Arloesi CEO | Independent. Proven. Practitioner. | Transforming How Brands & Agencies Create Content in the Age of AI
Summary of “The Same Deck”
The article argues that the global advertising agency industry is in structural decline, even though major agency groups keep presenting similar “transformation” strategies.
The trigger: Dentsu’s failed sale
The article begins with a striking example:
Dentsu tried to sell its $4.5 billion international agency business, but no buyer wanted it.
Private equity firms and rival agencies walked away.
This signals that large global agency networks are no longer considered valuable growth assets.
“The Same Deck” problem
Jones argues that major advertising groups are all presenting nearly identical strategic plans, such as:
WPP – restructuring into a single operating company with AI at the centre.
Publicis Groupe – “Power of One” integrated platform model.
Omnicom Group – integrated model and consolidation.
Dentsu – “One Dentsu” strategy.
Despite different branding, the strategies share the same themes:
Integration of agencies
AI transformation
Cost savings
Becoming a “trusted growth partner.”
The author argues these are repeated PowerPoint narratives, not genuine solutions.
3️⃣ The real problem: shrinking economics
Financial results show a different reality:
Declining revenues
Massive goodwill write-downs
Thousands of layoffs
Falling stock valuations
For example, WPP’s valuation has dropped dramatically, while profits and headcount have fallen significantly.
The author says cost-cutting disguised as “transformation” actually signals contraction, not growth.
4️⃣ Why the agency model is breaking
According to the article, agencies are being squeezed by two structural forces:
Platforms above them
Companies like:
- Meta Platforms
- Amazon
are building advertising systems that allow brands to run campaigns directly without agencies.
Automation below them
- AI tools can now automate tasks agencies once charged for, including:
- content production
- campaign optimisation
- localisation and adaptation.
Much of the execution work that generated agency revenue is becoming automated.
What might survive
The author believes agencies won’t disappear completely, but they will shrink dramatically.
The parts that remain valuable are:
- strategic brand thinking
- creative direction
- cultural insight.
These activities represent only a small portion of agency revenue today, meaning the industry structure will have to change significantly.
Core takeaway
The article’s central message:
Many advertising holding companies are presenting identical transformation strategies, but the underlying business model of large global agency networks is being undermined by AI, platforms, and changing client behaviour.
In short:
The decks change logos, but the outcome remains the same.
------------------------------
In January, Dentsu tried to sell its international business. A $4.5 billion revenue operation. Agencies, media, data, creative, production. Thirty countries. The full stack.
Nobody wanted it.
Not rival agency groups. Not private equity firm Apollo. Bain Capital stuck around longest, reportedly with "significant reservations." By February, the process had collapsed entirely. Dentsu posted a record loss of ¥327.6 billion ($2.18 billion), wrote down ¥310 billion in goodwill on its international operations, suspended its dividend, sold its historic Tokyo headquarters, and overhauled its leadership.
Let me say that again. A $4.5 billion revenue agency network was put on the open market, and nobody would write the cheque. Not trade buyers. Not private equity. Not at any price that made sense.
That should have been the biggest story in advertising this year.
It wasn't. Because everyone was too busy rehearsing the same PowerPoint.
The PowerPoint
On February 26th, WPP announced "Elevate28." A three-year strategic plan to move from holding company to single operating company. Four divisions (Media, Creative, Production, Enterprise Solutions). Four regions. AI at the centre. £500 million in annual cost savings by 2028.
Cindy Rose stood up and declared WPP was "no longer a holding company." Its new mission: to become "the trusted growth partner for the world's leading brands in the era of AI."
Sound familiar?
It should. Because I've now read this deck four times. With four different logos.
Publicis: "Trusted growth partner. Power of One. AI-driven platform company." Omnicom (post-IPG): "Trusted growth partner. Integrated model. Unified operating company." Dentsu: "Strongest growth partner. One Dentsu. AI and technology at the core." WPP: "Trusted growth partner. Single company. AI-enabled solutions."
Same consultants. Same three-phase plan (stabilise, build, accelerate). Same promise that individual agency brands will be preserved. (They won't. I watched it happen to the media brands. Then to the production brands. The creative brands are next.) Same confident assertion that "enterprise solutions" will unlock new revenue pools. Same insistence that AI is a growth driver, not the thing that's eating the business model from the inside out.
I've sat in the rooms where these decks were presented. I helped build one of the businesses that got "integrated" in the last round. The language changes. The logos change. The outcome doesn't.
The Math That Breaks the Narrative
Let's leave the strategy language aside and look at what WPP actually reported today.
Revenue down 8.1% to £13.55 billion. Revenue less pass-through costs (the metric the industry actually watches) down 5.4% like-for-like. Q4 alone: down 6.9%. Operating profit collapsed 71% to £382 million after £641 million in goodwill impairments. The dividend was slashed 62%. Headcount fell by 9,400 in a single year to 98,655.
