Inside Google’s global marketing operation

[00:00:00] Joshua Spanier:  One of the things I’ve come to realize is that marketers have been heroically good at delivering crappy bad advertising for decades and decades and decades. And the notion that AI slop was somehow going to overwhelm and replace the brilliant sparkling ads that we all see every day is just silly. And actually, what I’m seeing today is that AI models, Google’s Gemini models, Flow and Veo and Nano Banana is kind of raising the floor for everyone. I think everyone’s starting point is actually much better than where it has been, because honestly, our industry has made a lot of very bad, very unmemorable and, you know, boring ads for a very, very long time. Now, that said, yes, I think if you just turn on the machines and let the machines run by themselves, you could end up in a bad place where you just start optimizing creatively into a corner and looking like everyone else. But that’s where human judgment and taste and maturity comes in. The AI tools are that, tools. They are a starting point. They are a contributing part. They’re an input. They are not the final say. And when you hand over everything to the machine, it’s gonna end badly. It’s a partnership between humans and AI to actually create differentiation, and that’s what good marketing ultimately looks like.

[00:01:09] Jason Hemingway: Welcome to In Other Words, the podcast from Phrase, where we speak with business leaders shaping how organizations grow, adapt, and connect around the world. I’m Jason Hemingway, CMO at Phrase, and today’s guest is Joshua Spanier, VP of AI and Marketing Strategy at Google. Joshua spent, I don’t know, nearly 15 years at Google building some of the most complex marketing infrastructure on the planet. And he founded Google’s global measurement function and built their brilliant marketing AI lab, and he oversees pretty much integrated campaign planning across loads of Google products globally. He’s been really open about what’s gone wrong along the way. In ’24, he revealed that less than 7% of Google’s paid social was performing, and now he’s mentioned that that figure subsequently reached 70%. So, that’ll be an interesting thing that we talk about today. And also today, we’re going to find out what changed and how you measure marketing when you operate across very different markets. And actually, a subject that we always talk about: what happens to accountability as AI starts to do more of the execution. But without further ado, let’s get into it. Joshua, welcome to In Other Words.

[00:02:18] Joshua Spanier: Jason, thanks for having me. Happy to having, looking forward to this conversation.

[00:02:21] Jason Hemingway: Great to have you. So, as listeners will probably tell from the accents, you’re a fellow Brit, and you studied in the UK and started your career, I think, in London. But how does someone end up overseeing, you know, integrated campaign planning across Google’s products globally?

[00:02:37] Joshua Spanier: You know, the funny thing is it was all about love. I met the woman who would become my wife and the mother of my children at UC San Diego on exchange year for my British university. And we fell in love. And 30 years later, we’re still happily married, white picket fence and all that sort of stuff. But she’s American. I’m British. I moved to California 20-odd years or so ago. I’m now an American citizen as well as a British citizen. And, you know, when you live on the West Coast, when you live in San Francisco, you’re right by Silicon Valley. And you sort of stumble and fall and end up joining Google 15 years ago, as I did. And that’s how it all came about.

[00:03:13] Jason Hemingway: I mean, that’s a fantastic reason, isn’t it? That’s not your usual career path or journey. But it is interesting that you’ve kind of gone to one of the most famous brands in the world. And your perspective today would be fantastic. And so, let’s get into some more of the kind of meat of our discussion today. You recently raised the concern that AI could make advertising increasingly homogeneous, you know. From inside an organization already using AI extensively in marketing, are you seeing signs of that happening?

[00:03:41] Joshua Spanier: I reserve the right to change my opinion all the time, because it turns out both within this AI moment where everything is just in flux, no one knows anything. I think me least of all. So, I may have said that back in the day, you found some articles, some podcasts I was on, but you know, one of the things I’ve come to realize is that marketers have been heroically good at delivering crappy, bad advertising for decades and decades and decades. And the notion that AI slop was somehow going to overwhelm and replace the brilliant, sparkling ads that we all see every day is just silly. And actually, what I’m seeing today is that AI models – Google’s Gemini models, Flow and Veo, and Nano Banana is kind of raising the floor for everyone. I think everyone’s starting point is actually much better than where it has been. Because honestly, our industry has made a lot of very bad, very unmemorable and boring ads for a very, very long time. Now, that said, yes, I think if you just turn on the machines and let the machines run by themselves, you could end up in a bad place where you just start optimizing creatively into a corner and looking like everyone else. But that’s where human judgment and taste and maturity comes in. The AI tools are that – tools. They are a starting point. They are a contributing part. They’re an input. They are not the final say. And when you hand over everything to the machine, it’s going to end badly. It’s a partnership between humans and AI to actually create differentiation. And that’s what good marketing ultimately looks like.

[00:05:12] Jason Hemingway: You’re absolutely spot on because, you know, there was a lot of bad marketing in the first place, you know. And that’s a failure of human intervention, regardless of whether AI is doing it. It’s whether somebody’s got a bad idea or is executing on a poor advertising idea. So, thinking about that, you know, that idea of AI helping creativity, it becomes really complicated when one brand has to work across a very different market. As AI increases, or is a tool that you can use to increase the content you produce, and fingers crossed with the right oversight, you know, good quality stuff. How do you make sure that the brand still feels distinctive and appropriate to the different audiences in different markets?

