The concept of Fantasy Island, published by Rotten Books, sounds simple and quaint – 50 years of Irish photography by 70 Irish artists – while the final product is anything but. Flicking through this book is like riding a screwball carousel, a kaleidoscopic history book that offers everything from black-and-white film photographs of punk movements, queer intimacy and nightlife, to ambient scenes that play with light, colour and shape. Attempting to encapsulate this sprawling project, Belfast-based Joel Seawright, fine artist and founder of Rotten Books, and graphic designer Lucy Jackson, lift the lid on how the book came together.
“The goal was to create something that felt truly representative of Ireland as seen through the eyes and experiences of people who live here,” says Joel. Deciding to pause their previous project, the stylish Rotten Magazine, to focus on something more expansive, the main focus was authenticity. That meant collaborating with Irish artists who don’t just have lived experience in Ireland but also shot Ireland in all of its beauty, grit, tension and humour. “We avoided the stereotypical portrayals of Ireland that often appear in wider media,” says Joel. “Fantasy Island doesn’t attempt to tell a single, linear story of Irish history,” adds Lucy. “Instead, it acknowledges how layered, emotional, and unresolved that history is.”
With images starting in 1975, naturally The Troubles surfaces, but Fantasy Island diverts from the trauma-dumping of cliche-laden media surrounding Ireland. Making sure that violence and resilience have their moments, the book also humanises the country with creativity, culture and colour. “Rather than centring the conflict, we tried to let it sit alongside other narratives,” says Lucy. “Making space for the complexities that often get flattened in mainstream portrayals.”
To my fellow marketers, I first wrote this title in the summer of 2020, back when I thought, “Wow, surely things couldn’t get worse.” Needless to say, I was wrong.
Here’s the actual quote I started with last time:
“If you’re reading this, then it is early July, you’ve made it this far in the game of ‘Let’s See What Else Can Happen in 2020’.”
We have largely left the world of all-day Netflix and sourdough, and moved on to more pressing things like understanding the impact of tariffs on a brand’s willingness to run digital, and wondering how, five years later, my NY Jets are still so terrible.
With those changes has come a shifting dynamic in search, once called “PPC” (I have always disliked that term), more recently referred to as search engine marketing (SEM) and paid search, which is now simply “paid media.”
With this shift in ad types, ad placements, and management comes a shift in how we target audiences for our ads.
Why? Ad technologies change, ad units change, and thus, targeting changes. Not to mention, a shift in “what is demand?” affects more people than those who are actually qualified to see your ads.
I didn’t see economy-driven searches overtaking COVID-19 in my future (Screenshot from Google Trends, June 2025)
And once again, there are caveats:
Consumer sentiment is in flux as the economy rocks back and forth from concerning to good.
Google’s look-alike audiences (similar audiences) sunsetted (except for Demand Generation).
Audience targeting can easily be mixed up with various forms of AI targeting (i.e., Meta Advantage+).
Cookie deprecation started and then stopped, but first-party and modeled audience data became worth as much as gold.
The concept of the keyword match type (or even the keyword itself) is continuing to erode away.
Not everyone who views your ad is truly qualified. Whether it is in-market, demographic, geographic, behavioral, etc., not everyone should see your ad.
To put it bluntly (and I am trying my best not to sound rude), some individuals are not worth spending ad dollars on for a specific ad.
For high price point items:
Income often correlates with CVR based on category (Image from author, June 2025)
For more age-specific items:
Age is often a deciding factor as well (Image from author, June 2025)
With times being as uncertain as they are, brands must tighten their purse strings and become more selective in their prospecting efforts to help the bottom line.
One would think that this concept, focusing ads on a particular audience, would always be the case, but the reality is, mid to larger brands will still often do the “spray and pray” approach, with just small audience adjustments.
Why?
Tighter audiences help with return on investment and efficiency, but they can wreak havoc on volume and total revenue when done too excessively.
This leaves the advertiser with a decision to make: What is the best approach?
The most important (and first) step: Identify who your ideal customer is.
Important disclaimer: Identify who your ideal customer is/has been, not who you think it is going to be/should be.
Be sure to pore over your analytics and conversion data to decipher this. Otherwise, any future steps are pointless.
Learn exactly who your converter is (Image from author, June 2025)
Previously, to weed out the less qualified and still feed the top of the funnel and prospect, you would need to lean heavily into audience exclusion and audience targeting. That is still true, to a degree, and more specifically in the case of paid search.
