Digital Advertising Has Never Been Bigger. So Why Does It Feel Broken?

Updated: Oct 11

Follow the money and digital advertising looks unstoppable. US internet advertising revenue reached $294.6 billion in 2025, up 13.9 per cent, despite a calendar without a Summer Olympics, men’s FIFA World Cup or US presidential election. Europe’s market added €12.5 billion to reach €131.1 billion, an increase of 10.5 per cent. Globally, Google generated $294.7 billion in advertising revenue, while Meta reported $196.2 billion, equivalent to 97.6 per cent of its total revenue.
Now talk to the people signing the cheques and the picture becomes less reassuring. Attention is harder and often more expensive to buy, while few organisations can say with confidence which campaigns genuinely changed customer behaviour. The dashboards are full; the conviction is not.
Both pictures are accurate. The market continues to grow while the model underneath it becomes more concentrated, less transparent and increasingly difficult to measure, leaving marketing budgets caught between impressive headline numbers and uncertain economic value.
Two Markets, One Headline
The headline growth conceals a widening divide. Money is flowing towards large platforms, video and commerce environments, while the open web competes for a diminishing share of advertising expenditure.
Google’s 2025 accounts show this shift clearly. Search revenue increased by $26.4 billion and YouTube advertising added another $4.2 billion, while Google Network revenue, which includes advertising placed across partner websites through AdSense, AdMob and Google Ad Manager, declined by $567 million. The fall was primarily attributed to weaker AdSense revenue, partly offset by growth in AdMob.
Advertisers have not stopped spending on digital media. They have moved towards environments where one company controls the audience, the data, the algorithm and, increasingly, the transaction itself.
That consolidation buys convenience at the cost of leverage. A brand can reach millions of people within hours without knowing precisely how they were selected, why prices increased or whether the reported sale would have happened without the campaign. The results appear measurable because the reporting is detailed, yet the logic producing those results remains largely behind the platform’s walls.
Programmatic buying illustrates the problem. The ANA’s latest benchmark found that the share classified as working media in web and mobile programmatic buying increased from 36.0 per cent in 2023 to 47.1 per cent in the third quarter of 2025. That is meaningful progress, but it still leaves more than half of expenditure outside the study’s definition of working media.
Precision Was the Pitch
Digital advertising was built on a promise that television and print could never fully offer. It would find the right person, deliver the right message and prove the connection between advertising exposure and a completed sale.
That promise depended on an extensive tracking infrastructure that most consumers never saw. Cookies, device identifiers, pixels, data brokers and application trackers connected behaviour across websites and devices. Privacy legislation, consent requirements, Apple’s tracking controls and growing public concern have steadily reduced the availability and reliability of those signals.
Chrome remains the notable exception among major browsers. Google abandoned its plan to remove third-party cookies from Chrome and later retired many of the core Privacy Sandbox technologies intended to support privacy-preserving targeting and measurement. However, this decision did not restore the universal tracking environment on which digital advertising once depended.
Safari and Firefox continue to restrict cross-site tracking, while Apple’s application controls and stricter consent regimes have fragmented the information available to advertisers. As independent signals become harder to obtain, brands become more dependent on the first-party data and automated buying systems controlled by a small number of platforms.
The commercial exchange is increasingly clear. Advertisers receive speed, automation and scale, but surrender visibility into how decisions are made. Tools such as Google’s Performance Max can determine audiences, bids, placements and creative combinations, while the advertiser receives an outcome without necessarily receiving a full explanation.
AI Is Rewriting the Search Bargain
For two decades, search operated through a relatively simple exchange. A person entered a question, reviewed a page of results and clicked through to another website, creating a natural place for paid advertising within the journey.
Generative answers disrupt that exchange because they reduce the need to leave the search environment. When an AI system summarises, compares and recommends, the answer itself can become the destination rather than the gateway.
Pew Research analysed the browsing activity of 900 US adults and found a material difference in behaviour. When a Google AI summary appeared, users clicked a conventional search result on 8 per cent of visits, compared with 15 per cent when no summary appeared. Links contained within the AI summary were clicked on only 1 per cent of visits, and users were more likely to end their browsing session after receiving the generated answer.
Search advertising is unlikely to disappear because Google and its rivals have powerful incentives to place commercial content within or alongside generated answers. The commercial prize, however, is changing. Brands will compete less for a position on a results page and more for inclusion within an answer, comparison, shortlist or recommendation.
Paid placement will remain important, but it may not be sufficient. AI-generated answers may draw on product information, customer reviews, expert commentary, creator content and the consistency with which a company is described across independent sources. Visibility will therefore depend not only on media expenditure, but also on whether the wider information environment gives the system a credible reason to mention the brand.
Search optimisation is consequently moving closer to reputation management. SEO, public relations, product data, customer experience and content strategy can no longer operate as isolated disciplines because they increasingly shape the same recommendation environment.