[00:05:52] Joshua Spanier: I agree with everything you’re saying there, Jason. I’m just going to take out the word AI. How do you create differentiation for different audiences, different groups, different products across markets, whether it’s one large market like the U.S. with lots of subcultures or whether it’s across Europe with these 28 or 30 markets? You just do good marketing, right? You focus on your customer. You focus on their needs. You focus on their wants. You focus on your product. You understand your product to a deep level. And you look at the way culture, society, context, and usage is occurring. And the intersection of those three things leads to better insights, leads to more relevant content, leads to better placement. Obviously, you need measurement around it all. AI just accelerates that. And one of the things about AI is bad insights, bad creative, bad work will just be accelerated by AI. So, you can fail faster, but you can also succeed quicker with AI as well. When it comes to Google marketing, look, we’re a very large organization. We operate in 150 markets around the world. We have marketers all over the world. We are a large marketing organization. And so, there is a constant tension, there always has been, between the Global Mountain View, which is where I’m based in, sort of just south of San Francisco, the center of the company, and London and Singapore and Korea and every other market around the world. And that tension is pretty healthy. Our CMO has fundamentally believed in the power of local insights, that people closer to the market, closer to the customer, are going to be smarter about how to engage with them. So, we do a huge amount of work looking at all the frameworks that people typically look at, like GRL: global, regional, local. What do we set up globally? What is the sandbox? What are sort of the precepts and principles? And then, how do you regionally and then locally adapt and explore those things? And setting up that balance is complicated, right? And sometimes we get it wrong. A lot of times we get it right. But we generally feel that local markets know what’s best for them. Local markets don’t have the capacity around measurement, around resourcing, and some of the other things that you can supply centrally. So, freedom within a framework is another of our big sort of mantras. And then one of the other things we really focus on is accountability. It’s fine to give your French marketing team the autonomy and the authority to make their own decisions, but they have to report back the metrics, their impact, their growth, and they have to do so in a set of metrics which actually are understandable by the global marketing leadership team and by the business at large. You can’t have Germany and France saying, “We’re both great with different metrics.” So, you do need some standardization. But when everyone knows what those rules are or what that guidance is, then they’re free to go and deliver on it. And that really helps. So, constant tension. We generally trust our countries and our regions and our individuals in those markets to have more insight than me or anyone else sitting in Mountain View to tell the Australian marketing team how to do things.

[00:08:54] Jason Hemingway: Yeah, I think that’s brilliant. I think you have to give people flex in order to adapt things to their local markets. But let’s touch on that measurement idea because I know that’s something you’ve worked really hard on at Google and almost like the global measurement from scratch. And they scaled it to cover whatever spend you have, which is maybe billions. But what did you discover about conventional marketing measurement that this was kind of failing to tell you? Or how do you kind of put that all together?

[00:09:22] Joshua Spanier: How we got good at measurement, the flippant answer is we started 10 years ago, right? There is no one silver bullet. And even today, Google marketing is a very sophisticated measurement machine. You’d expect that. There is a phrase inside of Google: data trumps opinion, right? Although it’s funny how often opinion can get in the way of data, but that’s maybe a marketing thing. But we started 10 years ago, and we really started sort of 10, actually 11 years ago, because the company was doing great. You know, you think back to 2016, the shift to mobile was happening. We were spending a lot of money on behalf of the company. We were driving business outcomes. But there was a small part of me lying awake at night, and our CMO at the time and our leadership, where we were thinking: what do we actually know that’s worked? Like, if our CFO said, I’m giving you these giant marketing budgets. What am I getting for that money? We didn’t have a great answer. And we knew we didn’t have a great answer. We were uncomfortable. I literally laid awake at night thinking about that, and we were making some of the mistakes that a lot of marketing teams make, where you cherry-pick the result you want that you saw was positive, whether actually, you know, out of all the metrics available. So, we recognized that we needed to transform our ways of working. That meant hiring a true head of measurement analytics. That meant actually breaking some eggs to make an omelet. We had to invest in resources. You need data scientists. You need analysts. You need people who got statistics in their background and could actually help us. We couldn’t do this with just management consultancies or people with sort of, you know, average business degrees. We needed people with real expertise in data science and marketing science to do that. One of the things I learned along the way is my role before joining Google was at a creative agency in San Francisco called GoodBy Silverstein & Partners; it’s one of the storied, great boutique creative agencies, and one of the things I’d come to realize is a world-class analytics lead has incredible commonality with a world-class creative lead. And you would think there’d be opposite ends of a spectrum. One is all about creative, and emotion, and beauty, and one is all about science, and data, and mathematics, but it’s more of a circle where they meet back in the top of the circle very politely, very respectfully. These people are incredibly smart. They’re prima donnas. They are dealing with unknowns. They’re operating at the frontiers of marketing, right? They’re trying to create breakthrough creative and breakthrough measurement. And we don’t understand how any of this really works. Like neuroscientists who have spent their lives dedicated to understanding the brain tell us today they barely understand how the brain works. Yet we in marketing science and in the sort of creative side of the business are meant to be able to say, “Oh, this campaign worked to the nth degree.” It’s just none of it kind of makes sense. But these complementary disciplines of creative brilliance and sort of analytics brilliance requires a special type of management. So, one of the things I’ve really learned along the way, because I’m not a creative person in a pure creative, I’m not an analytics person. How do you manage and get the best out of those sorts of people? I think it’s a real challenge for CMOs. They speak different languages, they see the world in different ways, and then often, there isn’t a right answer. Often the instinct of the Creative Director is correct, and often the sort of mathematical instinct of this Data Scientist is correct as well. So, one of the big things I learned was, it’s not so much about the methodology or the creative; it’s about how you bring those teams together and how you get them to collaborate and actually fuss and come together at the frontier of what they’re doing. That was a really big learning for us. Third thing, and we’ll try and bring this all home in just a second, I mentioned standardization already. We had to create some sort of playbooks, right? One of the things that was really important was creating a playbook for, okay, we need common metrics, we need common standards, we need common ways of working, right? That obviously sort of makes sense. How you do that, though, is really important. And one of the things we recognize when you operate in 150 markets is not every market has the same data sets, not every market has the same maturity, not every market launches the product at the same time, and when you try and create a top-down, one-size-fits-all measurement methodology or one-size-fits-all creative strategy, it doesn’t work. So, what we ended up doing was creating what you might call it, and I think it was an Olympics year, so 2016, maybe. Was that the Olympics year? Either which way, the Olympics was on my mind. We created a bronze medal measurement methodology for brand, for performance. We moved on a little bit from that, which basically said, “If you’re doing any campaign, you’ve got to have the minimum viable measurement.” And it could be as simple as just clicks, right? You’ve got to measure clicks, right? You’ve got to have some signal of something. We never truly embrace clicks. But then we also said, okay, we need a – what’s silver? Once you go beyond the basic, what’s the silver measurement approach, and then what’s gold star, right? And over time, by creating this sort of pathway to what great looks like, if you’re an immature market, or you don’t have the right data set, or your budgets were intermittent, so you can’t do geo experiments in the same sustained way that other markets can. You can be acceptable in the eyes of our finance team and our overall global measurement team, because you’ve got something. And we’ve tried to move people over time. And finally, here, I guess the sneaky part is, once people know there is a pathway from bronze to silver to gold, eventually money starts flowing along that path as well, right? If you’re making a budget decision in your annual planning process a couple of years later, and there’s market A, which is using the gold standard, and market B, which is only using the silver standard, that is a signal that you can use. Not the final and only determinant signal, but our marketers were smart enough to see that, hey, if I get the check mark in the measurement piece, that’s going to give me more credibility when it comes to pitching for budget and actually allocating resources, which is obviously finite around the company in our annual marketing plan.

