How to Prepare Your Marketing for AI: Data, Strategy and Systems

Maelien: Welcome back to the podcast, and let’s start by saying that this might be the last one for a little while. We’re gonna take a little bit of a break.

Louis: At the moment, we’re in the middle of moving offices and building a new podcast studio, so we thought it was the perfect opportunity to take a step back, do a bit of a review, and come back better with a fresh new approach. And so with this in mind, we thought that the best kind of topic for today’s episode would be something that we can leave you with some food for thought over the coming weeks.

Maelien: And the thing that we’ve decided to talk about here was AI. So when I was watching this year’s Google Marketing Live a few weeks ago, they were talking about all the upcoming launches that they had with AI. It covered things like being able to chat with an AI agent directly through ads for certain verticals like automotive, real estate, and education; being able to brief in your strategy and tone of voice so any creatives that are created or ad text that’s made considers your strategy and the tone of voice that you want to use; as well as being able to directly explore websites from within Google Chrome’s AI mode without ever leaving the search results. So given all of this, if there was one thing to focus on over the next few weeks, I definitely think it is how AI is shaping the future of marketing and advertising.

Louis: Okay, cool. So to kick things off, Maelien and I have had a little think about one or two things that we’d focus on to get AI-ready. So let me get things started. The number one, and it’ll probably come as no surprise ’cause I’ve talked about this a number of times before, but it’s conversion tracking. Getting conversion tracking in place, making sure the data is clean and robust and trustworthy. Google will want to give positive feedback on what’s happening in the ads, whether or not conversion tracking is in place. So for me personally, if I’m gonna be getting feedback via AI on what’s happening in the advertising, I want it to be based on the most robust and valuable data possible.

Maelien: Just to kind of ask a question there, Lou, about valuable data. In your mind, what does that look like?

Louis: Yeah. So it’s a great question, and I think, you know, like if you haven’t got conversion tracking in place, well, you’re probably gonna get feedback on things like reach and impressions being positive exposure, or CTR showing a good amount of engagement, or video completion rate showing people positively engaging with videos, which is, I guess, all interesting. But if it’s not coupled with “and then this many people enquired,” then it’s not as useful data as it could be. So in terms of what a valuable conversion would be versus one that wasn’t so valuable, you know, I care a lot less about email newsletter signups than I do about demo requests or service inquiry forms being filled out. So that kind of illustrates that a little bit.

Maelien: And going back to this idea of being able to interact with websites within AI mode, what’s that gonna look like for conversion tracking? I think that becomes quite interesting since they’re not actually going to the site here.

Louis: Yeah, if I’m being completely honest, I need to explore it more myself. And now I know that, or at least I’m pretty sure from what I’ve read, when you are browsing a website within AI mode, it will come through Google Analytics as being from organic search and from Google. Um, but as to whether or not we’ll get the same level of tracking or if we’ll be able to distinguish that that was a session within AI mode… So are we gonna be able to filter between like full website sessions and those done in AI mode? And is tracking gonna work completely the same? And how will users behave differently? So while having the AI mode interface around a website, it makes it a lot easier to see other websites and go to other websites. So it stands to reason that it’s gonna be a lot easier to abandon and to visit other websites and carry out actions on other websites much easier as well.

Maelien: And a big one for me here is systems thinking. A way I like to think about it is having a helicopter view rather than viewing it hidden in the weeds. So in practical terms, what that means is that you’re focusing on the strategy and how everything connects together versus maybe focusing on all the details and the minutiae. Because at the moment, everything does feel a little bit more detailed when it comes to ad management. Because when you look at the things that are getting replaced, like with Meta going to Advantage+, which allows audiences to expand massively, things like audiences are being replaced, things like bidding becomes automatically done, keyword targeting… it becomes clear those detailed bits are slowly getting replaced. And so strategy and being able to point things in the right direction becomes really important.

Louis: It does make me wonder if this is the idea between the kind of AI brief function that they’re bringing out. Maybe the problem was that they were a little bit too detailed, you know, creating headlines and descriptions in isolation rather than being able to reference like a central system-level strategy and tone of voice doc. “This is my strategy. This is my tone of voice. These are my objectives. This is what I’m trying to achieve.” And then starting from there and looking down through the campaign, through the different levels to assess and kinda steer in that way instead.

Maelien: Because I think we’re gonna start to see a change now, aren’t we? Where like if we went into Google Ads, and we used AI-generated headlines and descriptions, eight out of ten of those, we would feel quite strongly that we could write much better.

