← The Digital Marketing Podcast
The Digital Marketing Podcast
The Meta CMO Interview - Digital Marketing Measurement, AI & Why the Basics Still Matter
00:35:03
In this special episode of The Digital Marketing Podcast, Daniel Rowles sits down with Alex Schultz, Meta’s Chief Marketing Officer and Vice President of Analytics, to unpack the future of digital marketing, and why the fundamentals still matter...
Loading summary
Transcript71 lines
- [00:00]
A
Welcome back to the Digital Marketing Podcast brought to you by targetinternet.com and in this episode, we have an interview with the chief marketing officer of Meta, Alex Schultz. So we get at least 20 emails a day suggesting guests for the podcast, some of whom are fantastic. You would have heard them. Some not so great, but a lot of them are authors who are approaching the launch date of a book. I want to share some of their insights, and wherever we can, we take a look at the book, see if we think they can actually provide some value. And one of the first things you end up doing is looking at the book cover and you have a quick look at recommendations. And as an author myself, I know it can be really hard to get these recommendations because you've only just finished the book and you need someone to read it urgently so they can give you a quote so you can kind of hit your publishing deadline. And therefore you end up asking your close colleagues and friends. Well, quite often for me, that might be me asking Kieran or something like, well, this particular author, who did he ask? Well, he's got quotes from Mark Zuckerberg, Facebook fame, Sam Altman from OpenAI, and then even Matthew Vaughan, the director of Layer Cake, one of my favorite movies. So who on earth was this person? We were pretty excited. So Alex Schultz is the longtime CMO of Meta and VP of that analytics, and he's all around an amazing guy you'll hear more from in the interview as well. But he's written what I think is going to become the definitive book about digital marketing measurement. And the book's called Click Here. There's loads of useful insights in here. It was a really fun interview, a very open interview as well. So sit back and enjoy the interview with Alex Schultz, CMO of Meta. Okay. I am here with Alex. Alex, tell us a little bit about the book and why you decided to write it. And I should say before we get going as well, this book is epic. I mean, I, I, we get a lot of invites to speak to authors, and this is the first one that I'm actually going to put onto a course text for my students at Imperial College, because listeners will know that I'm senior lecturer there. So no one else. Also has Sam Altman, Mark Zuckerberg, Daniel Ek from Spotify, and my favorite, Matthew Vaughan giving recommendations to the book as well. So I have to say congratulations. So congratulations on the book. That's what it's about.
- [02:19]
B
Yeah.
- [02:20]
A
That's why you decided to write it?
- [02:21]
B
Matthew Vaughan is. He shot our super bowl ad and I swear he didn't just endorse it because. Because I paid. He says he's a good book, genuinely. No, he's a great, He's a great lad. You know, there's two, there's two reasons why I did the book and I think, like, you're probably. One of them is like very banal, but it's true. Which is there was a gap in the market.
- [02:38]
A
Yeah.
- [02:39]
B
And I keep being asked for a book on digital marketing that kind of tells you a high level overview of every area and gets you. And gets you going and I just don't have one. There's a scattering. Like I tell people, viral loop, I tell people scoring points, the Tesco Club card story. But there's no one book. And so there's this gap. And I talked to Hachette's CEO, he agreed. They actually published Ogilvy from one of their, one of their imprints and so went for it and went for the gap in the market because I want a book I can give to people. So that's the easy answer.
- [03:10]
A
Nice.
- [03:11]
B
The less easy answer is I really believe in what we do. And there are so many people pooping on marketing in general. Right. Like marketing in general. You're trying to convince people to buy things they don't want with money they don't have to quote Fight Club. And the digital marketing realm in particular. You've got surveillance capital, you've got everyone kind of going on about this, you know, and we adhere to all laws that are put out there, but we get all the laws put in that are not helping it. And, and bluntly, I just think they're stopping people from getting personalized ads that make their lives better. Give them products for free and connect them to products that they actually like by showing them slightly more relevant ads for trainers. And on the other side, create business growth so people are employed so businesses can grow, so the economy can grow, so there can be taxes which funds so much of what we have as, you know, institutions in the uk. And so I also wanted to just be like, online marketing is good, it grows the economy and I'm proud of it. And someone needs to say something positive. So I wanted to say something positive.