[00:15:16] Jason Hemingway: I love it. I think that’s, I mean, what a great model. It actually touched on a question I was going to ask you about how you decide whether a campaign is working globally rather than allowing strong results elsewhere in one market mask the weaker results. So, if you’re running a global campaign, do you accept kind of some of them will be bronze to silver, and then some markets will be gold? Is that kind of how you run it? How do you think about that?

[00:15:38] Joshua Spanier: We have made so many mistakes. We’ve also made so many learnings, and it’s constantly in flux, especially on the measurement side: the complexity of the measurement given data regulations and compliance rules, and just trying to understand the human psyche and all the rest of it and all the new channels that have emerged. So, we do accept that. But, you know, within Google, we have something called OKRs. They stand for Objectives and Key Reports. And our CEO, Sundar, sets 6 or 7 OKRs for the whole company. Everyone’s work cascades from there. And obviously, as you go down, your team OKRs and your individual OKRs sort of get narrower and more finite to your discipline area. But having that sort of central strategy is super helpful because when it comes to a decision on investment in a marketing campaign, it’s very clear to say the Italian team want to do some marketing for this because it’s going to deliver towards this OKR, which ladders up to this regional OKR, which ladders up to this global OKR. And if that cascade or upward march doesn’t seem clear, it’s an easy decision to say, well, we’re not going to invest in that because it’s not mapping to one of the OKRs that we’ve all agreed is actually something we’re doing. One more thing. I love the OKRs, the discipline they provide; Google, at least, they are very numerically based. You cannot have an OKR which says, we want to drive brand lift, or we want to sell more phones, or we want to drive understanding of Google Cloud. It’s tied to a metric. And those metrics are put within the OKR at the start of the year at every level. So, Sundar has his, marketing has theirs, the Cloud team has theirs specifically. And so, that sort of rigorous discipline. And then we have these evaluation sessions. We grade every OKR. One of the other sneaky things that Larry and Sergey did when they set the company up was they made it very, very clear that OKRs are stretch goals. So, if you have nothing but green in all your OKRs, you set them too easy. If you’re too successful, you get dinged. So, it’s like a real incentive to say, “Okay, I need two-thirds of my OKRs. Yeah, I’m close.” But otherwise people are like, oh, you made it too easy on yourself.

[00:17:41] Jason Hemingway: Yeah, I mean, it’s fantastic to hear Google do that. We do a lot on OKRs, and we did a lot over the last few years at Phrase. We did quite a bit. John Doerr – is he the guy who wrote the original sort of way of thinking about OKRs? I can’t remember.

[00:17:53] Joshua Spanier: It’s a real discipline that I think has to be built from the top,

[00:17:56] Jason Hemingway: right? And those quarterly check-ins, like you say, are very good. And I think one of the things people do sort of forget is that stretch. It’s got to be a stretch. And I think that’s what keeps people motivated and looking to do better. So, interesting to see that you do. And I’m sure that from your side, the complexity of running OKRs across the business are tenfold, I mean, maybe a hundredfold compared to us.