Nine years ago, WPP was valued at £24 billion. Today: under £3 billion. It fell out of the FTSE 100 in December. The share price dropped another 7% on results day, despite the most radical restructure in the company's 40-year history. Another in a recent line of 10-year lows.
The 2026 outlook? A further "mid to high single-digit" decline in the first half, with "an improving trajectory in the second half."
The improving trajectory is always in the second half. It's been in the second half for three years running.
Bloomberg Intelligence analyst Matthew Bloxham didn't mince words: the strategy "doesn't go far enough to deliver the course correction required to address concerns about the existential threat from AI."
And WPP isn't even the worst case. Dentsu's goodwill impairment on overseas operations was $2 billion. They're selling the furniture. Literally. Their Ginza headquarters, a building that symbolised the company's post-war rise, went for ¥30 billion. That's not strategy. That's estate clearance.
The Cost-Cutting Tell
Here is the thing that every trade publication dances around but nobody says directly.
When your transformation strategy is denominated in savings rather than revenue, you are not transforming. You are shrinking. And shrinking is not a strategy. It's a trajectory.
WPP: £500 million in savings. Omnicom/IPG: $1.5 billion in savings (they doubled the target; the market cheered, up 15%). Dentsu: ¥52 billion in severance, another ¥26 billion planned for 2026. WPP's headcount dropped 9,400 in one year. Omnicom/IPG went from 128,000 combined to a target of 105,000 since the merger was announced. Dentsu cut 3,400 internationally with 1,300 more planned.
Not one of these companies is investing its way into the future. They're cutting their way toward survival. The market knows the difference.
When Omnicom doubled its savings target, the stock jumped 15%. Not because investors believe in Omnicom's growth story. Because the market is rewarding capital return, not capital deployment. The financial equivalent of praising a patient for losing weight when the reason is they're ill.
Now ask yourself: where does the new revenue come from? Not savings. Revenue. Net new money from clients for services they can't get anywhere else.
Not one of these strategies has a credible answer to that question.
The Balance Sheet Nobody's Discussing
There's another dimension to the brand-merging strategy that nobody in the trade press seems to want to touch.
Every holding company's balance sheet carries billions in goodwill. These are the acquisition premiums paid over decades. Ogilvy. JWT. Young & Rubicam. Grey. AKQA. Each time a holding company paid more than an agency was worth on paper, the difference went onto the balance sheet as an intangible asset. A legal assertion that these brands and client relationships would generate enough future cash flow to justify what was paid.
WPP still carries billions in goodwill and intangible assets on its balance sheet. At the end of 2023, that figure was north of £7 billion. Impairments and disposals have brought it down, but even after the writedowns, goodwill alone sits around £5 billion. The entire market capitalisation of WPP is now under £3 billion.
Read that again. The market is saying that the goodwill on WPP's books, representing decades of acquisition premiums, exceeds the value the stock market places on the entire company.
And here's the problem with the "merge everything" strategy. Every time you fold brands together, you weaken the accounting basis for the goodwill attached to them. WPP merged JWT and Wunderman Thompson into VML and took accelerated amortisation charges of £633 million on those brands. Created Burson from BCW and Hill & Knowlton. More restructuring. More value dissolved. This year: £641 million in goodwill impairments, on top of £237 million last year (mainly AKQA).
That's nearly £900 million in goodwill impairments alone in the last two years, on top of over £600 million in brand amortisation charges when VML was created. WPP Creative is now bringing Ogilvy, VML, AKQA, Grey, Burson, and Landor under one umbrella. Each merger makes separate brand valuations harder to defend. Each cost cut compresses the margin forecasts that support the impairment models. WPP's own sensitivity analysis at the half-year was telling: if Ogilvy's margins dropped just 1%, that's a £68 million impairment charge. AKQA was even more exposed.
This isn't unique to WPP. Dentsu just wrote down $2 billion. Omnicom is about to integrate IPG's entire agency portfolio, with billions in combined goodwill that will need testing against a unified cash-generating unit. The pattern is the same everywhere: merge the brands, take the writedown, repeat.
Let me translate from Accounting to English: when you can't defend the value of the individual parts, you stop reporting them individually.
The Reclassification, Revisited
Three weeks ago, I wrote about what happened when Publicis posted the best results in its 100-year history. Record 18.2% operating margin. 5.6% organic growth. Seventh consecutive year of outperformance. The stock fell 9% the same day.
Not because the numbers were bad. Because Anthropic shipped some plugins over the weekend, and the market sold Publicis alongside ServiceNow, Salesforce, and Thomson Reuters.
Let that sink in. The market didn't sell Publicis with the other agencies. It sold Publicis with the software companies. The stock market has reclassified advertising holding companies as software businesses with a labour cost problem, not creative businesses with a technology opportunity.
WPP's results today only confirmed it. The most radical restructure in WPP's history, announced on the same day as the worst results since the pandemic, and the market shrugged. Down 7%. Another 10-year low. The market has already priced in the outcome of Elevate28. Not because the plan is bad. Because the category has been repriced.
The Dentsu Signal
This is the part that should keep every holding company CEO awake tonight. And every CMO paying attention.