However, for more modern concepts, such as Performance Max, Demand Generation, LinkedIn, or Meta, we are leaning more toward the target, as the exclusion may not be as readily or easily available for use.
Audience targeting vs. exclusion: Yes, they are similar, but different. Here’s a quick refresher:
Targeting: The direct targeting of a specific group of consumers who fall within a certain characteristic(s), enabling everyone who meets it to see the ad.
For example: “I am selling a luxury car with a high price point, so I am only showing the ad to those whose household income is in the top 10%.”
Note: This is still valid in most scenarios. However, certain platforms and verticals do have limitations or restrictions.
Excluding: Indirectly targeting an audience by minimizing the ad units’ reach, based on consumers’ characteristics, by intentionally preventing ads from showing to those individuals.
For example: “I am excluding homeowners, so they are not served my apartment rental ads.”
Not doing one or both is as good for you as trusting a truthful outcome from Theranos.
How does one use these targets and exclusions to tighten one’s belt?
This is not rocket science, and more importantly, it doesn’t need to be applied account-wide, just high (sometimes mid) funnel initiatives.
Particularly in search, the more specific the query (often mid- to long-tail searches), the higher the qualification, the higher the likelihood of conversion.
But those are often few and far between (terrible for prospecting in terms of feeding the top of the funnel).
So, audience targeting becomes a necessity for high-volume search keywords. Otherwise, you’re spending your already limited budget on everyone (not ideal).
We break audience targeting into two types: actualized behavior and user traits.
The most common form (and easiest to use) of actualized behavior is retargeting.
Cart abandoners are the lowest-hanging fruit. It is a simple setup and deployment (I am a huge advocate of it via Google Analytics 4):
As much as I dislike GA4 UI vs. GA UA, they make audience creation fairly simple. (Image from author, June 2025)
But keep in mind: If you’re still getting those queries off a top-of-funnel query (generic, short-tail), then the qualification is already lower to start off with.
Frequently, we separate out retargeting past shoppers, retargeting site/cart abandoners, and prospecting (brand new visitors) from one another. Thus, controlling spend, creative, and user experience for each category.
At the same time, these lists can be used as exclusionary, ensuring there is no overlap, and a consumer receives an experience they were not intended for, which works well for prospecting audiences.
When thinking about user traits, these can be tied to platform-predicted behavior (i.e., affinity or in-market), or even self-identified characteristics (i.e., age, gender, income, etc.).
User traits are great at isolating targeting to your most qualified/relevant audience.
For example, anyone can eat at one of my fast-casual restaurant locations across the major cities of Connecticut.
But suppose I want to maximize the cost-per-customer efficiency for the “kids eat free” special. In that case, I will target parents of children under 12, not in the top 25% of the Herfindahl-Hirschman Index (HHI), but who have some disposable income, who enjoy eating, and are within a five-mile radius of one of our locations.
Make the audience that meets your typical customer (Image from author, June 2025)
But a nice little function these days is that Google and Meta are learning from current activity to help build out in-market audiences on a rolling basis.
It is great for all of Meta, PMax, YouTube, Demand Gen, etc.
Google is finally being helpful without a sales rep (Image from author, June 2025)
Using these tools, we have taken a step to prequalify the audience we’re prospecting. If they don’t convert at first (but do engage with the page), at least they’re pulled into our remarketing lists as a higher degree of qualification for later.
Net-net: These consumers are deemed worthy of seeing our ads.
To put it bluntly, exclusion is a vastly underrated, yet wildly glorified version of a search negative keyword list.
But rather than saying we don’t want to show if someone searches for XYZ, we say, we don’t want to show for you.
When we apply exclusions in any channel, we are saying, “I am open to anyone seeing my ads, provided they aren’t [fill in the blank].”
I know it sounds harsh, but it is highly effective and important.
Remember, not everyone is right for your brand, but they may still try and find a way to see the ads.
Exclusions can be simple, such as geography or time of day, or they can be much more specific.
One of the key times I see this needed is for YouTube and Google Display Network (GDN).
You want to capture a wide audience, but you know not everyone is right.
I should note, though, that certain verticals (those falling under Housing, Employment, and Credit or HEC policies in Google and anti-discriminatory policies in Meta) limit what can be excluded.