[00:18:16] Joshua Spanier: There is a whole team, an operations team, who helps sort of manage that. And we’re pretty well-oiled at doing it now. I like the discipline. And marketers and marketing generally, I’m going to speak broadly here, has a slightly fluffy reputation, right? We’re sort of, is it art? Is it science? You know, we just want to make beautiful things. And what I have found and seen in my 15 years at Google is that the marketers we hire, they want to drive business outcomes. They want to contribute to the success of a company as much as the engineers who are building Google Search or building YouTube or building Google Cloud, whatever it may be. And they’re hard-nosed, pragmatic business leaders. And so one of the things that marketers have to do, and where metrics and measurement over the long term become incredibly valuable, is they give you credibility. They give you the ability to go to your CFO and pitch for budget, as it were, and actually have the CFO not be questioning the fact that you’re asking. But like, what am I going to get for it? And when you can show that, it just kind of unlocks everything. And it goes from being a – I meet with a lot of CMOs in my own role. I run my podcast, Frontier CMO, that you can get at any good platform, including YouTube. There’s such a, potentially, crisis of confidence in marketing, and I think great CMOs are showing the way, especially in this specialist AI era, how marketing can be an engine of growth. But that’s built on measurement and credibility that they’re having impact in the world. We’re obsessed by incrementality, and Google Marketing has been for, you know, a decade. So, we do everything as much as possible with incrementality. And now we’ve moved on to sort of true outcome-based planning. So, incrementality is a part of that, but credibility gives you confidence, which conveys competence. I’m making this all up, by the way. And that just, you know, sets the stage and then you can attract more talent because people want to work on a team where you’re having an impact, and you can see that impact in the real world.

[00:20:05] Jason Hemingway: I’ve been doing marketing pretty much as long as you have. And I think that’s sort of changing. There’s much more focus on marketers having much more business acumen, being able to build good business cases for the investments they want to do. And that’s driven a lot by CFOs and teams saying, well, you know, there’s that whole thing at the moment, people talking about do more with less. And I’m a bit like, well, it’s not do more with less. It’s actually make everything you’re doing, you know, have a business case for it and make it effective. And I like that idea of incrementality that you talked about. That’s quite good because you can’t… One of the things I’ve noticed, in looking at your sort of stuff you’ve done before and obviously your poker player, and you mentioned it, reading incomplete information is quite an interesting thing, isn’t it? And making calls and, on measurement, kind of more specifically in what we’re talking about, what do you think people mostly get wrong on that kind of idea of measurement or accountability, let’s say, if you think from a CFO’s point of view?

[00:21:00] Joshua Spanier: Different companies with different levels of maturity make different mistakes. One of the things I think is challenging is if you’re not set up right. And you end up in a position where you have sort of opposition or mistrust in the C-suite; that creates real problems. So, part of our journey as a marketing organization has been to work closely with the finance group and to educate them. Not blindside, sort of, you know, browbeat them, but the Head of Marketing, one of the leaders of finance, sits in the leadership group of the CMO, right, comes to every meeting. There’s nothing sort of being hidden there. I think that’s a really important signal that we’re partners in this. So, that’s a super helpful thing. And I think oftentimes the finance team want to believe there is a single set answer and somehow the marketing team are either incompetent or hiding it or missing it. And it doesn’t help. There is an awful lot of smoke and mirrors and sort of people blowing smoke up the butt of everyone that somehow there is a secret answer to all this stuff. And we all know there isn’t. So, that fundamental partnership and trust and openness is not always there. More mature organizations get there, but that’s an important thing. Second, I think that measurement is incredibly valuable, but it can’t be definitive. And one of the things, one of the mistakes we made early on, is we started doing a lot of creative testing. And creative testing is some people like it; some people hate it, right? The challenge we saw is when you prescribe too closely, you must pass creative testing. What people started doing was making advertising and making marketing materials that passed the creative testing as the end result, not in terms of actually driving business outcomes in the real world. So, that is a mistake that we sort of made. The tail was wagging the dog. And we had to sort of course-correct and understand that people want a gold star. So, they wanted the gold star, which is you pass the creative testing. But if you slavishly follow that, you end up in weird places. So, that was another thing: the metrics are a signal, they’re a guide, and there needs to be some sort of flexibility and freedom within those because no one ever has the perfect single answer. I ate breakfast this morning. I didn’t know what I wanted for breakfast. And if you’d push me to say, what do you want, it’s like none of us knows that much about any of these small decisions, let alone these big decisions. I think there’s a third thing. I think in marketing, we get bored. I think we get bored of doing the stuff that works, the stuff which is straightforward, the stuff that we know is there. And so you often see it with new CMOs wanting to put their stamp and their sort of thumb on the scales and put their whole new campaign out there, changing agencies too often and too regularly. And, you know, we’re a little herd-like in marketing. We tend to follow fads. And that happens both at the measurement level. Oh, this new binomial way of doing research is all the rage, so we’re going to try and do that or whatever it may be. But actually, before you throw the baby out with the bathwater, before just because someone else seems to be doing something cool, take your time, understand it, and be structured. And the final point here is one of the things we’ve been very successful with in Google marketing is really establishing global and then sort of regional long-term learning agendas. We set out at the beginning of the year to say we would like to be able to answer these three questions. What is the plan to answer those three questions? And then what are the experiments? What are the countries? What are the places that we can do that? And sometimes those questions take multi year to answer, but by having an established plan, which we review, which we actually look at, which we understand, it gets us to avoid some of the vagaries of, oh, jump on that new thing, or let’s change everything, or let’s kill this or cancel that. And sometimes the long-term plan may be scuppered because of some major change. It’s like, okay, so how do we adjust it? But having that long-term learning agenda when it comes to measurement – the key questions you want to answer and then how you’re going to do it has been incredibly important for us in our incrementality journey, our attribution journey, our media mix modeling journey. Otherwise, it gets whiplash and sort of all over the place.

