Print to Pixel to Prompt: 500 Years of Marketing Media

Marketing has had four media revolutions in five hundred years. Print, broadcast, pixel, prompt. Every one of them was called a productivity upgrade by the people about to be displaced by it, and a new medium by the people who were about to displace them. The pattern is so consistent it is almost embarrassing. It is also, right now, the single most useful lens for understanding what AI is doing to UK marketing — and what to do about it.

1. Print (c. 1450 — c. 1920)

Before Gutenberg, a marketing message was a town crier and a hand-painted shop sign. After Gutenberg, a marketing message could be reproduced at roughly one tenth the cost per impression, and — crucially — it could leave the room. The handbill, the pamphlet, the trade catalogue, the newspaper advertisement, the direct mail piece: every one of them is downstream of moveable type. The print era ran for about 470 years, which by historical standards is a long time for any medium to dominate.

What did the print era actually change? Three things, and they will rhyme with every revolution that follows. It collapsed the cost of reproduction. It created a new craft elite — the typesetter, the engraver, the copywriter, the editor. And it rewrote the definition of a 'good' marketing message: from memorable spoken claim to scannable, structured, persuasive text. By 1900, the David Ogilvys of their day were not on Madison Avenue. They were at Sears Roebuck, writing 1,400-page mail order catalogues that did, on a per-pound basis, more retail conversion than any single marketing artefact since.

2. Broadcast (c. 1920 — c. 1995)

Radio in the 1920s, then television in the 1950s, did to print what print did to the town crier. The cost of reaching a mind collapsed by another order of magnitude. Suddenly one message could reach millions of households at once, in motion, with sound. The craft elite changed again — copywriter and art director replaced typesetter and engraver. The definition of good changed again — from persuasive text to memorable thirty-second narrative. Mad Men is the cultural memory of this era's apex, which is roughly 1955–1975.

Two things to notice about the broadcast era. First, print did not die. It shrank, and it specialised. The 1980 newspaper was a smaller, more focused product than the 1920 newspaper, but it was still a serious business. The lesson — every media revolution compresses the previous medium, it does not erase it — holds in every cycle including the current one. Second, broadcast was genuinely a new medium, not a new channel. It demanded a different job to make a thirty-second TV ad than to make a magazine spread. The print houses that tried to make TV ads using the magazine operating model lost. The ones that hired new specialists, with new craft, won.

3. Pixel (c. 1995 — c. 2023)

Mosaic shipped in 1993. The first banner ad ran on HotWired in 1994. AdWords launched in 2000. Facebook opened to advertisers in 2007. The iPhone shipped the same year. By 2015 more than half of UK media spend was digital, and by 2023 it was over 75%. The pixel era was the third great compression of the cost of producing and delivering a marketing message. It also did something the previous two eras did not: it made the message individually addressable. For the first time in 500 years, two different people could see two different ads from the same brand at the same moment, chosen by a machine, optimised for likelihood of response.

The pixel era killed the broadcast craft elite the same way broadcast killed the print one. The Madison Avenue creative director who could not brief a programmatic media buy was gracefully retired. A new craft elite — performance marketer, growth lead, paid social specialist, SEO strategist, CRO analyst — took the chair. The definition of good changed again: from memorable narrative to measurable conversion. The thirty-second TV ad gave way to the six-second pre-roll, the carousel, the retargeting funnel, the welcome flow. Mad Men became MathMen.

4. Prompt (c. 2023 — )

ChatGPT shipped in November 2022. By mid-2024 it was operationally mature for the bread-and-butter production work of UK marketing teams: brief decomposition, copy first drafts, content variants, ad creative iteration, landing page wireframes, audience research, analysis summaries. By 2026 it had been joined by purpose-built production agents, embedded AI inside every meaningful martech platform, and AI-mediated discovery surfaces — Google AI Overviews, ChatGPT, Perplexity, Claude, Gemini — that already account for a measurable and growing share of UK consumer information journeys.