In addition, the rapidly growing share of wallet ad unit, Performance Max, in both Google and Bing (I still refuse to call it Microsoft), you cannot exclude audiences (yet), but you can exclude keywords (Google only beta) and brands.
Some day… (Image from author, June 2025)
It is a glorified negative keyword (Image from author, June 2025)
You’ll get fewer visitors, but a more qualified audience. You also maintain control of who you’re spending ad dollars on.
We are in the early stages of exiting the world of keywords and focusing on the audience. At the same time, platforms continue to reduce control and transparency of who/what/when/why/how your ad is served. That hurts your wallet and your bottom line.
When you can’t use first-party audiences, learn your typical customer’s profile, and build audiences for it.
By ensuring you target the right audience and exclude the wrong ones, you can make sure your operation continues to thrive another day.
More Resources:
Featured Image: ICONMAN66/Shutterstock
The FTC approved Omnicom’s purchase of Interpublic Group this week, after the ad agency giant promised not to engage in any nefarious brand safety activities.
If you think I’m exaggerating, the proof is in the consent decree:
Omnicom cannot base its “advertising spend … on the Media Publisher’s political or ideological viewpoints, or the political or ideological viewpoints expressed in content that the Media Publisher sells advertising to run alongside.” It also cannot rely on third-party “exclusion lists” premised on political or ideological views to determine where it will direct advertising.
Isn’t that, in part, how brand safety works?
If you were wondering, this is happening under the fig leaf of “protecting free speech” to prop up X/Twitter and help that poor waif, Elon Musk. Here’s how it came about:
Omnicom — parent company of agencies like BBDO and TBWA — is buying Interpublic Group, which owns McCann Worldgroup, in an all-stock deal that values IPG at about $13 billion. The combination will be the biggest ad company in the world, with around $25 billion in combined net revenue based on 2024 numbers.
The deal was announced in December following President Trump’s election. At the time, executives from both companies expressed confidence that the new “business-friendly” FTC wouldn’t ask any pesky questions about anti-competitiveness would quickly approve the deal. But the folks at Omnicom got caught in the administration’s Woke Ness Monster hunt. Said monster being the reason for the outbreaks of empathy, tolerance and representative democracy undermining America’s greatness.
Dig deeper: The war on DEI is hitting marketing and hurting business
In March, the FTC asked Omnicom and IPG for more info about the deal. Under normal circumstances, that’s a standard request. However, the request came just as the commission started investigating big ad firms and progressive advocacy groups for not advertising on X. Elon Musk’s media platform. (Anyone here remember the 1st Amendment? It was a great one.)
X (née Twitter) was quite popular with the public and advertisers until shortly after Musk bought it in November 2022. He paid $44 billion — twice Twitter’s market cap — for a company barely turning a profit. Musk borrowed the money to buy Twitter, even though he didn’t need to, saddling the company with the debt, and why not? He was never going to be able to earn back what he paid.
He then did everything possible to scare off advertisers. He got rid of the content moderation team and the trust and safety council, and the site began to be overrun by fake and unverified accounts. In the first month of Musk’s ownership, a bogus Eli Lilly account said the pharmaceutical company would give customers insulin, causing its stock price to drop 4.37%.
By early 2023, 50 of the site’s top 100 advertisers had stopped advertising or reduced it considerably. Later that year, amid a significant rise in hate speech, Musk posted a tweet endorsing an anti-Semitic conspiracy theory. He followed that with a tweet about the thoroughly debunked right-wing conspiracy theory “Pizzagate.”
To be fair, he apologized for the anti-Semitic tweet at a New York Times event. However, he followed that with a profanity-laden rant about the companies that had stopped advertising on X, singling out Disney CEO Robert Iger. Immediately after that, Walmart stopped advertising on the site.
A report released in September found only 4% of marketers believed the platform provided adequate brand safety. Musk then took full responsibility for the fiasco and promised to make changes. HAH! He sued marketers and watchdog organizations like Media Matters for allegedly colluding in an illegal ad boycott.
Dig deeper: Hate speech on social media can significantly damage brands
He also made full use of his bromance with President Trump to strong-arm advertisers, threatening even more brands with lawsuits if they don’t spend more on X. It is unclear if the risk of action by government agencies was stated or implied. I haven’t seen any new survey numbers, but it’s unlikely these actions improved marketers’ opinion of X.
When I was a newspaper editor, the only rule regarding advertising was that stories about airplane crashes were not allowed next to airline ads. That’s the simplest form of brand safety, and it may also be the only form that doesn’t involve politics.