[00:25:03] Jason Hemingway: Yeah, and I think, look, we’re in one of those moments, aren’t we, with all the AI tools and things that you can do where you can jump onto the next fad that, you know, some of the underlying stuff isn’t there. You’re going to make mistakes. It’s going to compound the issues, like we said at the very beginning. So, on that sort of downer note, let’s line things up with a little mid-show moment that we ask all our guests. We ask everyone the same question. And because you’re, you know, you’ve done a global role for God knows how many years, what’s the most memorable language fail that you’ve ever come across?

[00:25:34] Joshua Spanier: At a company which, you know, literally translated the world’s populations, the world’s languages, you know, with Google Translate, it’s remarkable how many times back in the day we produced very, very weird ads by using automatic translation that was horribly, horribly wrong. And so, I know some of those were pretty embarrassing. Ultimately, I’m going to make this more personal, but I’m British. My parents speak French. My brothers speak French. I spent about 6 weeks every summer in France for the first sort of 15 years of my life on family vacations. And I learned French in school for 12 years. Today, I speak all of like 15 words of French. The failure is me. I am lost in translation when it comes to France. I can order, “Une bière, s’il vous plaît”, but that’s about it. So, that’s the embarrassing thing.

[00:26:20] Jason Hemingway: I think I’m in exactly the same boat as you. I think at school we have to do one language, don’t we, in the UK? And my French is exactly the same as yours, I think. But anyway, right, let’s go back into some of the things that you’ve done. In particular, one thing you published about how Google cracks social. And it was very honest, I think. You said that Google, you revealed that less than 7% of your paid social spend was performing. What made you comfortable about putting that number out in a public environment? I find that very interesting because not many people, like you said, marketers, can still hide some of the bad stuff. But you were very public and said, no, this is not right. We’re going to change, or we have changed.

[00:26:58] Joshua Spanier: Google Marketing is a sophisticated, smart organization with very talented people, right? Yay us. We have built trust and credibility with our stakeholders internally. So, when things go wrong, as they often do, we have the trust of our partners to, okay, so fix it. So, that’s what we do, right? In the same way that they have things that go wrong and they go, “Yeah, we got this wrong.” And then they go and fix it. And that’s part of the culture and DNA of the company. See a problem, fix it. Take accountability and ownership for it. So, me publicly stating, “We’re not working here. This isn’t quite right.” No one has a problem with that because they trust that we’re on it and we’re going to fix it and find a way. Now, when we first discovered that our social campaigns were performing that badly, I didn’t immediately rush out of the world and say, “Look at us, we’re terrible.” I did wait a bit because I wanted to find solutions and stuff to actually get there and know we’re on the right trajectory and where things were picking up. So, I’m not being completely selfless here. I’ll get to some of the things we did in just a second. But one of the nice things about my job and being on this podcast and hosting my own podcast is this, you know, desire to actually believe and love the marketing industry and the colleagues and peers within it. And we’ve always had this remit of: we want marketing to be better. I mean, that’s probably good for Google if people are better marketers and Google sells marketing solutions. So, that makes some sense. But it’s also bigger than that. Too often, I feel like in marketing, we are quick to thump our own chest and go to some conference and say how amazing we are. And we talk about what we’ve done and how amazing it is. What I’m interested in, what I know so many of Google’s customers and other marketers are to yourself, they’re interested in how things are done. What they want is not case studies; they want how studies. And so part of my job, part of our remit, is: let’s talk about the things that have worked and haven’t worked, and let’s actually publish them. We have Think with Google, thinkwithgoogle.com – an amazing resource full of incredible marketing, insight, and learning. Because if people get better at marketing, then the category, the industry, the discipline gets better. So, we just talk about it. It’s fun. It comes to social overall. We were making some of the classic mistakes, you know, a lot of people are making. We weren’t building for platforms, right? We were still too much in the, hey, we built an asset, a video asset, and, you know, for either YouTube straight or television. And then we were just porting it over to a social platform or a social context. And, you know, understandably, it wasn’t working. This is such a blindingly obvious thing. But we were being, cutting corners and not doing it in particularly smart ways. As we got more sophisticated, we started learning and really working with the platforms. One of the things I think is incredibly valuable and important is to kind of take a collaborative posture with your sales teams, meaning Google marketing gets supported by the YouTube sales team, right? The YouTube sales team actually pitches to, and, you know, helps Google marketing. Who knows the platform better than the sales team at YouTube? YouTube, right? So, one of the things we did was we went to the YouTube sales team, and we said, how do we make YouTube work better? What are we doing wrong? Help us learn from all your customers. And actually it’s true across sales teams and organizations. And there is this sort of general tendency for media teams and for creative teams to kind of be suspicious of sales teams because they’re always pitching, right? They’re always selling, and it feels disingenuous. However, these tools, these platforms, are understood better by those sales teams. They see the best of it and the worst of it. And so we literally said to the YouTube sales team, if we start running an ad which is not going to perform and you know it’s not going to perform, we want to stop the campaign. Stop the campaign, stop spending the money and tell us. And it’s okay that you stop it because we need to actually take accountability and not, you know, just say, just run it even though you know it’s not going to work. One of the really important things we’ve learned, which has helped us get from 7% to 70% effectiveness now in our social campaigns, is something we call a two-voice model. So, we all know about YouTube creators and how powerful and influential creators are. And what we have seen is when you run ads with creators, you know, those ads made by creators in their own voice to their own audiences work very, very well. We’ve also seen that when you run ads just featuring Google products, like a standard product ad on YouTube, they work as well. The combination of creator ads and product ads, works exponentially better. So, we actually saw this, and we tested this; we measured this in multiple markets around the world. And it takes time to do these experiments, make the right creative, but finding the ways to crack social, as we call it. We had a whole cracking social work stream, but this two-voice model of work with creators, run your own ads, the combination of the two in the right mix and the right markets, actually drove significant breakthroughs in performance and success. You don’t just veer off into, we’re just going to be all creators all the time. A year or so ago, the new incoming CEO, I believe, of Unilever basically announced that they were going to be all social all the time. This was the future. And then I ended up having – I interviewed Mark Ritson for my Frontier CMO podcast, and we had a good conversation about how careering that far in one direction is probably a mistake. It absolutely is. And I think the CEO is trying to make a point more than say everything is that. But I mentioned earlier, marketers are a little herd-like. We tend to sort of get our marching orders, and then we just like jump off the cliff together in one direction. Nothing is that straightforward. So, along this social transformation journey, understand and acknowledge we had a problem, try and root-analyze what’s going on, work with the platforms to truly hear from them, and learn from the people who know more stuff than we do. Test and experiment our way through multiple variants of multiple variations of doing things. Find the winners and scale that and create those best practices, which then become the sort of freedom within a framework. So, every market in the world knows, oh, if I do these 4 or 5 things in this general way, I’ll see better business outcomes. And then trust the verify, keep testing that stuff over time to make sure that as consumer patterns change, as the platforms change, we’re not stuck doing the same thing we agreed three years ago, but not adapting to the platforms as they evolve themselves.

