Online shrunk the media
Industry analysts hear mood music from the questions they are asked by insiders and outsiders. In 2024, outsiders want to know when (and why) journalism media will become investible again. Insiders want to know why subscription fatigue has hit the sector, after an impressive growth spurt from the first lockdown to late 2023 (by our estimates, more than 2.5m journalism subscriptions were secured in that period). Both insiders and outsiders ask about AI. The linking theme is not difficult to discern: where’s the growth?
Journalism has had a shrinking paying audience since the mid-1950s, when circulations peaked (commercial TV was born). But news businesses particularly thrived in the 20-year period from the mid-1980s, when print advertising exploded. Newspapers were one of the great consumer mousetraps, generating demand from advertisers and favourable relationships with agency suppliers. Newspaper cultures became commercial. Advertising sales leaders grew in stature, influence, swagger. Advertising became a proxy for commercial, in the broader business sense. Cover price rises offsetting declining volumes was consigned to humdrum operational background. The free Metro arrived in 1999.
On the wave of advertising success, came a technology and investment wave, as journalism became the anchor content of the internet. Online businesses and emboldened entrepreneurs filled the schedules of publishing executives. Content, distribution and advertising deals were done and editors were encouraged to chase traffic to maximise value. Further disruptive technology waves followed: search engine advertising, user-generated social media, social media advertising, ecommerce, the mobile internet. These waves continue to crash over the journalism sector, as a new one — Large Language Models (LLMs) — poses its own unique balance of jeopardy, opportunity, distraction and cost.
As many commentators have pointed out, mass media was a happy accident: expensive journalism was part of a profitable bundled package at a time when alternatives and distractions were fewer than today. Three decades have passed since the dawn of the consumer internet. The stubborn free-to-access, audience scale mantra has eschewed the establishment of an infrastructure and marketplace for online journalism. The highly effective distribution and merchandising collaboration that secured profits for the print era have not been replicated. Millions of media units delivered overnight to 45,000 retailers will become a thing of the past as we exit the industrial print era, and enter a new artisanal era for print (from around 2030 or thereabouts).
Meanwhile, we have learned what does and does not work online. If offline rewarded market generalists, online evidently rewards specialists: companies that do one thing extremely well. The hefty fixed costs of professional journalism don’t fit neatly with cyclical advertising. Online advertising alone will rarely fund a large independent newsroom. Journalism — a sector centred on the judgements of experienced people with good connections — will never have the digital network effects of search or social media. Besides, independent journalism should feel uneasy about an entirely advertising-funded solution.
The public generally pays for three benefits online: utility, entertainment and valued information that cannot be found elsewhere. Journalism often falls short: not quite useful enough, not quite entertaining enough. Most journalism does not create convenience or solve problems, or at least not reliably enough to compel payment. It rarely feels as unique and valued as — say — a football match or a tool to personally organise the world’s available music. The massive exceptions hint at the options: Which?, Autotrader, Good Housekeeping, the Financial Times, Good Food, Time Out.
AI and a new fork in the road
LLMs is a good moment at which to consider the journey media is on. Journalism is a human craft operating at various levels of independence from institutions and enterprises. It provides a vibrant balance of what people want to communicate and what people would rather was not communicated. It reports, contextualises and issues authoritative judgement. Journalism acts on behalf of citizens to help them make sense of the world, or to prioritise their time and expenditure. It champions consumers, cohorts, outcomes. It’s not anonymous. Using bylines and brands, it provides a highly transparent right-to-reply. It corrects errors. There are legal restrictions to what journalists can and can’t cover; there is a regulatory or self-regulatory regime; there are protocols (editorial codes) that journalists systematically honour. Journalists are paid salaries disconnected from audience or advertising revenues.
In sum, there are a variety of reasons why the craft is highly distinctive from user-generated content, or non-human-generated content. Yet the journalism sector has done relatively little over the last 25 years to make or reinforce a case to the general public, or to advertisers or to public institutions. At worst, it has actively downgraded its unique benefits. It has sometimes sought to employ or emulate the work of influencers or wilfully blended into the social media landscape, as if journalism’s motivations were nothing special. That turns the above advantages into onerous cost disadvantages. The error is perhaps understandable: titles such as the Sun or Grazia were the social media of their age.
