But do you see the other AIs as competition to search in terms of people going to other AIs to look for answers to what they might have used Google for? Well, we don't see it as a zero-sum game. I think it's expansionary for overall discovery across the web. We see the overall searches going up. We see an opportunity for everybody in the game. I had a fascinating experience. Just before Christmas, I went to SEMrush's conference and met their CMO. And they did this report about digital brand visibility. They were basically tracking. And what I thought was fascinating is that actually Google is the main source of information for the LLMs.

26:38Dan Taylor:So even if you're searching on another AI, it's using Google to inform it. So from a brand point of view, you need to be showing up in Google as well as showing up in LLMs. Well, one of the things that's been interesting is we've talked about AI overviews and AI mode. And one of the questions that we get is, how do I optimize for these new AI-powered experiences? And our core SEO guidance has not changed. You still want to build for people-first content. You still want to think about, you know, multimodal capabilities, images, videos. You know, where are people discovering and hearing about your brand on, you know, review sites or content creators and social and all the rest?

27:17Like all the places where you would think about your brand presence, earned, owned, right, et cetera. All of those things still are true. Now, we have started to think about what are some things we can help for discovery on some of these new AI-powered surfaces. So, for example, we recently announced in Merchant Center the opportunity to add a bunch of new attributes like frequently asked questions about your brand or accessories that might go with products that you sell and make those available in Merchant Center so that AI can more easily pull those things into, things like AI overviews. But overall, the core guidance is going to be the same.

27:53Dan Taylor:Yeah. In fact, actually, my conclusion from the report they did was it actually just reinforced the importance of building a brand, getting other people to tell your story, getting loads of reviews, getting content out there that people engage with. And because that's what every LLM is learning from. And so it's important to get all those things in place. Talking about AI mode, then, are we going to see paid search appear? Will I be able to buy ads in AI mode? So today we're showing ads in AI mode to 2 billion users in about 200 countries and territories globally. And we are testing ads in AI mode just in the U.S.

28:32today. And so we are thinking about the right consumer experience in AI mode and where in the journey of using AI mode it makes sense to introduce advertising opportunities. And so taking half a step back, when advertisers are asking me, how do I take advantage of these opportunities, in addition to kind of that keyword construct around, you know, you can't take advantage of, you know, these new search opportunities when people are searching longer, more complex queries without tools like AI Max or search campaigns or Performance Max. But what you don't have to worry about is as these new AI-powered consumer experiences, whether it's AI overviews or AI mode, start to light up within search, your campaigns are automatically extensible into these new surfaces.

29:24So with all that aside, we're only testing ads in AI mode because we want to get that consumer experience right. And we're only testing them in the U.S. today. What I found most interesting as we look at that, so AI mode is a total reimagining of search, right? It's a place where people are diving deeper, it's conversational, it's longer. And what we're finding is it's not about the format per se, it's about where in that consumer's journey of discovery does it make sense to introduce a commercial opportunity. Like in what you call maybe traditional search, you might be closer to a commercial opportunity than not.

30:02And in AI mode, you might be more in a research place. And so, you know, slapping an ad for shoes if you're a runner, but really you just want to research, how do I get started in running? You know, I just started running last year. I'm at the 5K stage. I'm not at the marathon stage. But, like, I need to know, like, you know, what kind of clothing do I wear and do I need a different kind of nutrition set up? Don't try to sell me a bunch of gear the first time I'm starting to ask these questions. But as they get closer to a commercial opportunity, where are those ads going to show and where are ads going to be the right answer?

30:38And so that's really what we're learning. A good example of that is we announced at the National Retail Federation in January direct offers, which is when we have a good understanding of where a consumer is in that journey and a specific exclusive offer just for that consumer would be the right thing to get them across the line, we introduce a direct offer in a shopping ad and they can redeem it right there. Those are the kind of things that we're experimenting with right now to try and see if that's the right model for AI mode.

31:08Dan Taylor:Yeah, that's clever that. My experience with it is similar. I love the AI overview because it kind of teases a question I hadn't thought of. And I sort of read and go, oh, I'd like to know more. And then, of course, I'm straight into AI mode asking a question. So it's quite a good tease into going deeper with the product or it helps me refine what I really wanted to know. It kind of refined the question and you go back and forward. So it's very clever how the overview almost gets you into the AI mode almost without you realizing, and then suddenly I'm literally in AI mode, hadn't even realized.

31:41Yeah, it's one of the things that the consumer team is experimenting with to figure out when is a traditional search engine result. I keep saying traditional, but search is constantly changing. When is a regular search engine result the right answer? When is an AI overview the right answer? and when is moving someone into AI mode for deeper research going to be the right answer. And that'll change for every query, I think.