- [04:12]
A
Love it. I mean, on that first point as well, I completely agree. There is a massive gap in the market because we spend a whole time. So Kieran, my co host, we spend the whole time going, people need to get back to basics, they need to get back to fundamentals. I don't think people know what those fundamentals are And I think this book is beautiful at going through and just giving that kind of right mindset and going through and saying, look, this is need to think about it, get back to those real core kind of pieces as well. I also think that thing of being proud of digital marketing is a really key one. If you look at how the ease of launching new businesses now, my, my master's students come into me and are kind of terrified on two fronts at the moment. One, constant horror stories. Digital is bad. You know, all this kind of stuff. There's that side of it. There's also they're terrified because of AI. And we'll get into that later on. And I keep saying to them, look, and it's a quote that came from you as well, was that the tools change, but actually the fundamentals stay the same.
- [04:58]
B
Yes.
- [04:59]
A
And I thought that was a really interesting. I picked out from the book as well. So. But let's get into some key areas in the book as well. And I think this is lovely one to start with, that difference between goals and metrics and what that means to people. Because this is kind of fundamental.
- [05:12]
B
Yeah, I mean the back to basics thing, when I was reading my audiobook, which I know you've gone through that torture as well.
- [05:18]
A
Exactly.
- [05:19]
B
I was just going, oh my gosh, I need to go and spend time with my team and tell them to do these things because like there's all these basics that you just need to do. And one of the very basic things is actually be super clear what your goal is and realize that a metric is not a goal. So, you know, our historic goal was connect the world online. Like people know that through Facebook. Our metric was monthly active users. And the thing is, a metric can be gamed. And so many people like turn a metric into a goal and then the metric doesn't actually do what they want as a company and they're really upset about it and that's because they forgot their goal. And so the example I use is if I get you to click on an email and you visit the site for 10 seconds, have I really connected the world online or did I just get a 10 second session? So we set the goal, connect the world online. We then have the metric monthly active users, and then we have a bunch of guardrails against it. So percent weekly, percent daily, all of those kind of things, lifetime value. To make sure we're not bringing minimally active users, we're bringing actual monthly active user users, full stop. And I got so many examples of it. The other one's eBay, where we had, we wanted the. For the acquisition team, the North Star was get everyone to use ebay. Right? I mean, pretty simple. The goal we used to set was confirmed registered users. But the thing is you'd get affiliates out there, incentive affiliates, incentivizing people to sign up for ebay and confirm. You get search affiliates sending people to the registration page and not the search landing page.
- [06:52]
A
Right.
- [06:53]
B
Which got registrations and confirmations. But none of those users activated and became long term active users. We changed the metric to aclu. We kept the goal, get everyone to use ebay, which is activated, confirmed registered users. All of those incentive affiliates basically went away. And the search affiliates changed their landing page to be the search results page because that's what got someone to bid by activate. So the, the goal and the metric have to be separate because you create metrics that never perfectly describe the goal.
- [07:25]
A
I love the fact that when you were going through and reading your own book out, you started thinking, oh, we really need to refocus on this. I mean, reading your book had the same effect on me as well. Like we preach this stuff all the time and we run a business and we're successful, but you look at it and go, yeah, we're not doing that. We should be doing more of that. And there's a gap there and we need to go. And I always think it's a nice thing because marketing is inherently challenging. Right? Because I think there's that whole piece that if it was just something you could repeat, there'd be a guy that you could buy for 99 online and everyone would buy it and then it wouldn't work anymore anyway. Yeah, so is that it's an art and it's a science. And I think there's that, that, that kind of key piece. And that brings me on to this thing that came out again in the end of the book is that, that incrementality piece. So if incrementality, everything, that's the key principle. What does that mean in practice for people?
