Why Every Marketing Revolution Eats Its Agencies

Every marketing revolution kills the agency model that served the previous medium. The pattern is so consistent across 150 years of industry history that it borders on a structural law. Print shops collapsed when broadcast arrived. Mid-century broadcast agencies collapsed when digital arrived. The mid-2010s cohort of social and SEO specialist shops is collapsing right now, in 2026, as the AI shift restructures the economics of every marketing service line. UK agency owners reading this have a binary choice in front of them, and the historical pattern says it is the same choice their predecessors faced in 1925, 1995 and 2010. Rebuild now, or be acquired in distress in five years.

1. 1925: the print shops

In 1925, the dominant US marketing services firm was a print house. They typeset, they engraved, they printed, they distributed — most of what 'marketing' meant in 1920 was executed by them. By 1945, the dominant marketing services firm was an advertising agency, and most print houses had either been absorbed, restructured into purely industrial printing, or quietly shut. The transition took twenty years and produced more agency closures, in the US alone, than any other twenty-year period until the 2020s. The print houses that survived were the ones that recognised, around 1928, that broadcast was a different medium requiring a different operating model — and rebuilt accordingly.

2. 1995: the broadcast agencies

In 1995, the dominant US and UK marketing services firms were Mad Men-era broadcast agencies. The largest agency in the world had revenues of around $1bn. The agency-of-record retainer was the dominant commercial model. Account services, planning, creative, production, media — the org chart was so standardised across firms that talent moved between them seamlessly. By 2015, half of those firms had restructured beyond recognition, a quarter had been absorbed, and a small handful had transitioned successfully into the digital era. WPP, Publicis, Omnicom and IPG survived through aggressive acquisition. Most of the firms they acquired did not survive on their own terms.

The 1995–2015 transition produced a generation of agency restructurings that are textbook studies in why operating-model redesign matters. The agencies that bought digital tooling and bolted it onto the 1990 operating model lost ground steadily. The agencies that built native digital operating models — Razorfish, Sapient, AKQA, R/GA — captured the surplus.

3. 2010: the SEO and social shops

The 2008–2014 era produced a wave of UK agency formation around two specialisms: SEO and social media. Both were rational at the time. Google and Facebook were demonstrably the dominant new distribution surfaces, and clients needed specialist help. Many of those firms grew quickly. Most of them are now gone in their original form, absorbed by larger firms, pivoted into adjacent disciplines, or quietly wound down. The reasons rhyme with every previous extinction event.

First, the platforms commoditised the specialist craft. By 2018, most of what SEO agencies had been selling in 2012 was either built into the platforms or available through tooling affordable to the client. Second, the in-house team got better. The clients who had been paying SEO agencies in 2012 had hired in-house SEO leads by 2018. Third, the dependency on a single platform created concentration risk that the holding companies eventually exploited through acquisition or replication. The social-shop story tracks the SEO-shop story almost exactly, just delayed by three years.

4. 2026: the AI restructuring begins

The fourth wave is in motion now and we have a structural view of how it will play out, drawn from running and rebuilding our own firm through it. The cohort under most acute pressure in 2026 is the mid-2010s digital agency that has not redesigned its operating model around AI. The signs of pressure are consistent: margin compression, talent attrition at the operator level, client churn to AI-native challengers, and pricing pressure on retainers that used to be sticky.

The cohort compounding fastest in 2026 is small, AI-native, operator-led firms with heavy senior weight, light delivery layers, AI production lines, and outcome-priced commercial models. Some of them are new. Some — like ours — are existing firms that have rebuilt the operating model from the inside. Both routes work. The route that does not work is layering AI tooling on top of the 2020 operating model.

5. What survives every restructuring

Three things survive every agency-side extinction event, and they are the only three things any UK agency owner should be compounding right now. First, senior strategic capacity — the ability to walk into a client's room and add value in a conversation about their business, not just their marketing. Second, brand authority — the firm's own market reputation, which compounds across decades and is almost the only durable asset on the agency-side P&L. Third, an operating model that matches the economics of the current medium. Get all three right and you survive every revolution. Get one wrong and you do not survive any.

6. Why holding companies are uniquely exposed

The holding company model — built on aggregating broadcast media buying weight in the 1980s and consolidating digital capability through acquisition in the 2000s — is structurally challenged by the no-cast era we describe in From Broadcast to Narrowcast to No-Cast. The aggregated media buying weight matters less when the buying is increasingly automated and the most important upper-funnel surface — AI-mediated discovery — cannot be bought at all. The holding companies will survive in some form, but expect significant internal restructuring through 2030. The early moves are already visible.

7. The 2026 forecast

Our central forecast for the UK agency landscape, based on internal panel data and the historical extinction pattern: roughly 40% of the 2024 UK agency landscape is, in its current form, not viable through 2030. Of those, roughly half will restructure successfully, roughly a quarter will be absorbed, and roughly a quarter will quietly close. A new cohort of AI-first operators — both new firms and rebuilt incumbents — will take a disproportionate share of the surviving market. We document the operating-model architecture in the pillar essay The Death of the Marketing Agency (As We Know It).

8. The honest message to UK agency owners

You have, on the historical pattern, three to five years to rebuild the operating model before the structural pressure becomes terminal. Most owners reading this will recognise the pressure already — margin compression, talent attrition, pricing pressure, client churn to AI-native challengers. The historical lesson is that the pressure does not reverse, it accelerates, and the firms that move first redesign from strength while the firms that wait redesign from distress. The 1925 print shops faced this. The 1995 broadcast agencies faced it. The 2010 SEO shops faced it. None of the firms that moved first regretted it. Most of the firms that waited did. The pattern is the pattern.

Suggested citation: Javed, Kasim. Why Every Marketing Revolution Eats Its Agencies. Creative Marketing Group, April 2026. Available at creativemarketingltd.co.uk/why-every-marketing-revolution-eats-its-agencies.