What impact will AI Overviews have on PR, SEO and PPC?
There’s been a lot of discussion in the marketing and media press recently about the impact that AI Overviews are having on search.
TL:DR… I don’t think AI Overviews require a radical re-think of most marketing models in the short term, at least in b2b. As always, my general advice for a marketer looking for an edge would be to focus on building a brand that makes people feel something, and connect with those people by making them think.
But nonetheless, here’s my take on AI Overviews, as they are causing a stir in medialand.
Publishers are at the sharp end and, sadly, in big trouble.
Over time it’s become harder for publishers to attract and retain paid customers through subscriptions, and as a commercial model, the relative importance of digital ad revenue has grown. For a while, some publishers really prospered from digital ads.
But more recently, Google’s triple whammy has put more pressure on the ad-first publisher model. They are paying less for impressions (dropping unit price), widening the pool of inventory (increasing the competition), and now cutting off traffic altogether with AI Overviews (reducing sales volume).
This has hit publishers hard, with ‘shocking’ fifty per-cent-plus reductions in click throughs reported at some titles like the Daily Mail, which is having an obvious impact on revenue.
There’s been some speculation that this has been part of the reason for TechCrunch’s exit from Europe and layoffs at Business Insider (where traffic has halved in five years).
Both titles have broader revenue streams than display ads. Business Insider has been focused on a hard paywall strategy (always hard to pivot from ad-funded to free) but the impact of AI can't be overlooked in terms of the timing.
Given that driving traffic to websites via search has also been a staple of marketing for years, is this going to change?
Big brands and differentiated brands will still get found, and may even thrive
Some numbers for context:
In 2022 an Semrush report said that more than half of users didn't leave Google after conducting a search.
The Stanford AI Index Report found that AI-generated responses now answer over 50% of common coding and factual queries, without requiring click throughs.
AI-generated answers in Google's Search Generative Experience (SGE) do not match any links from the top 10 Google organic search results 93.8% of the time.
So yes, perhaps the content sources AI systems draw from operate outside the traditional SEO playbook.
But I expect companies that have built up a lot of digital brand capital (i.e. SEO on core terms, well-established owned media operations) may even do better.
History suggests that as the media real estate shrinks, the already-big get bigger, and the distinctive and talked about stand out more. Like being on the homepage of Netflix, an established talent in a cast list, or a famous artist on Spotify - people default to what they know and I expect the AIs will do the same when compiling summaries.
Even if SEO loses its effectiveness over time, new brands shouldn’t be deterred from owned media, as it’s important to keep mid-funnel prospects engaged with useful content and show what you’re about - it’s perhaps more a case of quality over quantity, and optimising for humans rather than machines.
Paid search doesn’t look massively impacted by AI, just yet
Where does this leave paid advertising?
Most people think the impact will be minimal, because paid search is mainly used to achieve lower and mid-funnel objectives. While there’s not much hard analysis of this at this point, the thinking is that keywords people are bidding on would be unlikely to feature in the upper-funnel general searches AI Overviews tend to cater for.
Maybe. Spending on generic search is generally not done so much these days by big spenders in established categories (eg Nike on running shoes) as it’s seen as wasteful, but for companies in emerging categories or niches it has a place. For example a search on ‘physical AI’ - widely tipped to be a major tech category in coming years - delivers no ads for me, with first place being afforded to an NVIDIA landing page (who have sought to own this category).
Anyone else interested in buying that term could possibly feature on page on, but the result would still appear ‘below the fold’ of the AI Overview. That’s a missed revenue opportunity. I don’t imagine Google would seriously cannibalise a major-revenue-stream-of-today (in search) for a maybe-in-the-future revenue stream (in Gemini). So we’ll have to see.
Given that AI Overviews occupy important real estate at the top of the page, perhaps over time, people may end up clicking on paid search results a little less, and I expect Google will optimise around that as time goes on. But in a marketplace where marketers are charged for clicks, rather than impressions, this probably has minimal impact on their ROI.
It’s not clear how the Copyright War will play out, but this will matter a lot
As a PR specialist this is the area of most interest to me: it’s not clear at all how AI Overviews will be impacted by the ongoing arguments around AI firms’ unlawful use of copyrighted content.
Here’s how I see this three-way brawl:
In the Red Corner, Property Owners. These are people who want to make money from their property, so are always looking for ways to create new income and protect what they have (copyright is how they do this). They don’t want other businesses to steal their property and make money off it.
In the Blue Corner, Big Tech. Content is what people use their technology for, but it’s a cost of business. It’s more profitable to ignore copyright, or, better, destroy the idea. This is why they spend more money on lobbying than licencing (I expect).
Finally the Referee? Regulators. At best, they simply don’t understand copyright and have drunk the big-tech Kool Aid or at worst, recognise the futility of trying to defend citizens against the excesses of big business, and have given up.
I’m hardly a non-partizan observer given a lifetime working in the creative industries, and as a media observer and supporter of journalism (the most underrated pillar of a properly-functioning democracy).
I’m genuinely not sure how this will play out.
The optimist in me says the rule of law will prevail and AI firms amassing enormous valuations through what’s basically industrialised theft will collapse. On the other hand, despite repeated scandals over data privacy and now suicide grooming, stock prices continue to soar. The much-mooted big tobacco moment doesn’t feel any closer.
What I think suspect might happen over time is the continued erosion of the web, as people react to the current stasis
publishers start to withdraw quality content from the open web. If you can’t stop your IP being stolen, I can see even less content going onto open websites and more into other channels/ behind stronger paywalls, meaning quality content is gated-off where the bots can’t get it.
the creation of more second-rate content. Already a lot of Gen AI research I get is rubbish > not because the AI isn’t technically good, but rather it’s been tainted by the second-rate content that already exists, and I suspect bots are most attracted to websites that favour quantity over quality. So it’s a doom loop. There’s only so many times a thirsty castaway can drink his own urine before it kills him.
increased investment in anti-hacking technology to mitigate the chances of content being stolen in the first place, and I wonder what a counterattack might look like especially as the AI firms invest more to hack through improving defences. The EU has recently committed to spend five percent of its GDP to fend on physical warfare and the protection of digital property, perhaps, will see similar investments over time.
What does this mean for media relations and PR?
Earning coverage in credible media titles has always been great for leads. Quite apart from the benefit of having human beings validating you, and talking about you (and what this does in terms of brand recall and trust in the mind of the buyer, investor or whomever you are trying to reach) good PR validates your brand on the first page of Google forever (or at least until it gets bumped off by the next piece of coverage). For now, at least.
It certainly helps raise awareness, if you’re named in an article that talks about solving the problem you solve. It drives consideration if everyone can see you’re trusted and talked about. And it can drive sales if people are reading it whilst in market for what you do.
I don’t think any of this will change overnight.
In the mid-term, I think the quality publishers that choose to licence the AIs will become more important from a PR perspective.
Not only will AI firms favour the sources that allow themselves to be crawled, but they’ll be commercially incentivised to push the compliant publishers up the rankings.
Be fascinating to see where we are in a year or so.