The ad market is two markets, and most forecasts only see one
Ian Whittaker argues the ad market splits into a visible, agency-led layer and a hidden, fast-growing SMB layer that most industry forecasts consistently miss.
Display
A media and capital markets analyst today laid out a structural argument that cuts against the way most industry observers read advertising market data: the global ad market is not one market, but two - and the one that appears in agency earnings reports, trade forecasts, and industry conferences is the smaller, slower, and increasingly misleading one.
Ian Whittaker, who advises chief executives, chief financial officers, and investors on advertising economics and capital allocation in media, marketing, and technology, published the analysis on LinkedIn on May 12, 2026. It arrives shortly after a set of first-quarter results from the major digital platforms that produced some of the largest reported growth figures in recent memory, and directly after those same quarters revealed continued revenue declines at major holding companies. The juxtaposition, Whittaker argues, is not a paradox. It is evidence of a structural split that the industry has largely failed to name correctly.
Two markets, not one
The first market is the one that agency reports measure, that trade publications cover, and that financial analysts model when they produce forecasts. According to Whittaker's post, it is "the one everyone in the industry knows. Large enterprise advertisers, with marketing departments, agency relationships, and media plans built around a small number of holdcos." He describes it as "brand-led, mature, slow-growing, increasingly fought over." WPP, Publicis, and their peers operate inside this layer. So do Campaign, AdAge, and the holding-company earnings calls that define how most outside observers understand the health of advertising.
The second market is structurally different. It is composed of small and medium businesses that access advertising platforms directly, without an agency in the chain, primarily through self-serve interfaces. According to Whittaker, this layer is "performance-led, fragmented, growing rapidly." It does not appear in agency revenue figures because those businesses were never clients. It does not appear in most industry surveys because there is no agency to survey. And the trade press has limited visibility on it because it operates entirely outside the institutional structures that produce advertising industry news.
The scale of this second market is substantial. According to Whittaker, Meta has discussed having ten million advertisers on its platform. Google has approximately 1.2 million. Amazon's third-party seller advertising base is described as "similar in shape." Many of those advertisers, he notes, are Chinese-based, which further complicates the picture for analysts relying on Western industry data sources.
The earnings gap explained
The discrepancy between platform revenue growth and holding-company revenue trends has generated considerable confusion in recent quarters. Whittaker's framing offers a specific mechanism for why that confusion exists.
According to the post, WPP's revenue decline is frequently interpreted as evidence of an industry under pressure. "That is only half the story," he writes. "Yes, its clients - the large enterprise budgets - may be under pressure. But the other market is growing, and last month's Big Tech results showed it: Meta advertising up 33 percent, Amazon advertising up 24 percent, Google Search up 19 percent."
Those three figures represent the Q1 2025 results from Alphabet, Meta, and Amazon. PPC Land covered those numbers in detail, noting that media analyst Whittaker and podcast host Justin Lebbon had described a similar thesis on the Unfiltered Media Podcast on April 30, 2025. At that time, Whittaker estimated that small and medium business advertising accounts for roughly half of all global ad spend, with the vast majority flowing directly to platforms without an agency or holding company in the chain.
The platforms are therefore growing a market that the holding companies do not sell into. That, Whittaker argues, explains the divergence - not macro weakness, not advertiser budget constraints as a universal condition, and not the AI disruption narrative that has dominated holding-company commentary.
Three structural implications
Whittaker's post draws three implications from this analysis, each carrying different weight for different parts of the industry.
The first concerns the nature of the problem facing holding companies. According to the post, "this is a structural rather than cyclical problem for HoldCos. Their cost base was built to service enterprise clients in the visible market but the SME market does not need them. No restructuring of the agency model fixes that exposure." This framing matters because it shifts the diagnosis from a problem that can be solved through cost efficiency or technology investment, to a problem rooted in which customers an organisation was designed to serve. If the growth is coming from ten million small businesses using self-serve interfaces, there is no version of a holding company that captures that revenue through its existing commercial model.
The second implication concerns competitive moats. Whittaker argues that the platforms' real advantage is not their large-advertiser relationships - those can in principle be disrupted by a new entrant or by regulatory action. The durable advantage is in the SMB layer below the waterline. "That is where the platforms are genuinely irreplaceable (for now)," he writes, "and it is the part of their revenue that gets least attention." The logic is that there is no alternative infrastructure for an agency-mediated SMB market. No competing ecosystem exists through which ten million small businesses could run performance advertising campaigns without going through Meta, Google, or Amazon.
This connects directly to a pattern PPC Land has documented across platform development: Amazon, Google, and Meta commanded approximately 72 percent of US digital advertising spend in 2025, according to EMARKETER data published on July 29, 2025. That figure reflects a concentration built substantially through SMB volume, not just Fortune 500 budgets.
The third implication - arguably the most forward-looking - concerns AI agents. This is where Whittaker's analysis converges with one of the most active debates in advertising technology right now. In the visible market, AI agents may disintermediate agencies further, automating the tasks that holding companies currently perform for large enterprise clients. But in the invisible market, the dynamic runs in the opposite direction.
According to the post, "AI agents probably accelerate platform growth by lowering the friction for SMEs to advertise effectively. The same technology has opposite effects depending on which side of the waterline you are looking at." This is not a marginal observation. It suggests that the same capabilities being debated in terms of whether they threaten agency jobs in large brand campaigns are simultaneously reducing the technical barriers that previously prevented small businesses from running campaigns at all. The two effects compound the bifurcation rather than resolving it.
