Despite Google’s assertion that nothing has changed in search for publishers and brands, the studies showing AI overviews kill clicks and referrals keep coming.
At the same time, some segments of the ad industry are drawing up new strategies for how brands can reach people through AI, as others claim that traditional SEO tactics will still do the job just fine.
For media planners, these are disorienting times.
So we asked Aengus Boyle, vice president at VaynerMedia, to explain how AI is changing search and what that means for brands.
Boyle also offered his thoughts on whether creative really is the new targeting, and why brands are sleeping on Facebook.
You’ve spent a lot of time thinking about how AI is changing search. Why is it important to know and what have you found?
It’s important to know because a large bulk of businesses across the world rely in some way on acquiring traffic and customers through search [and] consumers are moving to places like ChatGPT, Perplexity, and Gemini to search, at least in certain instances.
[AI search] is now starting to steal share from Google, which means if a business relies on Google and they’ve got great organic search rankings and they’ve got a great paid search program, where before they were going to capture 100% of the traffic, they’re now going to see that share decline.
If they don’t figure out how to have their brand visible across these AI touch points, then what was a very reliable source of traffic and revenue for their business is just going to start disappearing. And if they don’t future-proof and build into this new AI space, their brand’s kind of going to disappear over time.
Can you think of any brands that have mastered AI search?
It’s so new that I don’t think anyone can really claim to master it. The people I’m seeing who are most at the cutting edge — or doing the most testing and learning and, fortunately for us, also sharing those learnings — are certain SEO blogs or newsletters that I follow because those people are so close to the data that they’re the ones who are first to act.
But it’s so early that I don’t think there’s anyone who’s figured it out fully. I think any brands that are starting to think about it and experiment are at the forefront because most brands haven’t even registered it. But I think that’s really exciting because it means it’s anyone’s game to win. It’s an open playing field and it’s whoever can figure out the game quickest.
Ad Age had a piece earlier this week about how AEO is no different from SEO and that it requires the same tactics that website owners have been using for years to appear in search. What do you make of that?
I definitely do not think that’s the case because it’s just such a different space, and the mechanisms by which the information comes to land on those pages are so different from traditional SEO.
There’s definitely a crossover. There’s a Venn diagram, and there are certain elements that mean there is a notable crossover, like making sure your brand has authority across the web and that the content is relevant, authoritative and answers questions.
If you’re wondering what time does a certain performer go on stage at a gig, you’re going to Google it, and what you would usually get in a result that’s optimised for Google is a page that’s stuffed with keywords, and you have to scroll down three-quarters of the page for about 15 seconds before you finally find the answer […] because Google rewards scroll depth. So they’ll have hidden it down at the bottom, below 17 versions of the keyword to try and optimise for the keyword.
Whereas what gets an answer pulled into an AI overview is a very succinct and to-the-point answer to a question. So it’s having the question in an H2 or H3 header in an article and the answer directly after it. It’s less about targeting specific keywords in a keyword stuffing manner, and it’s less about scroll depth.
It’s more about giving the most accurate answer in as concise a manner as possible, and I think that’s the big shift that I’m seeing happening. So while there is crossover and there are certain elements that cause you to rank on Google that also influence ranking on AI, I think the way things are structured and the way information is presented definitely needs to be different for an AI world.
Beyond AI search, what other changes, either in tech or media habits, do you think are interesting or consequential at the moment?
A big one for me is YouTube and the way behaviour is shifting. It used to be very mobile-focused, and now a significant percentage of YouTube viewership is happening on TV, so people are screencasting or watching YouTube on their TV. Also, from a media perspective, that makes it an interesting space because you can get the equivalent of a TV ad buy but with the targeting capabilities of Google’s ad stack on a TV screen for a much more efficient cost than if you wanted to do a linear TV deal.
That’s definitely one interesting space. And then, more broadly, just the entirety of organic social across channels like Instagram, LinkedIn, TikTok, and even Facebook. The amount of attention there is still undervalued. These algorithms are so advanced now at serving the right creative to the right people; it’s such an effective tool to find what resonates with audiences. And the fact that there’s so much attention on those platforms almost makes them like living focus groups.
Everyone talks about Instagram and TikTok, but Facebook almost seems to be forgotten. What are you seeing on this platform, and how can brands make use of it today?