Left-leaning brands like Ben & Jerry’s aren’t the only ones that don’t want their ads running next to pro-Nazi content. A brand’s value proposition is political. It is a statement of what that brand stands for. Standing for one thing means ruling out other things.
Dove is for women and emphasizes body positivity and the idea that beauty isn’t restricted to fashion models. These are all political positions. Until about nine years ago, they were the kind of saccharine, vaguely empowering political positions that few people would bother objecting to. Because of its brand values, Dove didn’t run ads in lad’s mags like Maxim or FHM, which are the antithesis of those values.
Under the FTC’s consent decree with Omnicom, the ad agency would be at risk of legal penalties if it couldn’t show the decision was made for non-ideological reasons. In this case, readership demographics seem like a surefire defense, but with the current trend of precedent-free legal decisions, who the [expletive deleted] knows? And, if the government is willing to go after a mega-company like Omnicom, it’s willing to go after anybody.
Once upon a time, companies could spend their advertising dollars wherever they wanted. The only check on it was the consumers’ opinion. That is how free markets are supposed to work. Once upon a time, Republicans were in favor of that.
If the struggle to get yourself a Switch 2 is reminding you of that year you spend trying to get a PS5, fear not because drops are happening all the time. And the latest one is coming at 1:30pm ET at antonline. With many retailers showing as out of stock, including Walmart and BestBuy, this will come as welcome news for those hunting Nintendo’s latest console.
When we reviewed the Nintendo Switch 2, we were blown away by its design and thought it an excellent upgrade on its predecessor. But we didn’t love the launch game line-up (including Mario Kart) so we are looking forward to seeing what comes out later. See more below on the drop from antonline.
Today’s best Switch deals
Boost your skills with Growth Memo’s weekly expert insights. Subscribe for free!
Today’s Memo is all about query fan-out – a foundational concept behind AI Mode that’s quietly rewriting the rules of SEO.
You’ve probably heard the term. Maybe you’ve seen it in Google’s AI Mode announcement, Aleyda Solis’ write-up, or Mike King’s deep dive.
But why is it really that revolutionary? And how does it impact the way we approach search strategy going forward? You might already be “optimizing” for it and not even be aware!
That’s what we’re digging into today.
Plus: I’ve built an intent classifier tool for premium subscribers to help you group prompts and questions by user intent in seconds – coming later this week (still need to iron out a few kinks).
In this issue, we’ll cover:
Let’s get into it.
Image Credit: Kevin Indig
Query fan-out is how Google’s AI Mode takes a single search and expands it into many related questions behind the scenes.
It can pull in a wider range of content that might answer more of your true intent, not just your exact words.
You’re hearing about it now because Google’s new AI Overviews and “AI Mode” rely on this process, which could change what content shows up in “search” results.
Query fan-out isn’t just another marketing buzzword. It’s how AI Mode works.
It’s crucial to start understanding this concept because it’s very likely that AI Mode will become the default search experience over the next few years. (I expect it will be once Google figures out how to monetize it appropriately.)
This is why I think AI Mode could become the search standard:
On the Lex Fridman podcast, Sundar Pichai said AI Mode will slowly creep more into the main search experience:
Lex Fridman: “Do you see a trajectory in the possible future where AI Mode completely replaces the 10 blue links plus AI Overview?”
Sundar Pichai: “Our current plan is AI Mode is going to be there as a separate tab for people who really want to experience that, but it’s not yet at the level there, our main search pages. But as features work, we will keep migrating it to the main page, and so you can view it as a continuum.”
He also said that pointing at the web is a main design principle:
Lex Fridman: “And the idea that AI mode will still take you to the web to human-created web?”
Sundar Pichai: “Yes, that’s going to be a core design principle for us.”
However, if AI Overviews are any indication, you shouldn’t expect much traffic to come through AI Mode results. CTR losses can top 50%.
And according to Semrush and Ahrefs, ~15% of queries show AI Overviews.
But the actual number is likely much higher, since we’re not accounting for the ultra-long-tail, conversational-style prompts that searchers are using more and more.
Even though AI Mode covers only a bit over 1% of queries right now – as mentioned in The New Normal – it’s likely going to be the natural extension of every AI Overview.
Important note here: I don’t want to pretend that I know how to “optimize” for query fan-out.