32:02Dan Taylor:And you must give brands a lot of advice about how to optimize their search performance. What are your two or three top tips to brands for getting the most out of search? The top thing that I give advice to brands is to start thinking about those opportunities to be discovered earlier in the journey. And I'll give you a couple of examples. So Aviva Insurance, that's a brand here in the UK, right? Insurance is a longer research play. right? It's financial services. It's not an impulse buy, but people search for it all the time. So with things like AI overviews, that's going to have a little bit more information.

32:35That's going to come up more often in the financial services category. But many financial brands aren't necessarily investing in these kind of more exploratory queries. And so Aviva invested in AI max research campaigns. And in doing so, we're being exposed to some of these longer, more exploratory queries and it generated 22 % more quotes for them as a result. So that's a good example. That's like the kind of the lowest hanging fruit for these new AI powered experiences is to make sure you're thinking about people are researching more with these AI powered experiences. And are you thinking about that as a search engine marketing strategy?

33:12Dan Taylor:That makes a ton of sense because you tend to think about a purchase way before you get into specifically asking details about availability, like learning with your genes about availability of a size and a color, you know, and a location. So influencing higher up the funnel makes a ton of sense, doesn't it? Yeah. Yeah. And again, I think with AI overviews, it's giving these brands the opportunity to get discovered earlier than before in search. So you can pick up the signals that somebody might be starting to think about a journey towards a purchase that you're in the category you're in. Yeah, exactly.

33:43Yeah.

33:44Dan Taylor:That's really smart. I was going to link actually, it might be a good time to ask you this, is what the link between kind of YouTube and Google is? Because obviously, you know, you own both, right? Google owns both. Google owns both, yeah. We wouldn't be chatting here, would we, if you owned both or you'd be flying your plane back. But what's the potential to bring those two things together from a brand point of view? Because one does kind of higher up the funnel, doesn't it? And one does more conversion. Well, I mean, we're finding that YouTube and or Google are involved in two-thirds of consumer purchase journeys when they discover a brand, a product, a service, or a retailer.

34:22And so we know that when it comes to discovery and decisions that YouTube and search play a role. Often they play complementary roles with each other. And so one of the things that we launched earlier this year, actually it was 2025 now because it's only February 2nd, is a metric called attributed brand searches, which allows brands to understand once they have an ad that has run on YouTube, let's say, did someone conduct a brand search as a result? And it really starts to help brands understand the connection between ad exposure on one surface and search volume later on. And then we can start to have a conversation about like, okay, these things are complementary and how do I optimize my campaigns in a way that makes sense and the role that it's playing across the funnel.

35:12Dan Taylor:That makes a lot of sense. Because I might be looking at kind of car review videos, like just looking at, just thinking about my next car, it'll be months, maybe even years before I'm actually in the market to buy one. And then suddenly I'll be Googling my nearest dealer to try and buy one. But it will be months and months of kind of like looking up different YouTube videos and reviews that would have gone on before. So I think connecting those sort of upper funnel and bottom of funnel things is critical. You talk about measurement there. How do you measure, you know, the impact of search? Because you're literally, you know, how do you ensure that what you're capturing is incremental to the brand and not just what someone might have searched for anyway?

35:50Well, I mean, we have a whole suite of products to help customers measure the thing that they want. I think increasingly customers are moving towards, and you hit this with your question, right? Like, I don't just want correlation. I want to understand causation. And so increasingly customers want to see, we're using programs like Conversion Lift, which is a study that you can run within your Google Ads campaign to understand how much lift did I get from the search campaigns that I'm running and which of those searches and conversions that happened were incremental. We also have GOX, which is a study that you can...

36:29It's a relatively time-tested technique to run a test here and a holdback experiment in a different geo to see how that's happening and then apply those best practices to optimization. So we're continuing to do that. We're also putting to work incrementality studies and working with third parties so they can test those results with their own sources of truth.

36:47Dan Taylor:Yeah, I mean, that sounds critical. That's always my question is how much, if I'm sponsoring a search, how much of it's capturing what was there already versus how much is getting in front of people that are new customers that wouldn't have been thinking about me prior to that point? Yeah. One of the bigger investments we made in 2025, we launched a first-party marketing mix model called Meridian, which allowed us to help inform customers' own internal business intelligence because a lot of our large customers build their own BIs to understand what's working versus not, but also third-party MMMs.