- [08:05]
B
Yeah, I mean fundamentally the thing for me is I just, I really want what I do in life to matter, to make a difference to actually the fact I showed up, things are different because I showed up. I didn't just collect a paycheck. That's really just how I think about the world. And I think too many marketers, we. Ogleby has the useful idiot in his book of the person who, you know, doesn't get modern marketing. Well, I've got a useful idiot, essentially, which is someone who uses post click tracking and never looks at the incrementality. So what it means to me is you should just do testing and depending on the channel, right? Search doesn't let you do lift studies and conversion tracking in that way. So search, you have to do matched market tests, you have to do pre post tests. Like that's kind of what you have to do with search Social. Most of the social companies allow you to do holdouts because all their users are logged in, they've got user IDs. And so we offer lift studies, but all the other guys offer various different things. So there you should leverage those and you should actually see what is the uplift. So yes, it's really simple. It's like do a test, see if you got more results in the test group than the control group. And that dates right back to Claude C. Hopkins a hundred years ago with scientific advertising. This is not a new principle in marketing.
- [09:20]
A
I think it's interesting because the more and more talk there is about a no click environment and there's more challenges of attribution, modeling, all those kind of things, kind of getting the basics right becomes more and more important. So just let's come back to that. Just why is it so important to get those basics right?
- [09:34]
B
I mean the fraud is one example. Fraud and waste. So, you know, from a fraud perspective, the worst examples I saw back in my day at ebay, well actually Uber's a really good example, right? So Kevin Frisch was head of marketing at Uber and they ended up suing their banner agency, their dsp. I think it was because they realized when they turned off hundreds of millions of spend, the conversions didn't drop for Uber. And actually the way they found it, which funnily enough, we found something similar, the way they found it was their ads were showing up on the wrong sites. So they were trying to avoid controversy. They weren't trying to be on Breitbart at the time. And their ads kept showing up on Breitbart. And what was actually happening was that they're either a subcontractor of their agency or their agency was putting ads out in places where they weren't supposed to be put out so they could get extra impressions. But the worst thing that was happening is they were putting them out below the fold and getting no results and just scattering cookies. And by scattering cookies saying that they were driving results. And so when he turned it off, he saw no drop in top line growth. Now I recently did a similar experiment in India where I turned off all of our Marketing for Facebook, the app in India, and during the period I turned it off, Facebook shrank. It had a topline impact on Facebook overall to turn off all of our marketing. And so that's the difference is like if Uber at that stage they could turn off the advertising and nothing changed. And in India, we turned off the advertising and Facebook shrank. That tells you the difference between actually having incremental impact and not for me, that's all you need. That's like what you need for your job. But actually that matters for finance, that matters for the CEO, that matters for the board. Like that India result's only a couple of years old and it's a legendary result inside the company. Every time I go to the board, the CEO, every time I go to the cfo, I'm able to point to that and say, I'm not lying to you when I say I drive top line results for the company.
- [11:25]
A
Right. And I think that that proof of concept is actually said, let's do it. Let's do an experiment of some type to actually prove this is working. Let's scale from that kind of point of view. I mean, there's a couple of quotes I've just put out the book here that I thought were relevant for the audience, which is start with a marketing funnel. And I love the fact that, you know, there's. There is no funnel. There is. And I've always been a big advocate of a funnel is a great way of thinking about things. Are enough relevant people even aware of your product or company exists? And I'll tell you why I've just pulled this one out. We did a load of analysis and actually upon doing it, we worked out that about, you know, it was only a very small percentage of people that got to our website were actually getting to our sales pages, but when they weren't getting there, they were converting it at about 12, 15%. She's. Oh, the sales page is amazing. The thing is, is there's not enough people getting here in the first place. And without taking that step, we wouldn't have got there. So that that marketing funnel piece came out quite a lot. And do you mind just talking to us about that a little bit?