And query fan-out is a concept, not a practice or tactic for optimization.
With that in mind, understanding how query fan-out works is important because people are using longer prompts to conversationally search.
And therefore, in conversational search, a single prompt covers many user intents.
Let’s take a look at this example from Deep SEO:
Deep Search performs tens to hundreds of searches to compile a report. I’ve tried prompts for purchase decisions. When I asked for “the best hybrid family car with 7 seats in the price range of $50,000 to $80,000”, Deep Research browsed through 41 search results and reasoned its way through the content.
[…]
The report took 10 minutes to put together but probably saved a human hours of research and at least 41 clicks. Clicks that could’ve gone to Google ads.
In my search for a hybrid family car, the Deep Search function understood multiple search journeys, multiple intents, and synthesized what would have been multiple pages of classic SEO results into one piece of content.
And check out this example from Google’s own marketing material:
Image Credit: Kevin Indig
This Deep Search kicked off four searches, but I’ve seen examples of 30 and more.
This is exactly why understanding query fan-out is important.
AI-based conversational search is no longer matching a single query to a single result.
It’s fanning out into dozens of related searches, intents, and content types to synthesize an answer that bypasses traditional SEO pathways entirely.
Here’s my understanding of how query fan-out works based on the wonderful research by Mike King, as well as Google’s announcement and documentation:
Keep in mind, entities are the backbone of how Google understands and expands meaning. And they’re central to how query fan out works.
Take a query like “how to reduce anxiety naturally.” Google doesn’t just match this phrase to pages with that exact wording.
Instead, it identifies entities like “anxiety,” “natural remedies,” “sleep,” “exercise,” and “diet.”
From there, query fan-out kicks in and might generate related sub-queries, refining based on prior searches of the user:
These aren’t just keyword rewrites. They’re semantically and contextually related ideas built from known entities and their relationships.
So, if your content doesn’t go beyond the primary query to cover supporting entity relationships, you risk being invisible in the new AI-driven SERP.
Entity coverage is what enables your content to show up across that full semantic spread.
Here’s a good way to visualize this is the relationship between questions, topics, and entity expansion (from alsoasked.com):
Image Credit: Kevin Indig
If this all reminds you strongly of the concept of user intent, your instincts are well-tuned.
Even though query fan-out sounds cool and looks innovative, there is little difference to how we should already be targeting topics instead of keywords via entity-rich content. (And we all should’ve been doing this for a while now.)
Interjection from Amanda here: I’d argue that this kind of process (or a similar one) has been going on behind the scenes in classic SEO results for a while … although, unfortunately, I don’t have concrete proof. Just strong pattern recognition from spending way too much time in the SERPs testing things out. 😆
Back in 2018-2019, I noticed this pattern happening often with in-depth, entity-rich content pieces ranking – and performing well – for multiple related intents in search. The more entity-rich a content piece was, and the more the content tackled the “next natural need” of the searcher, the more engagement + dwell time increased while also concluding the search journey…
And the more the content did those things, the more the content was visible to our target audience in classic rankings … and the longer it held that visibility or ranking despite algorithm changes or competitor content updates.
When you keep query fan-out in mind, there are a few practical steps you can take to shape your content and optimization work more effectively.
But before you give it a scan, I need to reiterate what was mentioned earlier: I’m not going to claim I have a clear-cut way to “optimize” for Google’s AI Mode query fan-out process – it’s just too new.
Instead, this list will help you optimize your content ecosystem to fully address the multifaceted needs behind your target user’s search goal.
Because optimizing for conversational search starts with one simple shift: addressing searcher needs from multiple angles and making sure they can find those multiple angles across your site … not just one query at a time.
1. Passage-first authoring.
2. Semantically-rich headings.
3. Outbound credibility hooks.
4. Clustered architecture.
5. Contextual jump links (“fraggles” or “anchor links”).
6. Freshness pings.
Google’s AI Mode and the query fan-out process mirror how humans think – breaking a question into parts and piecing together the best information to solve a need.
People don’t search in a silo – when they search, they’re searching from a perspective, a history, and with emotions and multiple questions/concerns attached.
But as an industry, we have long focused on single queries, intents, or topic clusters to guide our optimization. Sure, this is useful, but it’s a narrow lens.
And it overlooks the bigger picture: Optimizing your content ecosystem to fully address the broader, multi-faceted needs behind a person’s goal.