37:19There's an entire industry out there that does that to help understand how to look at paid search relative to organic search because a lot of the methodologies don't do that very well. And so we open-sourced that and made that available, and it's been a real breakout hit. And so that's a pretty big investment, and it's been really helpful to help people understand the incremental value of Google search.

37:39Dan Taylor:I think that's essential because I think any marketer needs to be able to justify how much they're spending. Yeah, I think one of the things I think that's been both really interesting and healthy, and again, we talked a little bit about the 20 years that I've been at Google. Again, coming from traditional media, which is really hard to measure, and we're like, wow, I can measure clicks? That's amazing. Well, we're so far from those vanity and proxy metrics. We're now at a point where it's like, I don't want to just measure clicks and visits or even leads. I want to measure which leads converted.

38:10I want to measure profit, not just conversions or e-commerce transactions. I want to measure incrementality, not just total sales. I want to be able to look back and make a marketing mix model investment based on a six-month look back. And so modern measurement has really become this view of like, I want to understand what's incremental. I want to be able to make a marketing mixed decision. So return on ad spend is required, but it is not sufficient today. You need all these other things in order to do it well. And that's become a real CMO's challenge today is how do you manage all of those things?

38:47I am hopeful in some places where we're investing is I think that AI and some agentic capabilities will start to be able to make sense of some of that faster than we do today. because today it's kind of, I mean, you're juggling a lot of reports and trying to make sense of it all. And I think that we have a real opportunity in an agentic space to get that time to insight to move a bit faster, and that's a big area of focus for us this year.

39:11Dan Taylor:That could be huge. I mean, I remember it's been a few years since I was a CMO, you know, spending the money, but I remember, you know, the MMMs would be months. I'd have to wait months and months and months, and then, you know, the insights were helpful. But by that point, I'd already spent$20 or$30 million, and everything had changed. So the ability to get real-time measurements and Intel that you can action straight away, I think, is huge. I sort of think about attribution helps your hands on keyboard today, right? Incrementality helps your media planner tomorrow. And MMMs help your CMO in two months, right?

39:46And how do we munge all that together into something that you can make decisions on a regular basis when you're sitting down with your CFO to talk about growth, not just expenses, right? Yeah, yeah.

39:59Dan Taylor:That's exactly the way to position it. Flipping to the other side of it, so we talked a lot about media optimization and what AI is enabling us to achieve. What about the creative side? What are the implications of all this advancement on the kind of media side? What are the implications for how we do the creative? I personally think, you know, 20 years at Google and a little bit more in media, okay, maybe a lot more, that we're really just scratching the surface on creative. We spent so many years optimizing audience selection and inventory and even how much we're willing to pay for our conversion, et cetera.

40:33I feel like we're just getting started with creative, and so those tools are really going to help us a lot. One of the things that I think I'm worried about is AI is the how, it's not the what. And I think a lot of the concern is like, oh, well, is AI just going to create AI slop creative? Right. And it could. Right. But much like I think YouTube empowered creators of all sizes, you know, put the power of a Hollywood studio on their mobile phone, essentially. It showed that great ideas, given the right tools, can be knockouts. Right. Great hits. I think the same is true with creative. And so the idea is actually going to matter more than it used to, not less.

41:23And so I think that's getting a little lost in the conversation. And so if you can productionize and improve your scale and reduce your budget on a great idea, that's an equalizer, right? And so your idea better count. I think that is getting lost a little bit. It is again.

41:44Dan Taylor:I'm strangely optimistic on AI. I was very worried about it to begin with, thinking everybody is going to rush to make lots more, 10 times more, 10 times cheaper, and we're just going to see a race to the bottom. But I think the fact that it must be the word of the year, AI slop, for some reason, like every other post is on AI slop. The fact that we can name that as the problem, and then we get over that and go, right, now let's focus on the idea. and the tools are just getting exponentially better. The quality of tools to be able to produce great creative in a short period of time. Funnily enough, actually, I don't know if you saw this, but the Pepsi Super Bowl ad, have you seen the Pepsi Super Bowl ad?

42:19No, not yet.

42:21Dan Taylor:It is hilarious. They've got the Coke polar bear doing a blind taste test of Pepsi versus Coke and preferring Pepsi. And then, you know, the polar bear is so shocked by the result that the polar bear goes to therapy. And then you see him walking the streets forlorn, looking at everyone drinking Pepsi and then meets another polar bear. They go to a concert and they appear on a kiss cam, a bit like the kind of Coldplay concert. And he's doing like this. Yeah, yeah. And he's like, oh no, I'm so embarrassed. It's very, very funny. Yeah. But the reason I tell you that is not, is, well, they've come up with a very good idea and they're playing with the whole history of the Pepsi Taste Challenge.