- [12:13]
B
Yeah, I mean, first and foremost, this is another one of those basics. I'm going back to what you said at the beginning, which is a core principle for me of the book. Like, principles are timeless. Channels change. So the marketing funnel has been around since the 19th century. I think Elias St. John, I think is the. The guy who's attributed with it for me, I like to use the old one, awareness, intention, decision, action. You can use whatever you like. The only thing I do is I add virality. Like there is a huge word of mouth element that drives awareness. You know, I look back at our contact importers, so I'm in a similar place to you. I don't think I even have this story in the book. This is an incremental one. Blake Ross, the guy who founded Firefox, came to work for Facebook. He is a genius. I love him. Really amazing guy. And also made me feel very welcome at the company as a gay man. So there's a, an element of that that makes him very special. He was tasked with growing the contact importers and as a good engineer, what he did was he ranked the contact importers on the contact importing page. He put them in the right order. Like, if you're a young professional, I'm going to put LinkedIn first. If you're like an older person, I'm going to put AOL and Earth link first. If you're a younger person, I'll put Gmail first. And you did all of this. And he got some nice uplift. And then I came along and I put a link on the homepage of Facebook and the thing went up tenfold. And he wrote, he literally wrote a post about this internally. Like, don't over optimize your thing. Understand if people even know your thing exists. And that is the difference. Like, you know these glasses which I'm holding and you don't have a video, but these glasses that we're selling, we, we did a Super bowl ad and like I'm a digital marketer, I'm a direct response guy. Why am I doing a Super bowl ad is because the awareness was low. We drove awareness in the US to over 50% of our ray Ban meta glasses. I particularly like the Oakley metas. That's, that's what I like. But we drove them to over 50% and sales went up a third in the week following the super bowl and stayed up because people just didn't know they exist. Like, awareness is a thing.
- [14:07]
A
Yeah, right.
- [14:08]
B
Do it.
- [14:09]
A
There's a, there's a lovely story from Kieran that he kind of came in an E commerce brand as well, working with. They came in on a Monday morning and sales have gone up about two and a half thousand percent. And they were like, what's going on? They look at the analytics and it says, well, search has gone up, social media has gone up, email's gone up, directs gone up. And oh, this is really strange. And then they go, oh, they were mentioned in the Times newspaper, two paragraphs. And it. And that's the thing. It's like, yeah, let's take a step back. We're digital marketers. But actually, this is a marketing mix.
- [14:33]
B
Right?
- [14:33]
A
So I think that was a great insight. I love this other one as well. Mediocre marketing with a great conversion flow is way better than brilliant marketing with a broken one. Talk to us about that.
- [14:43]
B
I mean, the whole way through. And then. Then you've got to say, with a great product or a bad product, like, that's the next level down. Yeah, look. I mean, a classic example of this was the brilliant Coinbase ad that they did in the Super Bowl. They did it. They did this ad which was just a QR code bouncing around the screen. Now, I am slightly upset about this because I wanted to put a QR code in that and wasn't allowed to another time. But I don't know, they got. I don't know how they got it done, but they done. And that's brilliant. And they drove so much top of the funnel to the site, and the site crashed and so nobody could convert. The conversion funnel was gone. And it really does matter that the basics work. The flow actually takes you through the flow and doesn't block you. There are so many. Like, when I started at Facebook, our, like, signup form was five pages long because we got you to fill out the profile in the signup form. We reduced it to five fields on the homepage. Guess what? All the numbers went up. So. So, yeah, you put the wrong conversion funnel in front of someone. It doesn't matter how good your marketing is, people will not get through that conversion funnel. You have a bad product. It doesn't matter how good your conversion funnel and your marketing is. We do not make up for that with marketing.
- [15:52]
A
Yeah, let's move on. Flex, there's another quote and there's two pieces to it. So I'll read you the whole quote and then I'll just go to the two bits. It says, first of all, channels. Pick channels by scale, defensibility and your capabilities. Then work down the list. Allocate budgets on marginal, not average, ROI and fit the channel to the conversion you want. So let's just go to the first big pit channels by scale, defensibility and capabilities. Can you just talk to us about that for our audience a little bit?
- [16:16]
B
Yeah. So think the first one's pretty obvious, which is what are the scale of results you can get through the channel? If you're absolutely brilliant at buying ads on a paper airplane's website with a thousand visitors a month, that's great. You're not going to get. Ah, thank you. He held up a paper airplane for.
- [16:33]
A
Yeah, I, I, we'll get to this in a moment, but there's some history of this. But I, I did that. Which was supposed to be one of your favorite designs. And I have to say I am very, very impressed. And we will put this into the show. Notes Target, Internet.com forward/podcast. Let's come back to it. I interrupted you. Please go ahead.