We know Google’s AI Mode draws from:
So, here’s my step-by-step (unproven) concept:
Here’s a step-by-step guide:
Not only do paid subscribers get more content, more data, and more insights, but they also get the intent classifier tool I built to help save you some time on this work I’ve listed out above (coming to premium subscribers later this week).
If you’ve been doing SEO pre-AI-search era, it’s likely you’ve already been doing some version of this work.
The key thing to remember is to group questions and queries by intent – and optimize for intents across your core topics.
Think through what would’ve been a “search journey” or “content journey” for your user in classic search, and recognize that’s now happening all at once in one chat session.
The biggest mindset shift you’ll likely need to make is thinking about queries as prompts vs. searches.
And those prompts? They’re inputted by users in a variety of ways or semantic structures. That’s why an understanding of entities plays a key part.
But before you jump, I need to emphasize a core factor to creating content with query fan-out in mind: Make sure you do the work to take your collected questions that you plan on targeting and group them by intent.
This is a crucial first step.
To help you do that, I’ve created an intent classifier tool that premium subscribers will get in their inbox later this week. It’s simple to use, and you can drop your collected list of questions to group by intent in a matter of minutes.
Featured Image: Paulo Bobita/Search Engine Journal
Omnicom and Interpublic Group’s (IPG) $13.5 billion super-merger took a giant leap toward completion this week, following a qualified approval granted by the Federal Trade Commission (FTC).
The conditions attached to the FTC’s green light are far from normal, however. They constrain the combined holding company’s ability to steer clients away from publishers or media environments that might compromise their reputations. The agreement grants the U.S. government unprecedented power over the flow of advertising dollars to publishers, according to industry observers.
“It gives the [Trump] administration… and future administrations, for a 10 year period, a vehicle through which they can decide at an industrial scale, [that] the largest U.S. media agency will fund or not fund media owners on behalf of their clients,” said Ebiquity CEO Ruben Schreurs.
Taken alongside the FTC’s ongoing legal tussle with lobby group Media Matters and Congress’ recent scrutiny of the digital advertising sector, Forrester analyst Jay Pattisall told Digiday that the move “represents an unprecedented politicization of free speech and commercial creativity.”
Both companies have agreed to the conditions of a “consent order,” a legal mechanism usually deployed to force companies to divest from business units to avoid creating a monopoly.
But this order, which remains active for the next decade, prevents Omnicom from directing advertising dollars away from media owners that publish “ideological” content, unless brand clients specifically ask it to. The company is obliged to provide the FTC with an annual report detailing its compliance with the order for the next five years. The intention is to stop a post-merger Omnicom using its enlarged position in the market for political purposes.
While executives at both holding company welcomed the decision, the FTC’s imposed neutrality could end up costing the combined entity and weighing on its clients.
In a statement, Omnicom boss John Wren said it was “an important step toward the completion of the proposed acquisition and creating a new era.” It’s also an unusual step for the FTC, one that conjures up a series of knotty problems for the U.S. advertising industry.
Let’s take each one in turn.
The FTC’s order stops Omnicom from directing client ad dollars based on the “political or ideological viewpoints” hosted by a given publisher, and it won’t be able to use exclusion lists — either its own or those provided by companies like DoubleVerify or Integral Ad Science — unless an advertiser asks it to do so.
Pattisall said that caveat means the order won’t touch most media investment decisions as they play out in reality. “Nearly all media plans and buys are decided by advertisers. Clients are the ultimate decision-authorities for buying media, not agencies, which act as advisors. The consent order ignores how advertising works,” he said in an email.
The order shouldn’t affect “inclusion lists” — databases containing publishers that have been stamped for pre-approval by a client’s team, and which are considered the gold standard approach to brand safety among most media buyers.
“Brands are the ultimate decision makers, as they should be, about where their dollars get spent,” said eMarketer analyst Jeremy Goldman.
One snag? The FTC’s order doesn’t offer a definition of ideological or political publishing. That vagueness serves a purpose, given the FTC’s stated antipathy toward advertisers or agencies exerting power over publishers and social media platforms — and how that could change from administration to administration.
“The FTC is not happy that anybody has any degree of ability to unionize media buying. They don’t want to see any more consolidation of that,” said Goldman.
Potentially. The FTC’s request to document each choice to exclude a publisher from a media plan might not be a major headache for an organization this size, but it’s red tape they could do without.