[00:33:01] Jason Hemingway: There’s a compounding effect when you put two different things together, and the kind of World of Ritson, which is bothism, isn’t it? Which you talked about: you could do the long-term stuff and the short-term stuff, couple them together, and you’re actually getting good business results out of the two. But also talking to the sales teams, the ads you’d expect. I always think a good salesperson is somebody who tells you something that you didn’t know about what they’re talking about, and they can be your expert in the room. And you said it about the YouTube guys. It’s very interesting stuff.

[00:33:27] Joshua Spanier: It is important that the sales teams recognize that end. Always be selling is kind of the mantra, right? But actually, the best relationships, whether it’s with your agency, your vendors, your sales teams, are built over the long term. The human connection and the trust that comes from that. And you know, sometimes that means taking, you know, one step backwards. Like, hey, let’s pause this. It’s not getting you where you need to get to. But here’s what I think if you do this, this, and this. And so those have always been the most fruitful business partnerships. It’s funny how in all this talk of AI and transformation and all that blahdy blahdy blah, the human part of this is still so unbelievably important, just critical. You’re not going to succeed just using the machines and thinking you can, and just using the data. There is an art and a science, and there is a human piece of this, which is just incredibly powerful still, and I think always will be for marketers.

[00:34:17] Jason Hemingway: You’ve said, as lots of people said, and me included, the CMO role is in kind of revolution mode. You know, AI takes on much more execution in particular. But what do you sort of mean by that, because we talked a lot about humans being in the kind of oversight and having accountability. But what do you mean by the kind of execution? What are you seeing when you’re talking to people or indeed at Google? And AI takes on more execution. What do you mean by more execution?