- [16:47]
B
That's great. That looks like the nagatamaka lock or the Knicks plane.
- [16:50]
A
It was Nick's plane. That's the one. Well done.
- [16:52]
B
Yeah. So, yeah, so, so you can buy it on like some niche website and that's awesome. But it's going to limit your maximum scale you can achieve. So you need to just have something big enough you buy on that. If you succeed at it, it will matter. I think one of the things I get a lot is a statistically significant result that isn't actually significant for the business. So you need scale. So the scale piece, I think is actually like really, really interesting, but easy. The defensibility piece is fascinating. So when I started doing advertiser growth for Facebook, we copied all the keywords that Google used acquire advertisers for themselves for our purchases on Google, which is awesome. But it shows how non defensible it is sometimes to have just a bunch of keywords you buy because someone can go and scrape Google or buy it from a third party and basically build a keyword list and immediately compete with you. When I was buying on Facebook, what I did was we actually created custom audiences, the original custom audiences. Danny Ferrante, in the core data science team of the company, he built this tool where we could take user IDs of people who'd entered the checkout flow for ads or people who had pages. And we were able to do these really clever data cuts to target people on Facebook with ads for Facebook ads. And that was completely defensible because nobody could recreate that list. Nobody could target the same people. And it put us in a place where when we wanted to acquire our advertiser, I mean, you look at numbers we put out, we have well over 10 million monthly active advertisers. Nobody else is putting out any similar numbers. We have a superpower in dealing with SMBs, and a ton of that is our acquisition channels, whether buying our own ads or in product merchandising. Which obviously no one else can put a link in the Facebook page asking a Facebook user or a link in the Instagram app asking an Instagram business to buy an ad. Those things have driven us to be able to have that huge number of active advertisers who are all small businesses and they're completely defensible channels that nobody else can get access to, which puts us in the lead. So scale and defensibility are really, really important.
- [18:59]
A
And then the last piece of that was capabilities. And I think that's just, you know, it is leaning intro capabilities as well, right?
- [19:04]
B
Absolutely. Like superpowers for us, user based targeting superpowers for us. Like in product conversion, you know, Harry Stebbings has this concept of channel market fit that he talks about for startups and I really think it's true. There is a channel market fit. You have a channel that is the right channel for you. Ideally it's defensible, but whatever happens it has to have scale and it has to lean into your capabilities if you can make it defensible too killer.
- [19:30]
A
And then the last piece of it is brilliant, which is allocate budgets on marginal, not average, ROI and fit the channel to the conversion you want.
- [19:38]
B
Yeah, I mean, this one is an old ebay thing. I know I keep saying ebay, but it was an amazing, amazing teaching ground for me in Richmond in southwest London. I was based at ebay. Yeah. So if you're thinking about incrementality and return on investment, I think the thing that you need to think about is marginal incremental returns, if you can do it. Marginal returns are so obvious. I feel embarrassed on a podcast like this saying it. But it's that the last hundred dollars you spend will make you a lot less money than the first hundred dollars you spend, if you're being logical. Online auctions, so Meta. But Snapchat, YouTube, Google, all of the places you can buy online at auctions. Second price auctions, they act very much like perfect economic systems, which is really cool and really interesting. And that means you get these diminishing return curves, which starts with zero. Zero. You spend zero dollars, you get zero. And then you get a beautiful diminishing return curve where your last dollar, you are getting usually nothing back for your last dollar as well in that theoretical model. And so you have to look at the slope of the curve at the spend that you have rather than the average slope to the spend. And that means that you'll find, and then we found this at ebay, that actually you're spending ROI negative way before the average goes negative. And if you allocate between channels based on the marginal return being the same across every channel, you'll actually find that you massively move up ROI just by budget allocation. And within channels, you can subdivide them so you can have clusters of users and look at the marginal return for, let's say British users versus French users, for example. And then you can optimize for the same marginal return between each user cluster or keyword cluster, whichever one you're working at. So, yeah, marginal returns. And if you can get really clever, you want to look at marginal incremental returns. That's pretty hard to do even at the scale we operate at. So I currently separate incrementality and marginality, but that's the way we look at it.