Two things make the prompt era different from the three that preceded it. The first is that the cost compression is at the production end, not the distribution end. The previous three revolutions made it cheaper to deliver a message. This one makes it cheaper — almost free — to produce one. That is a different shape of disruption, and it lands hardest on the parts of the marketing org chart whose primary deliverable is first-pass production. We documented the labour-market consequences of this in our UK AI Marketing Jobs Tracker: junior copywriter postings down 44% year-on-year; junior SEO writer down 41%; media trafficker down 39%. None of those people are bad at their jobs. The job has been compressed.

The second difference is more consequential. For the first time in 500 years, the medium itself reasons. A printed page does not know who is reading it. A TV ad does not know who is watching it. A pixel ad knows who you are but cannot reason about you — it can only match you against pre-built segments. An AI assistant can read the brand's positioning, read the user's stated need, reason about the gap, and produce a personalised, persuasive answer in one pass. That is not a faster pixel. That is a different medium.

5. The losers, in every era, are the same people

The pattern across all four revolutions is so consistent it borders on a law. The losers are not the practitioners of the old medium. The losers are the practitioners who tried to do the new medium using the old job description. The print houses that lost in the 1920s were the ones that produced radio ads as if they were newspaper ads with sound. The agencies that lost in the 1990s were the ones that produced banner ads as if they were small TV spots. The agencies losing right now are the ones producing AI work as if it were faster pixel work.

The corollary is also consistent. The winners are the ones who ask, before doing anything else, what the new medium actually is and what new craft it demands. We unpack what that craft looks like for the prompt era — and what the new agency operating model behind it has to be — in The Mad Men to Machine Men Shift and in our pillar essay on The Death of the Marketing Agency (As We Know It).

6. What survives every revolution

Three things have survived all four media revolutions, and they are the only three things any UK marketer should be quietly compounding right now. First, distribution leverage. Whoever controls the surface where attention is allocated wins. In the print era it was the printer and the postal network. In the broadcast era it was the network and the agency holding the buying weight. In the pixel era it was Google and Meta. In the prompt era it is increasingly the AI assistant — which is why we wrote Who Owns the Customer in an AI World?

Second, brand memory. Every revolution compresses production, and every revolution therefore inflates the value of the one thing that cannot be produced at zero marginal cost: a recognised, trusted, distinctive brand. The case for brand investment in 2026 is structurally stronger than it has been since the 1960s, for the same reason it was strong then — the medium is doing the production, so the differentiator has to live in memory.

Third, taste at the edit stage. In the print era, anyone could buy a press; the people who edited the output well sold more books. In the broadcast era, anyone could buy airtime; the people who edited a thirty-second narrative well sold more product. In the pixel era, anyone could run an ad account; the people who edited the funnel well sold more units. In the prompt era, anyone can produce a thousand decent ideas before lunch; the people who edit them well will sell more of whatever they sell. Editorial judgment is the single skill that has compounded across all four revolutions and is about to compound again.

7. What this means for the next ten years

If the historical cadence holds — and we have no reason to think it will not — the prompt era's first decade will look structurally similar to the first decade of the pixel era, 1995–2005. Expect a new craft elite to consolidate (we are watching it form: AI ops, GEO, causal measurement, AI-led creative direction). Expect the previous craft elite to compress, specialise and survive at a smaller footprint (the senior strategist, the brand director, the human creative lead are all stable and growing in our 2026 data). Expect platforms to consolidate around the new distribution surface (the AI assistants are doing this in real time). Expect a handful of category winners to be defined by 2030, mostly invisibly, mostly through compounding rather than launches.

The marketers and brands that internalise this arc — that recognise this is Gutenberg, not the iPhone — will spend the rest of the decade quietly building durable advantage while the rest of the industry argues about which AI tool to buy next. Five hundred years of pattern suggests the pattern will hold one more time.

Suggested citation: Javed, Kasim. Print to Pixel to Prompt: 500 Years of Marketing Media. Creative Marketing Group, April 2026. Available at creativemarketingltd.co.uk/print-to-pixel-to-prompt.