[00:34:41] Joshua Spanier: So, this is a really important question. And I’ve met with over 200 marketing leaders this year, listening and talking about AI transformation through the lens of what we’re doing inside of Google and then hearing what these other companies are doing. And there’s probably 3 things which I think are most important. Only one of them is a thing you’ve referenced, but I’m going to talk about it. The first is, having met those 200 marketing leaders, and Jason, you probably feel the same way, I can tell within 60 seconds if they are walking the walk or just talking the talk when it comes to AI. There is a level of personal accountability and personal leadership, which if you are not investing your time in using these tools, in speaking with your teams, in listening to podcasts, I have 5 hours a week booked on my calendar, reserved, blocked off for me to actually just play around with these tools. It’s a critical part of what you’re doing. If you’re a senior marketing leader or any marketer at this point, if you are relying on other people to walk you through this AI transformation, you’re falling behind. So, that’s first, right? Second, and this is where it gets, I think, really interesting. The first era of AI transformation, I think, kind of, really happened within silos. And it was really driven by economics. Do more for less. Your CFO wants more done for less. So, your social team, your events team, your digital team, your media team, your creative team, basically in their silo, scrambled around, looked for where can we take our current way of working, and can we actually make it more efficient and save 10%, 20%, 30% by applying some AI tools and technology. Speed us up, save money, right? That gets you a certain distance. And yes, for a couple of years, you could save some money for your CFO. But the thing about AI, and the thing which is really transformative, is not thinking about AI just in terms of how it can save me money and make my current way of working better. Actually, how can I reinvent my ways of working? And what I’m calling this, for want of a better word, is sort of connected AI or brand systems architecture. I’m going to use an analogy or metaphor. I get confused between those two words, but I think it’s an analogy. One of the things that I’m really proud of, even though I have nothing to do with it, but from Google, is the work my company has done to help doctors in cancer screening and other medical scan readings. They built, initially, machine learning models, now AI-powered models, which can read a MRI scan or a cancer scan and look for anomalies and potential cancers. And when you train a model on 10 scans and then 100 scans and then 100,000 scans and then a million scans, they can become incredibly accurate and really incredibly useful. Because as good as our doctors and medical communities are around the world, they just don’t have the persistent memory of being able to look at a million medical scans and spot for every possible thing. That persistent memory built into those models means the junior doctor today is incredibly more powered by these AI and machine learning capabilities to help and treat patients. This is happening to marketing as well. It used to be when someone left your team, you lost all their knowledge, you lost all their best practices, you lost their ways of working. What we are seeing today is you can start applying AI to your systems, your processes, your protocols, and smart marketers are connecting it through everything. They’re moving from those silos just being efficient into a connected AI system. So, this brand system architecture is for the CMO to look at every function, every discipline, every job, every task, work out where AI can play a contributing role, and then how you modularly fit these things together to create this persistent memory. The starting point of your marketing programs isn’t we’re just putting our finger in the air and hoping we’re doing well. It’s actually grounded in your data, your best-performing campaigns, your best-performing ideas, everything that you’ve learned, and allows us humans to add our level of sophistication, intelligence, and taste on top in a way which is really powerful. This isn’t coding, or vibe coding, or, you know, true, sort of, architectural sort of, you know, soldering of computer points together. But this notion of marketing’s role and the leadership role to actually connect these different systems in a connected, seamless way is really, really interesting and really, really exciting. And I think the next sort of wave of where AI transformation is going, being able to see the big picture and how you actually connect all these systems and tools, but with a persistent memory layer. That’s really important. Third and final thing, just because I’m on a roll here, is personal responsibility as a leader to invest your time and learn these things. This brand systems architecture of connected AI is sort of a framing and a way of working. The third thing is really about your teams. Another defining trait of the early part of this AI transformation for every company around the world has been fear. Am I going to lose my job? Uncertainty. What is even my job? And what we’re seeing inside of Google marketing is when you can get the light bulb to go off, when you can flip the switch, when you can transition people from thinking from a fear-based, “What is AI going to do to me?”, to one of, “What can I do with AI?” That changes the game. It changes the mindset. And that gap between people’s fear versus the opportunity is the green space for CMOs and leaders to create for their teams, but it’s actually for marketing to take an even more central place inside of organizations to actually lead the industry and lead their businesses. We’ve never been constrained by ideas in marketing. We’ve been constrained by resources. In an AI-powered and enabled world, our ability to do more, to have more influence, to drive more connection and engagement with customers and across our entire organizations is actually really incredible. So, I kind of think we’re going to enter. We are entering potentially a golden age of marketing where the CMO could be a true Chief Growth Officer as well as the Chief Marketing Officer and drive growth for the whole organization. That’s what’s really exciting for me in all this sort of opportunity. It’s like things are great and they’re going to get better. There’s disruption, but this is the opportunity we all need to be embracing. Let’s become Chief Growth Officers, not just Chief Marketing Officers.

[00:40:36] Jason Hemingway: Personal experience of this. You know, I’m in a tech business. We’ve gone through that moment of fear as everyone has. It’s hilarious. You’ll talk about a connected system, a branded system. We’re calling it the AI marketing operating system that we’re building. So, you’ve got to think as a system, not as individuals. We were calling them teammates, all the things that we were building. But actually, you need a connected infrastructure. The memory layer, I think, is fantastically important because you’ve got to be able to make it some kind of self-learning and have a system of record almost built in. And then you’ve got the harness layer, the orchestration layer of all these things. But ultimately, at the same time, you’ve got to have people making good decisions. You can have all the operating systems. Anyone can operate an operating system, but you can still make bad decisions with it. So, we’ve got that. And I think the one thing that we found really interesting was you called it moments where people kind of recognize something. We called them all: everybody needs to discover their own ‘wow’ moment. And you give them the tools to go away and do it. And also what you said, you said it from a leadership perspective, but I think it’s important to give people time to go and experiment and the tools. Okay, we’ve got everyone in the business Claude; you’ve got time every week that you can dedicate to it, and we want to, you know, we want to see those things, and we make a big song and dance about it as a business. And we talk about it across the entire business. So, I think that’s interesting. And if you embrace it, and particularly in a smaller business like ours, we’re not Google, but you know, we had all the jobs to be done of everyone at marketing. And then I would kind of think, you know, if we have 10 marketers, what happens if I could scale this to act as if we had 100 marketers in our team? And that’s the thing that you kind of want, I think. So, it is exciting. I think lots of people, in the right way, I think you can make it exciting. So, that, I mean, fantastic. I’m going to, you know, we’re coming relatively close to the end now. It’s been a fascinating conversation. I just want to tap into that kind of headquarters versus local markets a little bit more because you operate in markets where language and culture and consumer behavior, you said they look nothing like each other, and you need those regional teams. But what is the point to which, or when do you decentralize? And is there anything you lose when you do that? Or how do you maintain? You talked about tension, but tension can quickly get out of hand, right? So, how do you kind of maintain? Is it communication? Is it looping everyone in? Is it the OKRs? How do you kind of keep that going?