“There’s no doubt going to be a level of legal scrutiny that has to go on those reports, which is costly,” said Gartner analyst Andrew Frank.
More importantly, cautious clients that want to avoid running ads against political media without catching heat from the Feds can now look to Omnicom’s rivals — think WPP Media, Publicis, Stagwell, Havas or Dentsu — for a means to do so. The order “undermines their [Omnicom’s] competitive position,” said Schreurs.
Pattisall noted: “Given that all the other global media agencies are under international ownership, the FTC may have unwittingly handed London, Paris and Tokyo-based agency holding companies a short-term win.”
That said, the biggest agency reviews rarely come down to a single factor. “If you’re talking about a major multi-billion dollar account changing hands, then you’re judging it on like 80 different parameters,” Goldman said, adding that he was “hard pressed to imagine a situation where this is the tie breaker.”
The consent order only binds Omnicom, but that doesn’t mean its holding company peers are beyond the FTC’s sights — the agreement is a signal for them to step in line, too.
“It’s almost like [pulling] over one person on the highway because everybody else is going to see it,” said Goldman. “It’s a very public show.”
It means Omnicom won’t be able to use a boilerplate approach. Now it’ll need to show that when its clients choose to exclude a publisher, they’re doing so for their own unique reasons. Accounting for those decisions in a way that satisfies the FTC will generate mountains of paperwork (it’s unclear as yet how the FTC plans to audit those reports, or what response non-compliance would provoke) but it’s the way the wind is already blowing.
“This is already being done on a brand-by-brand, client-by client basis. But it will now be the norm,” said Pattisall.
It’ll likely influence other holding companies to take a similar line, he noted. “Dentsu, Havas, Publicis and WPP Media… will adopt the same tactics to stay out of the crosshairs.”
On the face of it, there’s a positive angle here for news publishers. In recent years they’ve been caught out by blocklists that mislabelled legitimate reporting on important topics as unsuitable for brand advertising. Although some brands will continue to avoid running their ads on The New York Times’ or The Guardian’s websites by explicit choice, this decision means the agencies can’t cut them out from the word go — right?
Not necessarily. The FTC’s consent order cuts both ways. Omnicom can neither direct ad dollars away from a publisher for political reasons — or toward it. That means there’s a chance that news and current affairs publishers still end up losing out, if their coverage is deemed to hold an ideological bent.
“It does seem on the face of it, that it makes it harder to monetize news content,” said Frank. How the letter of the FTC’s order is translated into reality — today, five years and 10 years from now — is up in the air.
“It is totally dependent on how aggressively the regulators want to interpret,” added Frank.
“Stop-motion takes a long time to shoot, even when it’s quite simple,” says Bart. “So when it’s super complex, it’s a constant battle to try and figure out shortcuts and tricks, to find ways to get things done quicker.” Blinkink were able to make the laborious process slightly easier by sourcing stars from UK’s unique claymation industry such as Wes Anderson collaborator Andy Biddle to Andy Spradbury who made the puppets for Aardman Studio’s Chicken Run. “The puppets’ faces were also a real challenge to animate in clay” says Rebecca. “They were incredibly tiny and detailed so every time a player needed to have a change in expression it would take hours for the sculptors to resculpt the faces.” Centering the names of this year’s upcoming footballer stars, it was important that the stars of the advert were recognisable as well as their visual metaphors, each athlete’s signature skills communicated through fireball goals and melted faces from pure velocity. “Through this campaign and the BBC’s coverage of the Women’s Euro, we hope to help close the gender fame gap and make these amazing female athletes household names,” says BBC Creative’s creative director, Jess Oudot.
“In a time when AI is reshaping so much of the creative world, we wanted to go in the opposite direction – producing something tactile, raw and handcrafted,” says Bart. “It’s physical, full of character and energy – in many ways, just like football.” When creating the advert, the Aitana Bonmatí sequence really stood out. A flurry of animation techniques reflect Bonmatí’s own technical skills in a segment that sees opponents left twisted in their own clay-limbs from the footballer’s sharp turns and pro footwork. “We had to make a special head mounted camera rig to shoot the reference clip and then find a way to rig the clay puppet’s head to the stop motion camera when we were filming it,” says Rebecca. The ad secures Blinkink’s reputation as the absolute top of its game, along with the star-studded line-up of this year’s biggest event in women’s sports.
