Is Search Traffic About to Be Cut in Half? How the Media Industry is adapting in the Age of AI

43%. That's how much media executives worldwide expect search engine traffic to fall over the next three years, according to the Reuters Institute report.

Google Search has long been the single largest traffic source for digital media — a river that never stopped flowing, feeding the ad revenue of thousands of news websites. But the arrival of ChatGPT, Google AI Overviews, and similar services has built a dam at the source. Users' questions get answered inside the AI chat window. The motivation to click through to a news site is disappearing fast.

The traffic crisis is already causing real damage:

• Business Insider announced a 21% headcount reduction in June last year, citing "uncontrollable, extreme traffic decline" as the direct cause
• The Washington Post cut approximately 30% of its staff in February this year, with financial pressure and traffic loss as the key drivers
• Reader behaviour has shifted profoundly — people now expect a chatbot to summarise the answer, and fewer are willing to sit with a full article

When even the top-tier outlets can't hold on, the entire industry faces an existential question: when information itself becomes cheap, what can actually bring readers to your site? 

The world's top media companies aren't sitting still. Their strategies differ, but the underlying logic is consistent: instead of mourning lost traffic, make your product genuinely irreplaceable.

🔹 Bloomberg — Refuse Licensing, Double Down on AI Internally

While most publishers rushed to sign licensing deals with AI companies, Bloomberg made the counterintuitive call: refuse to license content to AI companies, and instead use AI to improve the on-site experience.

• AI-powered on-site search redesign led to 15% growth in usage and click-through rates
• Editor-reviewed AI summaries added to article tops improved page views and time-on-site
• Subscriptions reached 707,000 users in 2025, up more than 10% year-on-year

🔹 Washington Post & Financial Times — Make AI Your Own Weapon

If AI chatbots are the competition stealing your traffic, why not build one yourself? The Washington Post launched Climate Answers; the Financial Times launched Ask FT — both powered by each outlet's own deep archive of specialist content. The strategy is clear: when every chatbot can give a "good enough" answer, the on-site experience itself becomes the real competitive advantage.

🔹 The Economist — Use Video to Create What AI Can't Copy

In October last year, The Economist launched a new video product — Insider — producing long-form programmes of 30–60 minutes, with senior editors on camera interacting with audiences in real time.

• Video content is significantly harder for AI to replicate than text
• The atmosphere of engaging with a topic alongside an expert is something AI simply cannot imitate
• Each senior editor becomes a brand in their own right — readers aren't just after information, they want the expert's thinking and point of view

🔹 Condé Nast — From Screens to Real Life

Home to Vogue and Vanity Fair, Condé Nast has turned its attention to the one arena AI has the hardest time touching: live events.

• Vogue World is estimated to have generated over USD 30 million in revenue in 2025, up 50%+
• The CEO put it plainly: "Editors need to move beyond the fixed mindset of the magazine page and think across video, digital, and live experiences together"

🔹 Axios — "Zero-Party Data" and the Art of Radical Personalisation

Axios has gone premium — its Axios Pro newsletter carries an annual fee of USD 599, offering exclusive reporting and depth that AI simply can't replicate.

Zero-party data at sign-up: During the onboarding flow, Axios directly asks readers which topics they care about. From day one, subscribers receive highly targeted, localised newsletters based on preferences they've explicitly shared.

Personalised send times: Axios personalises the delivery time of each newsletter to the moment each individual reader is most likely to open it. This single strategy lifted open rates by 8%.

• Axios deliberately optimised its unsubscribe flow to make it easy for inactive readers to leave — counterintuitively, this increased loyalty among those who stayed

If the stories above show us individual publishers fighting back, zooming out reveals a structural ecological shift happening across the entire news industry. Traditional media is being squeezed from two directions simultaneously — AI technology and independent creators. Search engines have gone from "traffic controllers" to "answer engines," generating complete responses directly in the chat window and leaving publishers in a painful position: used, but invisible.

• Content licensing and revenue sharing looks like a way out, but the reality is tough — only around 20% of publishers expect to see meaningful income from it
• 76% of publishers now rank subscriptions and membership as their top priority — even above growing ad revenue
• The industry is shifting from a traffic economy to a relationship economy — no longer chasing every click, but cultivating loyal readers willing to pay over the long term

In an era when AI can generate a "good enough" news summary in seconds, journalism's traditional culture of keeping reporters in the background is breaking down. 76% of media executives say they will ask journalists to operate more like "creators" this year — building personal identity and recognisability.