Dentsu put a $4.5 billion revenue business on the open market. Trade buyers looked and walked away. Private equity looked and walked away. As one FT analysis put it: the market is asking a colder question now. How much of that complexity is an advantage, and how much is simply overhead?
If a $4.5 billion agency network can't find a buyer at any viable price, what does that tell you about the implied value of every holding company's creative, media, and production operations?
It tells you what the stock market has been telling us for two years. International agency networks are not growth assets. They are declining revenue streams attached to expensive infrastructure, with margins under pressure from AI commoditisation, platform disintermediation, and client in-housing. You cannot financial-engineer your way out of secular decline. PE knows this. That's why they passed.
The Dentsu sale didn't fail because Dentsu is uniquely broken. It failed because the asset class has been repriced. Like trying to sell a newspaper chain in 2012. The buyer wasn't wrong about the price. The buyer was right about the future.
What They Should Be Saying But Can't
Every holding company is telling the same story: simplify, integrate, embed AI, emerge as a "trusted growth partner." But the honest version of that story, the one that no CEO will give on an investor call, goes something like this:
The traditional agency value chain is being compressed from both ends. From above, the platforms (Google, Meta, Amazon, Adobe) are building self-serve tools that eliminate the need for agencies in media buying, basic creative production, and campaign optimisation. I wrote about this in January, when Google's Personal Intelligence turned the agencies' trillion-dollar data partnerships into an expensive rearview mirror. The platforms don't approximate understanding. They have the actual data.
From below, AI is automating the production, adaptation, and localisation work that generated the bulk of agency revenue at scale. I've watched automation platforms reduce content variant production from weeks to hours. Not in demos. In live client programmes.
What's left in the middle is genuinely valuable: strategic counsel, creative direction, brand stewardship, orchestration of complex multi-market campaigns. But in my experience, across dozens of enterprise client relationships, that middle accounts for perhaps 15-20% of what clients currently pay holding companies for. Look at how any major retainer breaks down: the strategy and creative direction sit at the top, and everything below it, the production, adaptation, versioning, trafficking, localisation, is execution. The other 80% is execution. And execution is being automated, platformed, or insourced.
The holding companies can't say this. It would undermine their entire pitch. So instead they talk about being "AI-enabled" while quietly cutting the humans who do the work that AI is replacing.
The strategy decks say transformation. The P&Ls say managed decline.
What Actually Survives
I want to be clear: this isn't a eulogy. I don't enjoy writing about an industry I've spent three decades helping build. Some of what exists inside these holding companies is genuinely world-class. The best strategic thinkers. The best creative directors. The best data scientists. The people who can sit with a CMO and reshape how a brand goes to market in a world where the rules change quarterly. Those people are irreplaceable. They know it, which is why the talent exodus is accelerating.
What doesn't survive is the structure around them. The layers of management. The duplicated back offices across hundreds of agency brands that exist primarily to justify separate P&Ls and leadership teams. The holding company model was built for a world where complexity was the product. Clients paid agencies to navigate the complexity of media buying, production logistics, and market-by-market execution. AI is dissolving that complexity. When complexity dissolves, so does the premium you can charge for navigating it.
I wrote last year about the revenue per employee gap: tech companies at $2-3.6 million, agencies at $130-175K. A 20x differential. You don't close a 20x gap with a new org chart. You close it by fundamentally changing what you sell, who does the work, and how value is delivered. None of these strategies do that. They reorganise the existing model. They don't replace it.
Sorrell, who still can't resist a comment on the company he built, called the current approach "carnage" and accused WPP of "slamming brands together willy-nilly." He's not entirely wrong. But his critique misses the bigger picture. The problem isn't how the brands are being combined. The problem is that the model that justified having separate brands in the first place no longer exists.
The View From the Client Side
If you're a CMO reading this, you already know most of it. You've watched your agency partners announce transformation after transformation while your experience of working with them has barely changed. Multiple contacts across multiple brands for what should be a single conversation. Layers of overhead that add process but not value. Promises that AI will transform everything. Next year. Always next year.
The brands I work with who are getting this right aren't waiting for their holding company to figure it out. They're building their own intelligence layers, their own orchestration capability, their own brand-controlled AI infrastructure. They're getting specific about which parts of the agency relationship are strategic (worth paying a premium for) and which parts are operational (and should be automated, insourced, or contracted at commodity rates).
Here's the question I'd ask at your next QBR: What does my agency do that OpenAI or Google couldn't package as a feature by next quarter?
If the answer is "everything," they're not paying attention. If the answer is "nothing," they're not being honest. The real answer is somewhere in the middle. And the sooner both sides of the table get specific about where that line falls, the sooner we can stop pretending that a new org chart solves a business model problem.
February 2026 wasn't the month one holding company had a bad quarter. It was the month the market stopped pretending. Four companies. Four strategies. Same deck. Same outcome.
The question isn't whether the holding company model will change. It's whether the holding companies will be the ones doing the changing.
No comments:
Post a Comment