Louis: Versus if it did have that layer, that kind of AI brief—strategy, objectives, tone of voice—if it knew the advertiser better, you know, chances are it would get a lot closer to something that was usable. So, um, I guess the thing that I wanna put on the table next is systems integration. I kind of feel like we’re entering into probably one of the most challenging, different times in marketing that I can ever remember. I also think we’re entering into an age with the biggest kind of opportunity as well. Previously, if I wanted to kinda take Google Ads data and Google Search Console data and HubSpot data and Google Analytics data, and then if I wanted to blend that, I’d have to, like, do it all at the data level. So I’d have to find common metrics and common dimensions and find a way to kinda normalise them. Whereas now, there’s a lot of support that we can get from AI to blend that together. Like the metrics and the dimensions don’t have to match up perfectly because with AI being an intelligent model, it can kinda join the dots. Uh, that’s definitely something that I’m focusing on right now and something that I’m really starting to get to grips with is, you know, like there’s lots of gaps between the data that we have. How do we kind of reduce those gaps a little bit and learn more in less time really?

Maelien: So the next one on my list here is having independent thinking, and the reason I think this is so important is because if you’re to leave AI to its own devices, there’s a real risk of it kind of going rogue. And I think a point that illustrates this really nicely is I often have Reddit notifications around PPC. And one of these threads here was about someone saying, “Look, my ads were working great. I was getting calls through every single day. But I followed the recommendations that ChatGPT gave me and now nothing’s working. I’m not getting calls through, but I’m still spending money.” And they were asking for advice, like, what should I do? And obviously, a lot of people are kinda diving in and saying, “Well, the first mistake you made here is listen to all those recommendations and just act on them and put them in place without really thinking any more into that.” But the other one also is that you put so many changes into place at once that it became hard to isolate, like, which change or changes caused the impact. So all of that is kind of pointing out how far astray it can lead you and how it can take something that could be really good and turn it into something really bad.

Louis: Yeah. Completely agree with that. You want the best data going in and feeding AI in order to give you some really juicy insights that then you can put your own stamp on, but you wanna kinda take that, and you wanna advance that, and you wanna put your own spin on it. You know, if you don’t start taking up AI, you get left behind. Um, but equally, if you rely too much on AI, you get left behind as well, um, which kind of creates this kind of notion of conviction, right? You can use AI to do most things in business now. But the question isn’t can I do it, it’s should I do it?

Maelien: Yeah, absolutely. It’s really important to have that human in the loop. It’s like AI is great, it can do great things, but we really need to be able to keep it in check and really need to be able to think of your perspective. So if you’re an expert at something, you should probably think to rely on yourself first. If you’re not so much of an expert on something, yes, use AI to help, but also do the research as well. But if it’s replacing you and your thinking, you should really start questioning, you know, is it better for me to put some time into the thinking here?

Louis: Yeah. Yeah. Such an important point. So I guess now’s probably a good time to do a bit of a recap. Uh, so I think the first thing, if you’re not sure that all of your measurement and your tracking is in place and is robust and is comprehensive, now’s the time to really get that in place. You really wanna be in a place where you can trust your data because when everything has moved further towards AI, which is happening every day, you want to know that when you’re uploading your data that it is correct and accurate. You also, with having that data correct and accurate, you also want to make sure that you’re really clear on the strategy level, the tone of voice, your objectives, and everything that’s gonna enable you to do systems thinking better. We’re moving from a place where performance marketing and advertising is very detail-led to where it is more systems-led. And, you know, there’s gonna be elements of detail-led management, but we’re moving to a much more top-down approach as everything becomes more AI central. The next thing we talked about was making sure that all of the different data that you’re collecting is formatted, and if you can, blended together to give you insights. That’s gonna help you be as informed as you can possibly be and keep on track with how things are changing, what’s working, what’s not working. So make sure that your Google Analytics is properly configured, make sure your HubSpot is properly configured and is kept up to date, and then, you know, export it to something like BigQuery, and then experiment with how can you query that data, how can you gather insights, how can you build reports that’s gonna keep you on the pulse of how things are moving forward. And then finally, it’s really important then to not rely 100% on AI. It’s really important to think independently, to add that 20% of special sauce that only a human can do, to treat AI more as an experienced junior, and then acting as the specialist, the expert, and taking what they find further, taking the work that they produce further, but also questioning whether it’s accurate and going back and spot-checking and making sure everything is good as well.

Maelien: That’s a great place to wrap up. So it’s been great creating the past 30 episodes, and we definitely couldn’t have done it without you, so thank you very much. To make sure you don’t miss us when we’re back, please hit the subscribe button wherever you’re listening to this. We’d really appreciate it. And if you’ve got a topic that you’d like us to cover, then head to webmarketeruk.com/topic. When we come back, we’ll be doing this from our brand new studio. Before then, take care of yourself and we’ll catch you before you know it. Thank you very much.