- [21:42]
A
Love it. I think that's. That's missed so much. So let's talk about the capabilities to do these things then. So how do you think that marketing organizations should start thinking about their team building, collaboration? How does that play into all of this as well?
- [21:53]
B
I mean, there's a lot of stuff in that. Which teams do you want to collaborate with?
- [21:56]
A
Let's talk about that because let's talk about that marketing engineering thing that you talk about a lot as well. I think that's really important and I just think that the role of that in these kind of businesses that are successful is really important.
- [22:07]
B
The engineering team. When I was at ebay, this was one of the negatives. I've said a lot of positive things about what was happening at ebay. This is a negative on what we did. We were in this place where we had a direct mail team and what they did was they extracted the data from the database, they transferred it to the marketing server at that time, then they loaded it into their marketing email engine or on site merchandising engine, and then they trafficked it. That process would take 48 hours, something like that, because the data would be processed overnight, like extraction, transfer and loading would take some time and then you'd send out the emails the next day. So it'd be two days. And then ebay had a notifications team in engineering who obviously would send the, you've just been out bid emails, right, Things like that that were critical for using the product. Why weren't they one team? Why weren't we in this place where if you had a brilliant marketing email, it would be able to be delivered by the massive infrastructure ebay had set up to send all of their notification emails? Honestly, the reason was the teams just didn't collaborate, the teams didn't kind of get excited about each other. So the marketing folks, and this was changing towards the time I left, which was 20 years ago. But the marketing folks didn't really value how important time is when you are doing database marketing and how critical it is to do something right at the moment the action has happened that you're triggering on. And the engineering teams, they didn't value the marketers, they didn't value the marketing insights and the ideas and the, the clever things that they had. And so they just didn't work together. And they built completely different email systems, which gave deliverability problems for marketing, which gave, you know, time lag problems for marketing, conversion problems for marketing. But on the other hand, the engineering teams had lower conversion rates than they would have had, despite the fact they had great deliverability and great timing because they didn't have great ideas. So at Meta, we built it differently. Like I am, my friend Naomi runs the product and engineering teams. I run the marketing and analytics teams. We are very close friends. We talk all the time. Our teams are fully integrated. In fact, exchange people between our teams and we operate as a single unit across engineering and marketing to do growth. And I think that format has led to what people call product led growth at this stage.
- [24:13]
A
Right.
- [24:14]
B
And the companies that do that, I think do a lot better than the ones that use the old model.
- [24:19]
A
And then how do you. We won't go into all the teams, but how do you think comms fit in with this as well? Because that's always a bit of a challenging one.
- [24:25]
B
Yeah. I mean, again, so much of this stuff boils down to interpersonal relationships.
- [24:31]
A
Right.
- [24:31]
B
You know, and I've advised a bunch of companies, I'm on a few boards. Fundamentally it boils down to can the head of marketing get on with the head of comms? So something that happened on our side is a lot of my teams wanted to do the social media for meta. It makes sense. I get why they want to do it. But Zuck does his social media through comms. That's where he goes to, that's who he believes in. And so I transferred the entire social media apparatus out of marketing. It was kind of at that stage, it was like a third in comms, two thirds in marketing. And I just moved the whole thing over to comms proact.
- [25:02]
A
Interesting.
- [25:03]
B
No one even asked me for it. I just did it. And you know what, that was really important for building a good relationship with the comms team because they had a vision that they were going to do modern communications. And of course they would manage. Meta has a standing press corps, which is kind of unusual for a company, but of course they would manage the press corps, they would work with our journalists, but they wanted to talk directly to our, you know, constituencies themselves. And for them that mattered to do social media. And so that move ended up solidifying the relationship between our teams and made us way more collaborative than we would have been otherwise. So taking these actions to actually like be a good partner is the number one thing to be able to work well with comms. Because comms is genuinely a fully people driven business. They aren't doing massive scaled marketing. It isn't loads of analytics, right? It's people, it's relationships, it's communications. So communicate well with them, be good partners.
- [25:57]
A
Brilliant. All right, well, let's move on to the subject of the moment. And does AI sadly or gladly, I don't know either way. But how, how do you feel? It's changing marketing, how is it going to change marketing? And I, I was really interested in how you reflected that on the book. So let's talk about that.