Machines can now deliver journalism to drive traffic. The thinning distinction between commodity journalism and other online content has been surgically closed. Few disagree that fakes — deep and shallow — will proliferate in the online world. Yet the craft of human journalism is also enhanced by AI tools, and so its distinction from the plethora of online content should widen. AI will not replace journalists, but journalists with AI capabilities will become more valuable. More distinctive journalism services are more valuable, and not just for audiences. Distinctive services will also drive a licensing economy. AI licensing could evolve into a sustainable revenue stream, and revenue diversification is a crucial factor in the online age, reducing journalism’s reliance on advertising.
AI will not only support the external-facing layer (new journalism research opportunities, notably in data) but also an internal-facing layer (product and audience development). AI massively improves audience targeting, conversion and pricing. AI can help design the marketplace for a publisher and for journalism more broadly.
Finally, audience payments and advertising are often seen in tension. A premium advertising economy — measured for awareness and pricing power as well as click-throughs — is best built from distinctive services. So-called ‘brand safety’ (online content tagging) systematically de-values journalism by maliciously labelling content as (say) “death” rather than “important, serious, live reporting”. Why are many app adverts traded at humble rates, when they should be scarce, premium inventory? Incremental adjustments won’t overhaul the thinking that defines all these outcomes today. It will take the upheaval of industry reinvention.
In all, journalism as an online marketplace — for audiences and advertisers — remains a privileged opportunity for upmarket and specialist services and audiences, unless the industry defines and delivers a broader opportunity, and convinces the public to value it. The AI era provides the tools to do so.
Back to the future
Successful online audience economies by traditional news media have so far been secured by services that would be recognised by audiences in 1995. The Economist is not so different from the Economist.com or its apps, which is not to belittle the publisher’s exemplary innovations in podcasts, newsletters, daily publishing (and so on).
But native news media look different: Politico, the Athletic, Tortoise, Bellingcat, Puck, the Ferret, the Manchester Mill. They are narrower and more specialist in remit; they have explicit purpose; their formats and content hierarchy are highly distinctive; their tone is intimate; they blur the boundaries between consumer and trade (B2B) in terms of publishing techniques, targeting, revenue streams and business structure.
Future successful mass market media, specialist media and local media will have new formats, new services, new entry points, new charging mechanisms. Content packages were unbundled online, but media are now re-bundling — the New York Times with sport, food, games, for example. There will be new consumption patterns.
Publishers rightly worry if AI innovations will always affordably address real issues for them. Many won’t, and others might require a collaborative solution. Our research broadly argues an imaginative interpretation of how the collaborative print infrastructure might manifest itself online is long overdue. Being swept around by the blustery algorithmic winds of search and social is an unhappy place, and as recent studies estimate, new AI products could reduce referral traffic. Certainly, the value chain for news distribution is not the controlled and aligned one that publishers enjoyed for many decades. Online should bring journalism and audiences together; the present infrastructure seems more inclined to keep them apart. Platforms are crucial merchandising tools, a means to amplify — not restrict — the ambitions of journalism businesses.
In a product production sense, AI also completes a transition from the print manufacturing era to the online service era. Functional silos are the spirit of quality manufacturing, but the rhythm of service providers is deep, trusted, cross-functional collaboration to create compelling experiences that achieve impact and stimulate regular engagement. The commercial centre of this workflow is not advertising sales but product design — an editor — with decision-making criteria designed to maximise benefits. ‘Church and State’ is an anxious friction for journalism businesses that lack trust in a shared business purpose. That was a productive tension in the silo era, but today feels an unhealthy expression of mistrust among teammates.
Sustainable growth can only be secured in any industry by leaders who believe in a specific, distinctive value they can reliably create, merchandise and sell. The most welcome development in media today is the return of entrepreneurialism in journalism — an origin spirit, with editors and business partners defining communities and intuiting (and tracking) what is useful to, and what leverages, those communities. We are optimists. There are green shoots all around, from Mail+ among the traditional giants to the Mill among the native upstarts. The journalism revolution will be greater over the next five years than it has been over the last 25. AI will be but one — albeit crucial — driver on that journey.
This article was first published in InPublishing magazine. If you would like to be added to the free mailing list to receive the magazine, please register here.