I completely agree. I think Facebook has kind of been forgotten, and people are really sleeping on it. But I think more and more, it really does have potential. It does skew older than a platform like TikTok, but it probably skews less old than people think. Equally, even if it does skew a bit older, that older skew tends to be towards higher value consumers, given the nature of how age and life progression works. So I think there is a really interesting opportunity on Facebook and I think brands are probably not tapping into it as much as they should be.
One thing that I know is really driving reach on Facebook is text posts with a coloured background, where you can have a text post but it takes up space like a photograph. We’re really seeing those drive a lot of success.
Indeed is one of our clients and they did a text post that got well over a million views. That was such a strong signal of something that really resonated that it ended up being turned into a billboard. So they turned that into an out-of-home campaign.
Those million-plus views in their own right are hugely valuable and can inform other tactics. And that billboard is clearly going to resonate more than something on Facebook. But Facebook is like a proving ground to then help inform and fuel campaigns across the board.
Ultimately, yes, I think Facebook is undervalued by marketers and has a lot of potential from an organic reach perspective and from a paid media perspective in terms of their ad stack, which is very advanced across the funnel, from awareness through to driving conversions.
VaynerMedia is big on agile media planning. What does that mean, exactly? Why is it useful?
As a company, our whole ethos is that we want to spend our clients’ money as if it’s our own money. It means if one week something is working well but the next week it’s not, we pay attention and we notice that and we adapt. I think part of it should just be what media buying is, and I think in a lot of instances, it is how media buying operates. In a lot of instances though, there is the sad fact that media agencies build out a media plan and then they just blindly execute it and consider the job done.
I think more advanced media agencies take that media plan and then look at the data each week, and they shift budgets between tactics within that media plan. I think that an integral part of what we do is really sweating the details of the data, having updated data flows and understanding day-to-day and week-to-week where we are seeing performance.
The other kind of more secret sauce thing that we have is that we are an integrated agency. We work with the strategy team and with the creative team and we really understand what the latest thing across the platforms is, what they are pushing, what is working organically, what the algorithms are favouring, and where that intersects with paid media.
Each week, we take the most holistic view we can of what’s going on in the market, what’s going on with our clients’ data, and how our media plan should change based on the realities of this week versus last week because it’s such a fast and ever-evolving space.
How do you feel about the future of media planning at the moment? How is the discipline evolving and where do you think it will continue to add value in the future?
I think media planning is hugely valuable to get to the starting point, but I think it needs to be an ongoing process. It’s almost like media planning becomes a living, breathing thing that’s for the whole length of a campaign.
In terms of looking to the future, the thing that I think probably scares a lot of people that I find quite exciting is how AI is going to become a tool and a partner.
There’s so much experimentation and learning to do, but I think there’s certain tools that allow you to ingest a huge amount of proprietary data and use [it] as almost a partner. You can then train the model on all this data, and you can have a conversation with that data or with a super-intelligent being that can assess all of this data and give you the trends that it sees, and the potential opportunities or blind spots.
I think the layer that’s going to get added on is taking owned data and then also the data across the internet — the latest trends, updates, etc. — and using AI as a tool to fuel and fact check, so that planning can kind of go to the next level.
Earlier this year, we heard a lot of conversations about creative being the new targeting. Can you expand on what this means to you and whether you think it still applies?
I think it definitely still applies across social platforms, and even places like YouTube, to a slightly lesser extent. I think the point applies across every platform, but Meta has openly come out and said that 56% of all auction-based outcomes are driven by the creative.
That alone is a pretty astounding fact: over half of your effectiveness relies on the creative. So if you have a perfect media execution with a terrible piece of creative, the best score you can get in the test is 50%. Whereas if you have a mediocre media setup and really effective creative, you’re probably going to get 70%.
In terms of why that is, it’s probably a couple of things. One is privacy, and the fact that previously, the targeting granularity and the level of depth and specificity you could target with on Meta was insane. And since The Social Dilemma came out, and Cambridge Analytica and these various things that have led to real privacy regulation, targeting has been getting less granular and less direct.
At the same time that’s been happening, the algorithms have become more and more effective at honing in on each user and what sort of content resonates with them. This now means that, because we can’t target with the granularity that we used to within the ads platforms, the algorithms can find the sort of people who resonate with certain types of content.
It’s what you used to be able to do manually within the ads manager platform. The only real way to reach very specific audiences now is to have a creative that speaks to those audiences, that the algorithm will then find and serve that content to.
Featured image: Joshua Hoehne / Unsplash