OpenAI CEO Sam Altman says ChatGPT has moved beyond being a Google alternative. Instead, the platform is increasingly focused on helping users complete complex tasks and workflows.
Speaking at Y Combinator, Altman described how ChatGPT has changed over time.
“For a long time ChatGPT was like a Google replacement… it still felt like a more advanced version of search.”
Now, he said, users can ask the AI to perform complex work like a junior employee.
“You can really give a task to code interpreter for example or to deep research… and come back to you with like a proposal.”
This shift signals a new direction for ChatGPT that could affect how businesses and marketers use AI.
Altman emphasized that ChatGPT is no longer just about retrieving information. The goal now is to help users get work done.
Altman said:
“It’s like a very junior employee that can work on something for like a short period of time.”
While the platform gets considerable traffic, Altman said ChatGPT.com is now the fifth most visited site in the world, he downplayed the idea that it’s competing with Google Search.
Instead, OpenAI is building a tool that can connect to user data, complete tasks, and act proactively.
A step toward this vision is ChatGPT’s memory feature. Altman called it his favorite feature so far this year.
This lets the AI remember previous conversations and user preferences, acting more like a personal assistant than a chatbot.
“I think memory is the first time where people can sort of see that coming.”
Altman described a future where the assistant knows when to notify users and when to take action automatically.
New models like GPT-4o and O3 are designed to handle more complex reasoning and workflows.
“Right now we’re in an interesting time where the product overhang relative to what the models are capable of is here…”
Altman said the technology is moving faster than most businesses can adapt to it. He sees untapped potential in how AI could support work like marketing, data analysis, and content development.
While Altman outlined an ambitious vision, there’s reason to be cautious.
Tools like ChatGPT face limitations like hallucinated outputs, lack of persistent memory across all contexts, and occasional reasoning failures. This is all detailed in OpenAI’s own reports.
That means, even with tools like Code Interpreter or GPT-4o, complex tasks still require hands-on oversight.
The shift away from search competition may also reflect the difficulty of challenging Google’s market dominance. Instead, OpenAI may be trying to define a new space for AI-powered task automation.
As AI tools like ChatGPT gain new features, they may change how marketers, developers, and everyday users complete tasks.
However, much of this vision depends on overcoming current limitations and delivering reliable performance across different use cases.
Altman shared that ChatGPT will soon support the Model Context Protocol (MCP). This allows it to pull data directly from tools and platforms businesses already use.
These integrations further support ChatGPT’s positioning as an assistant platform rather than a search engine alternative.
For now, marketers should focus on utilizing AI tools alongside, not instead of, traditional platforms like search engines. The two can serve different purposes in the same strategy.
Listen to the full interview with Sam Altman below:
Weather can be dramatic enough, what with tornadoes, thunderstorms, flooding, high winds, no winds, too much rain, too little rain.
For years, one constant has been the local meteorologists who have kept viewers informed about what is happening with the weather and how the community can stay safe when extreme weather hits.
But in the chaos known as modern times, media groups have attempted to hub weather forecasts to save money, the Trump administration has paring back core weather forecasting services like the NWS and NOAA, which makes chatting about the weather more than just small talk.
Madison, Wisconsin news outlet Madison Magazine writer Doug Moe took a look at the effects of changing weather staffing in the Madison market and how the local forecast is much more than just telling people whether they can picnic on the weekends. It often comes down to viewer safety.
“When that relationship is severed,” said Alex Harrington (pictured) chief meteorologist at WISC/Channel 3 (News 3 Now/Channel 3000). “You have community members whose lives are at risk because they don’t have that local connection.”
“The television weathercaster has a unique level of trust,” former KMGH and WKOW meteorologist Mike Nelson said. “At the scariest time for them, we are the ones saying, ‘Here is what is going on.’ ”
Click here to read his article.
There’s a lot of buzz around Amazon Prime Day at the moment, as retailers get prepped for one of the biggest sales events of the year. We’ve already started tracking the best Prime Day 3D printer offers, but actually, I think Bambu Lab’s 3rd Anniversary Sale is a much bigger deal (with up to 40% off), and definitely worth getting excited about.
Bambu Lab manufactures some of the best 3D printers on the market, and my fave 3D printer – the Bambu Lab X1 Carbon – has an amazing discount right now in both the US and UK. It’s reduced to only $999 / £829 for the printer with no AMS, or you can bag the full package for $1,249 / £1,049 – which is a big chunk off the usual retail price.