• The Economist, which once kept reporter names hidden, has begun letting editors speak for themselves through podcasts and newsletters
• Wired has turned prominent journalists into influencers — actively producing short-form video and participating in live streams
• Mainstream outlets are placing correspondent headshots prominently in layouts, turning reporters into brands

Why does this matter? AI can synthesise information quickly — but it cannot stand in a room and feel the atmosphere, conduct the deep interview that changes the direction of a story, or use twenty years of experience to smell a major story in an unremarkable data set. First-hand observation and professional judgment are journalism's last moat in the AI age.

79% of media executives say investing in video journalism is critical to their business. This is no longer a trend prediction — it's industry consensus.

• In 2026, vertical short-form video has fully hit the mainstream, with even older demographics now accustomed to the "swipe up" viewing habit
• The BBC and The Economist have both integrated short-form video into their own websites — it's no longer just a social platform format
• Video conveys a journalist's expertise, on-the-ground atmosphere, and human warmth in ways text cannot — all elements that are hard for AI to replicate
• Tomorrow's journalists will need not just writing ability, but camera literacy, editing skills, and community management — the talent brief is being rewritten

The report introduces a forward-looking concept: "content liquefaction." News is no longer a static article in a fixed format — it flows like water, adapting automatically to whatever container the reader's context requires.

• 🧬 Atomise long-form reporting — break it down into its smallest units: core facts, key quotes, data context
• 🤖 These "atoms" can be automatically reassembled by AI and pushed to different platforms in different formats
• 📧 Email newsletters are getting particular attention — they reach readers directly in their inbox, completely free of algorithmic interference
• 🎧 Podcasts, social media, and live events are becoming critical channels for building sustained loyalty

The future competitive edge isn't just content quality — it's the ability to build stable, long-term audience relationships. 

AI-era news production is moving towards two poles — on one end, high-quality in-depth journalism with high production costs but strong brand loyalty; on the other, low-cost AI-automated content that quickly satisfies basic information needs.

• Publishers stuck in the middle are facing the greatest survival pressure — not distinctive enough in quality, not efficient enough to compete with AI
• Every publisher must make a clear choice: invest in professional depth, or embrace automated efficiency
• From traffic economy to trust economy — in a world of information abundance and AI proliferation, trust is the scarcest and most valuable currency

The previous two sections showed us the media industry's traffic crisis and its breakout strategies — but even the best transformation blueprint falls apart if the underlying infrastructure can't keep up.

Reach plc is a textbook example of this challenge. As the UK's largest commercial news publisher, with 150+ news brands — including Mirror, Express, Star, Liverpool Echo, and WalesOnline — across 55 core websites serving more than 40 million readers every month, the scale is there and the brands are strong. But technical debt was acting like an invisible chain, holding the whole organisation back.

Three specific constraints:

• Innovation held hostage by vendors. Every new feature required separate negotiation and extra fees. Just running through the trial, contract, and setup process could eat months before a single feature went live.
• Manual processes consuming engineering bandwidth. Every website change required engineers to handle it by hand. Development teams queued up waiting, and the pace of updates and new features was severely throttled.
• Fear of the domino effect. With 55 websites sharing one architecture, any change could trigger a chain reaction. As one engineer put it: "We were worried that one change would bring all our sites down at once."

For a news group serving millions of readers every day, this wasn't just a technical problem — it was a survival problem. While competitors were already using AI chatbots and personalised recommendations to win audiences, Reach couldn't change a single line of configuration without joining a queue.

Reach decided to stop patching and started fresh — partnering with AWS to migrate all 55 core websites to the cloud, targeting three core problems:

🔹 Deliver a Smooth Experience to 40 Million Readers

Post-migration, whether a reader is scrolling on the Tube in London or reading the morning paper at a café in Belfast, they get a low-latency experience. When breaking news drives a traffic spike, the system scales automatically — no more worrying about the site going down under load.

🔹 Free the Development Team from Manual Work

Where every change previously required a manual handoff to the ops team, developers can now review, test, and ship on their own — fully automated deployment means no human bottleneck between idea and production. The team can also run A/B tests quickly, validating in real time which page designs and headline approaches keep readers on site.

🔹 Actually Understand Reader Behaviour With Data

Previously, Reach had limited visibility into how readers were actually using their sites. Post-migration, the team can analyse browsing behaviour with ease, optimise search engine rankings, diagnose site issues quickly, and build the data foundation needed for personalised experiences in the future.

Looking back across all three threads in this piece, Reach's story answers the most practical question of all: when a publisher decides to transform, is the foundation actually ready?

Bloomberg's AI on-site search needs stable, fast content delivery. The Financial Times' chatbot needs elastically scalable compute. The Economist's video pivot needs globally distributed content delivery. Behind every one of these innovations is a reliable, agile, scalable cloud infrastructure, making it possible.