42:54Dan Taylor:But 10 hours later, Coke had come out with a response. 10 hours. Yeah. They were already out there with as well produced, very funny, you know the polar bear basically kind of claiming back the real taste of coke that kind of thing it was just brilliant and like 10 hours like even two years ago there's no way coke would would turn something like that around in 10 hours no way you know but they've done that but again it's like you said it's the idea because somebody came up with a really clever idea got it produced and it ai has enabled it to live and get out there on social media in a way it couldn't done before yeah one of my colleagues uh josh spanier who uh who does um marketing for google he's got a podcast too now, it's the Frontier CMO.

43:36He talks about this notion of a concept in the morning, ship it at night or that night. And we're really seeing that, right? Like there's a company called Loop Earplugs in Belgium and they created a series of video campaigns using VO3 and five days, 10 markets using VO3 and AI dubbing. And they were just out to market in five days because they had a great idea.

44:00Dan Taylor:Yeah. And your point is so spot on about like, if everyone's got those tools and everyone can do the same thing, the equalizer as you describe it is going to be how good is the idea? Yeah. That's where it comes back to, isn't it? It all comes in a way, it makes the idea stand out more because that's the only way of cutting through is to get a better idea. Yeah. I really believe that to be true. I have a similar analogy when it comes to AI and overall like media campaigns. And so, you know, one of the things I've been focusing a lot more over the last couple of years is data-driven strategy. And so, you know, if AI is doing much more of the media decisioning and inventory selectioning and figuring out, like, what's the right return on investment, right, a lot of that is being automated these days, right?

44:43Well, AI is only going to automate towards the outcomes that you tell it you want or the outcomes you tell it that you're getting. And so I'm investing so much of my team's time now in getting the data foundations and the data strategy, AI will steer to the outcomes that you want. It's not going to tell you the outcomes to get. Similar on the creative side, it's not going to come up with the best creative for you. It's going to scale your great ideas. Yeah, yeah.

45:07Dan Taylor:That resonates. I've spent so much time in the last year, 18 months, talking to big tech companies who are investing in AI. And off the record, they all tell me exactly the same thing, which is the technology is so advanced, the issue is the quality of data it's being trained on. That's right. And that literally is a universal thing I hear absolutely everywhere. So whoever's got the best quality data and can connect that data to other data sources is where the big leaps are going to happen. Yeah, it goes to the same point that you made earlier. If marketers all have the same tools, then the data that you bring to bear is going to be the differentiator.

45:39It's going to be the difference, isn't it? Yeah. It's true. And the creative. Exactly.

45:42Dan Taylor:Maybe a tough question to end on, but I've got a new newsletter out called The One Thing. Oh, okay. So I'd like to ask everyone. Please do, yeah, yeah. And anyone listening and watching, yeah, go to the uncensoredcmo.com. 20 years of Google, amazing time. Pick out one thing that you learned in that 20 years that you think could be most useful to people listening and watching. One thing that could be most useful after 20 years,

46:13don't wait.

46:17someone told me long ago when you're faced with a fork in the road choose the boulder path and i think in in times of technological change there's always more people that take a wait and see attitude instead of an experimentation and a move quickly approach and that never works out The wait and see approach never really works out. I think we're in a place, especially in the software and the media and technology industry, where the pace of innovation is so fast. The opportunity to experiment and innovate and iterate is there. So you may as well get started. I don't think there's enough opportunity to pivot and tweak along the way that waiting does not help you at all.

47:03Dan Taylor:I love that answer. And you learn so much from trying, don't you? The feedback loop you get from actually trying sets you up for success in the future, whereas the wait and see... It hasn't served anyone. You won't learn anything by waiting and seeing. That's a great answer. Love that. Dan, thank you very much, mate. Really appreciate you coming and chatting. And yeah, good to see you. Thanks. So I hope you enjoyed that episode of Uncensored CMO as much as I enjoyed making it. Now, by the way, I've got a new newsletter. So if you'd like to get my thoughts on the one thing that I take out from each episode every week, then do subscribe to the One Thing newsletter.

47:38Dan Taylor:I'd really appreciate it. Also, I have another podcast. Just launched Uncensored Renegades with the fabulous Cori Marchesotto. She is one of the world's best CMOs. She's an absolute rock star. Every week we pick one topic, spend 20 minutes trying to fix it. So check out that. It's in your feed, Uncensored Renegades. And finally, I want to give a huge thank you to my sponsor System One. They generously provide so much support for this podcast. It would not happen without them. So big thanks and lots of love to System One. I'll see you next time.