- [26:10]
B
Yeah, I mean, the quote that's doing the rounds at the moment that I really love is AI is not going to take your job in the near term. Somebody else using AI is going to take your job.
- [26:20]
A
Right.
- [26:21]
B
The thing people don't understand about us is that we have been completely disrupted in six years, like six years ago, every single ad on, well, every single piece of content shown to engage people on Instagram and Facebook were connected. You joined a group you'd like to page, you'd followed someone on Instagram, you'd friended someone, everything was connected now. And we just spent a bunch of time testifying in court on this. Now, the majority of the usage of Facebook and Instagram is unconnected content. It's mostly these short form videos and creators you didn't follow. And the only reason, the only reason we can do that is because of the modern AI systems allowing semantic understanding of content and also semantic understanding of a user's intent at that moment. Like if you come in and you're like, oh, I want cooking videos today you get a ton of cooking videos. If you come in and you want, you know, comedy videos, you get a bunch of comedy videos and it's amazing how fast the systems adapt to your mindset on the day you visit the product. So we've been totally disrupted. We are a completely different company now than we were six years ago. You layer in app tracking transparency from Apple, one of the More brilliant and cynical moves I've seen in the industry. It is only, it's only personalization. If it comes from the Cupertino region, it's sparkling personalization, sparkling tracking. Yeah. So we've also had to model a bunch of conversions and get to the place where we're using AI very heavily to use less data to make decisions in the service. And so those two things means both our engagement side of our business and our ad side of our business have been utterly transformed by AI in six years and they will be transformed again. We are going to see this and then over 10 years. Ten years ago, I would never use an automated campaign. Now all of my campaigns are auto optimizing, app install campaigns and similar. And the clever creative thing I'm doing is the data, thinking about what data we're feeding through, how do we tag it, how do we send it through, how do we actually optimize the flows, build the campaign structure and so on and so forth. Those changes aren't finished. So the industry is going to be totally transformed in the next 10 years in my mind, in three categories. One, you're going to do existing things more efficiently. The example there is, you know, you're not going to create 47 variants of the banner ad yourself. You're going to chuck it into an AI and the AI will produce them. That will destroy jobs. That bit is definitely going to cause there to be work. And that's where someone who uses AI will outsmart you. So you've got to be learning to use AI and it's within the capabilities today. This isn't some future, oh my God, this is going to happen. This is today. Two, we're going to be in the place where things that were possible but weren't efficient suddenly get done. So the example I use there is we offer customer support now in the United States, chat support. If you lose access to your account on Facebook and Instagram that is enabled by AI, you start by chatting with the chatbot. It took something that was prohibitive to do to make it actually viable to do. And we've always wanted to do it and now we can do it. And that has created jobs because there's still more people needed than zero to do that. But the AI has meant it's less than the ridiculous number of people it would have been before. And then the third category is things that were utterly impossible before suddenly become possible. And that's semantic understanding of content and content based ranking like, like TikTok short form video, YouTube shorts, Instagram, Facebook, Reels and so those three categories are coming along and they are going to change marketing. But marketing always changes. Like 20 years ago you made all your money from a creative agency. 40 years ago there was no Internet marketing. You know, 60 years ago you don't just had soap operas come along. 80 years ago there wasn't TV. 100 years ago there wasn't radio. Like every 20 years marketing is completely changing itself and that's awesome. And we're really lucky to be in a field that's actually very good at changing itself. So, yeah, there's going to be a lot of change. I think it'll happen in those three categories. Hopefully those examples give people structure to it and a person with AI will take your job. AI is not going to take your job in the next few years.
- [30:34]
A
I love it because all my students come to me in an absolute panic when I show them all the AI stuff in the first lecture and actually say, look, we're getting this stuff first. We seem to be getting the tech way in advance of. In many other industries and actually there is the opportunity there to engage with it. The book's out on the 7th of October. We'll put everything into the show, notes you can link through and. And kind of get the book from there as well. I just want to get into one final thing to explain to people why I was showing that you were paper airplane earlier on. Did you want to talk to people about that? A little bit of background on what that was. Little nice little nugget at the end.