[00:42:56] Joshua Spanier: I don’t have a perfect answer to this. If you read Harvard Business Review, you know, every other issue is like, you know, how to manage an organization and the pendulum’s always swinging. What’s worked for us is having a consensus culture. What’s worked for us is having a culture where people have to explain what they’re doing and why they’re doing it, and get buy-in. That’s actually been powerful for us because it sort of forces conversation and debate. But a large part of what I’ve done 15 years inside of Google is just try and force people to collaborate, to actually have the – it’s very simple, but you know, I’m running a global team. I was running a global team. So, we would do a status call with the North American team and the EMEA team, right? And so, we’d do that every two weeks, right? And it would be at 9 in the morning, which is 9 in the morning East Coast time – yeah, mid-afternoon in Europe; that’s no good for the APAC team, right? So, we’d have a separate call with the APAC team, and we’d do that at 9 O’clock at night. And I would ask my teams to rotate doing early morning and late night because you’re trying to force these teams to actually talk to each other. Otherwise, you end up in a situation where the Asian team and the European team and the North American team or the Americas team don’t talk. So, just simple, you know, things like that. And it’s a little inconvenient when you’re on a global team, but it’s kind of part of the upside is you get to learn. You get to learn what people are doing in China or in Singapore, in Korea, which is fascinating if you’re sitting in Italy or in London and vice versa, right? And that’s, it’s so funny, we’ve talked a lot about this, right, about the human part of this. You need your people to see the value of their peers and their colleagues and that learning mindset. So, I want to go back to your original question. There is no right answer on how you make a decision, right? You have to build your infrastructure and your systems, your protocols. But the thing above it all is we want to do great marketing. What’s the blocks and barriers to that? Are you able to do that? And if you’re not, then leadership is not clearing the path enough. So, everything we’re about is, how do we get these incredibly talented, smart, wonderful people to do the best work of their career? Every level we can throw, we try and do that. And those little things, like setting up global meetings or regional meetings at slightly odd times, sometimes that’s necessary, but it’s worth it.

[00:45:11] Jason Hemingway: We’re very close to the end, and it’s been brilliant to have you on. I’ve just got 3 quick-fire questions for you to test you out right at the end. The hardest thing about scaling globally – creativity, trust, or measurement, or something else?

[00:45:25] Joshua Spanier: Trust has to be earned, and it’s so quickly lost. And, you know, us humans are dumb. We lose trust very, very quickly because we get frustrated and misinterpret things. But if you don’t have trust, you’ve got nothing. So, that’s true of everything in life.

[00:45:39] Jason Hemingway: What does language intelligence mean to you?

[00:45:41] Joshua Spanier: I have no idea. Should I know what language intelligence is? Is this a buzzword I’m unaware of?

[00:45:45] Jason Hemingway: No. It’s just for you to think of, you know, what would that mean if you were language intelligent to you?

[00:45:49] Joshua Spanier: I’m going to be very literal. The ability to communicate clearly and concisely.

[00:45:54] Jason Hemingway: Well said in the end. I mean, you had no idea, but you got it. And the best poker lesson that applies to marketing.

[00:46:00] Joshua Spanier: I’m just going to give you the best one, which applies to life, right? So, there’s this thing in poker called the theory of implicit collusion. Implicit collusion. When things aren’t going great at the poker table, it feels like the whole table is ganging up on you, right? That everyone is trying to get you and take your money. And you think there’s something weird going on. But it’s not. What’s really just happening is everyone is playing their game, and you’re assuming they’re all against you. And actually, what you need to do is to not think that everyone is out to get you. You need to actually start thinking about how to play your game, how to focus on what you can do, what you can control, and ignore the sort of seeming like everyone is out to get you because they’re not. That’s not the way the world works. They’re not in cahoots. You are seeing shadows in the dark. It’s not there. Focus on what you control. Make your smartest decision. And if that’s not the right decision, then that’s okay. You did the best you could

[00:46:56] Jason Hemingway: in that circumstance.

[00:46:57] Jason Hemingway: And what a brilliantly positive note to end the podcast on, Joshua. Thank you so much for opening the door to Google and telling us your career history and everything that kind of matters to you and the Google team. It’s brilliant. And I, for one, love the Frontier CMO podcast and all the things you’re doing to kind of help change the marketing industry and actually business in general. Thanks for joining us, and I hope to see you again soon.

[00:47:20] Joshua Spanier: Thanks, Jason. It’s been a pleasure.

[00:47:21] Jason Hemingway: Joshua, thanks again. You gave us what I can only describe as a rare insider’s view of how marketing at one of the world’s most recognizable brands actually works. And there’s some brilliant lessons in this conversation for anyone scaling brands across those different markets and indeed running a marketing organization for today and the future. As I said in the podcast, I’d also encourage everyone to check out Frontier CMO if you haven’t already and all the good things that Google are doing.

[00:47:46] Jason Hemingway: That’s it for another episode of In Other Words, a podcast from Phrase. I’ve been your host, Jason Hemingway, and a big thank you again to Joshua Spanier for joining us today. If this got you thinking about how your organization measures and manages content across languages and markets, come find us at Phrase. For more conversations on leadership growth and how businesses connect with customers cross markets, do subscribe to In Other Words, on YouTube, Spotify, Apple Podcasts, or indeed your preferred podcast platform.