Confident, Not Correct
Digital advertising sold itself on measurability. What it perfected was counting, and the distinction between the two is becoming impossible to ignore.
Consider a typical purchase journey. A customer may watch a video, read an independent review, open an email, search for the brand and finally click a retargeting advertisement before completing the purchase. The retargeting platform may claim the conversion even though it appeared at the end of a decision formed elsewhere.
Place several platforms across the same journey and each can produce a persuasive report showing that it influenced the sale. The numbers may be technically accurate within each platform’s attribution rules, but they do not establish what caused the customer to buy.
The confidence gap is visible in the research. Nielsen found that 85 per cent of marketers were confident in their ability to measure return on investment, while only 32 per cent measured ROI holistically across traditional and digital media channels. A detailed dashboard can describe what happened within a platform, but it is not, by itself, evidence of causation.
As individual-level tracking becomes less reliable, the more credible tools are often the slower ones. Controlled experiments, geographic holdout tests, customer lifetime value, contribution margin and marketing mix modelling provide fewer instant answers, but the answers they produce are more likely to survive scrutiny from finance and senior management.
Cheap Content, Expensive Attention
Generative AI has sharply reduced the cost of advertising production. Hundreds of headlines, images, audience variations and video edits can now be created in the time previously required to develop a small number of campaign assets.
The efficiency gain is real, but it is available to almost every competitor. As feeds fill with polished but interchangeable material, production becomes cheaper at the same time that genuine attention becomes more difficult to earn.
This changes where value sits in the creative process. The advantage moves away from producing more content and towards creating work with distinctive positioning, emotional relevance, recognisable brand assets and a clear point of view. AI will not end advertising, but it will reduce the value of routine production and increase the premium attached to original ideas.
The implication is not that every advertisement must become a creative masterpiece. It is that efficiency without differentiation simply increases the volume of material competing to be ignored.
Closer to the Till, Closer to Trust
The fastest-growing parts of the advertising market point in two related directions. Media is moving closer to the transaction, while influence is moving closer to people and sources that audiences already trust.
Retail media represents the first direction. Estimates vary according to definitions and forecast dates, with WPP Media placing global retail media at $169.6 billion in its mid-year 2025 forecast and Forrester’s later estimate reaching $184 billion. The underlying logic is straightforward: retailers and marketplaces can connect audience data, advertising exposure and purchase behaviour within the same commercial environment.
This proximity to the transaction makes retail media attractive, but it does not remove the need for scrutiny. Retailers frequently operate their own audience data, media inventory, measurement system and attribution model, recreating many of the same transparency questions associated with large technology platforms.
Creators represent the second direction. US creator advertising expenditure was projected to reach $37 billion in 2025, an increase of 26 per cent and approximately four times the growth rate of the wider media industry. Creators do more than distribute a message because they explain, demonstrate and give products social context before audiences that may ignore conventional advertising.
Retail media pulls advertising towards the point of sale, while creators pull it towards established trust. Connected television, social commerce, brand communities and AI-generated recommendations combine these forces in different proportions, making the boundaries between media, content, commerce and customer experience increasingly difficult to locate.
The Old Model Is What Is Dying
Digital advertising is not disappearing. It is becoming less independent, less transparent and more closely integrated with a small number of commercial and technical ecosystems.
The old model rested on four assumptions. Companies could track people across the web, buy attention relatively cheaply, direct traffic towards their own websites and attribute each sale to a click. All four assumptions are now under strain.
Tracking is fragmented, attention is scattered and AI-generated answers increasingly keep users within the platform. At the same time, click-based attribution is becoming harder to defend as a reliable account of what caused a commercial outcome.
Reducing advertising expenditure indiscriminately would therefore be the wrong response. The more useful question is what paid media is expected to achieve and how it connects with brand, customer experience, CRM, content, pricing and sales. Advertising can capture existing demand and support future demand, but it cannot indefinitely compensate for an undifferentiated proposition or a lack of customer trust.
The companies best positioned for the next phase will own meaningful customer relationships, build brands worth recommending and appear naturally within the environments where decisions are made. Their advertising may resemble an expert explanation, a creator video, a marketplace listing, a useful tool or a place within an AI-generated comparison rather than a conventional campaign.
Growth and health are not the same thing. Digital advertising is moving from an economy dominated by impressions and clicks towards one shaped by evidence, relevance and trust. Companies that understand this transition will continue to extract value from digital media, while those that respond by simply producing more advertisements will pay more each year to be ignored.

Comments