- [31:01]
B
Yeah. I mean the way my whole career started was GeoCities for those of you old enough.
- [31:05]
A
Yeah.
- [31:06]
B
Remember that? Yep. Before it was bought by Yahoo and I created a website with Formula One paper airplanes, Fun science experiments. Because I was a nerd. The paper airplane bit started to rank on AltaVista because one company linked to me and so I then created, using free NetName, who was a web I remember well in the UK. Yep. I created my first URL, paper airplanes.co.uk and I created a website all about paper airplanes. And for the next 20 years I had the number one paper airplane site on earth. And it made me actually a lot of money. It made me enough money to pay for all my accommodation at college and my rowing training camps. I ended up having YouTube videos with it which have about 10 million views of me making paper airplanes. And they haven't been updated in a decade, genuinely. And it went really well.
- [31:58]
A
Yeah, it's a great example because I went on there and the first thing I did was make the one that was the top and it flies better than any airplane I've ever made before. So thank you for that. If nothing else as well.
- [32:06]
B
You're very.
- [32:07]
A
The last one as well. And we can edit this out if you don't talk about it. Was that on your LinkedIn profile you've got your grades for your degree and your MSC program?
- [32:17]
B
Yes.
- [32:17]
A
And I thought this was amazing because A level results have just come out. There'll be lots of, lots of young people around the UK and around the world stressing about their exam results as well. And as someone that teaches at Imperial College, you know we've got some of the highest grades in the world, like Oxford, Cambridge, all those kind of places actually you put on there. Look, this was going on my life. I got a tutu and actually, and I, I really like that transparency for the fact to say, look, actually this is what's going on. This is where I am now. So it's kind of irrelevant. And I just thought that was a. Given the time of where we are, I thought that was quite a nice thing. So I thought it was quite a brave thing to put up with your LinkedIn Prof. Which I guess is. Doesn't really matter now, so it's not that brave, but at the time, I guess it's interesting.
- [32:52]
B
Yeah, no, I mean, I did well at undergrad. I, to be honest with you, I really enjoyed myself at university. I, at high school, I did, I worked very hard, I was very nerdy. And then I came to university and I just really had a great time. I got a decent grade, I got two one at undergrad. Then when I hit my masters, I, I was coming out as gay and I got dumped by my boyfriend right after I came out to my parents and so like, it kind of felt the world was over and I totally mucked up my masters. I got a two, two and everything.
- [33:22]
A
I'm not sure that's totally messing up, if I'm honest. I mean, you know, I've done a lot worse than some of the things that I've done.
- [33:26]
B
That is fair. That is fair. Maybe that's, that's an unfair way to put it. I did very also, I did very bad at theoretical. I was very good at experimental, but my theoretical physics was terrible. So I'm not, I'm not a brilliant theoretician, I'm a good experimenter and that's what. And majority of the grade comes from theory, minority comes from experiments. So, you know, I went into the thing that was, was good for me which was experimentation. And I leaned into that.
- [33:52]
A
And what I took from that is look at your career and look at where that experimentation took you. And that's what marketing is. So I thought that was. That was great.
- [33:57]
B
Yeah. And look, one of the best people who's ever worked for me is this chap called Brian Hale. And he will have no problem with me saying this. He's currently the head of Growth at Doordash. Absolute genius, like, really brilliant guy. He didn't go to college. It's one of his LinkedIn too. Like, some of the best people who've worked with me didn't have the best traditional education or didn't even go to college. I mean, Brian genuinely was my right hand for, I don't know, over a decade. And he's now the head of growth, Chief Growth Officer at Doordash. Like, credentialism isn't that important.
- [34:29]
A
Right. Love it. Okay. I think it's a beautiful point to end on. So thank you for joining us, Alex. I really appreciate your time. I know you're a very, very busy man. Everything will be in the show. Notes Targetinternet.com podcast. You can buy the book. Please do. It's phenomenal. It's really, really worth. And you'll come back to it again and again and again and again. Alex, thank you so much for your time.
- [34:48]
B
Thank you so much.
- [34:50]
A
For more episodes resources, to leave a review or to get in contact, go to targetinternet.com podcast.