PPC Land has tracked the agentic AI buildout across platforms extensively. In November 2025, Amazon unified its DSP and sponsored ads console, launched AI agents for campaign management, and extended its Creative Agent to television formats. Google made its Ads Advisor and Analytics Advisor available to all English-language accounts. The IAB Tech Lab introduced its Agentic RTB Framework. Each of these moves lowers the operational complexity of running advertising campaigns - which, in Whittaker's framework, primarily benefits the SMB layer by making it easier for businesses without marketing departments to access performance advertising.
The measurement problem
One dimension that Whittaker's post does not address at length, but which is present in the comment thread beneath it, concerns how the SMB layer is measured and whether platform reporting gives advertisers an accurate view of what that spending is actually producing. Alex Tait, identified as a marketing mix modelling and marketing professional in the comments, notes that "a lot of non-top-300 advertisers are likely leaving substantial incremental profit on the table by relying too heavily on platform measurement." Platform reporting is described as useful for day-to-day optimisation but "not designed to give a clean view of incrementality or true" return on investment.
This is a material issue for anyone interpreting the second market. If the SMB layer is growing rapidly partly because small businesses are accepting platform-reported metrics without independent validation, that growth carries a different quality than growth driven by measured incremental returns. The self-serve model removes the agency, and in most cases also removes the measurement infrastructure that sophisticated advertisers use to evaluate whether their spend is generating real business outcomes. Platforms like Criteo have attempted to address parts of this gap - opening their GO platform to full self-service for SMBs on March 31, 2026 - but the measurement question remains unresolved for the vast majority of the ten million small businesses running campaigns without specialist support.
Implications for the visible market
The post also draws a response from James Connelly, identified as a board advisor and CEO, who argues that the enterprise clients the holding companies serve "are all being squeezed by new entrants in their market. New, innovative challengers. The growth is coming from these SMEs. They would never work with a holdco. They do work with independent agencies though. They are faster, less cost layers, transparent, and adaptable."
Andrew Jude Rajanathan, identified as working in brand marketing, adds two points from the visible-market side. First, that large enterprise advertisers still need agencies to deliver global campaigns at scale - the parts that cannot be brought in-house. Second, that what has structurally weakened is not large-advertiser demand per se, but the broadcast model that holding companies were built around. "What's actually broken is the TV/broadcast model the HoldCos built their reputations on. That model used to compound. It just doesn't anymore."
This aligns with a pattern visible in platform revenue data. The UK advertising market grew 6.4 percent to £46.7 billion in 2025, according to the Advertising Association and WARC Expenditure Report published on May 5, 2026. But two of every three pounds went to Google, Meta, and Amazon - leaving approximately £15 billion for the entirety of television, audio, publishers, out-of-home, and other channels. The headline growth figure obscures a structural redistribution that the visible-market framing alone cannot explain.
What the comment thread reveals
The post attracted responses from a range of practitioners, which together sketch a more detailed version of the argument. Edward Papazian, President at Media Dynamics Inc., argues for a four-way split rather than two: long-term branding in large and small advertiser segments, and short-term sales promotion performance in each of those segments. "In reality, you've got four not two splits - all marching to different drummers," he writes.
Julien Boyreau of Marty TV frames the right segmentation differently, suggesting it should run between brand and product sellers rather than between large enterprise and SMBs. "You have plenty of SMB-size DNVBs and app games shops spending millions on Meta on performance metrics," he notes. This is a technically significant observation: the invisible market is not defined purely by company size, but by direct-to-platform behaviour and performance orientation. A venture-backed direct-to-consumer brand with ten employees can spend as much on Meta as a mid-sized regional retailer - and both are invisible to holding-company data.
James Connelly's observation about independent agencies is also relevant here. The dichotomy Whittaker draws - holdcos on one side, platforms on the other - elides a third category of intermediary that is growing: independent, specialist agencies that do work with SMBs and growth-stage brands, and that operate more efficiently than the holding-company model. These firms are part of neither the visible market in the traditional sense nor the invisible self-serve layer, but occupy a position in between that is difficult to measure precisely.
Why this framing matters now
The timing of this analysis matters for practitioners. Global advertising revenue reached $1.14 trillion in 2025, according to WPP Media projections published in December 2025, with commerce advertising surpassing television for the first time. That milestone was driven substantially by the same platforms whose SMB base Whittaker is describing. If the invisible market is growing and the visible market is maturing or contracting, then industry-level figures that blend both will systematically overstate the health of the institutions that most observers use to evaluate the market - and understate the dependence of aggregate growth on a layer of advertisers whose behaviour, measurement practices, and durability are poorly understood.
For the marketing community, the practical implication is that WPP's revenue trend tells one story and Meta's revenue trend tells a different story, and neither alone describes the full advertising economy. Strategists advising on media allocation, technology investment, or competitive positioning need to be explicit about which market they are analysing - and which one their data sources can actually see.
Timeline
- April 30, 2025 - Ian Whittaker and Justin Lebbon discuss the invisible SMB market thesis on the Unfiltered Media Podcast, following Q1 2025 earnings from Meta (+33%), Google Search (+19%), and Amazon (+24%). PPC Land coverage
- July 29, 2025 - EMARKETER publishes data showing Amazon, Google, and Meta control approximately 72 percent of US digital advertising spend in 2025. PPC Land coverage
- November 10-13, 2025 - Amazon, Google, and the IAB Tech Lab launch agentic AI campaign tools and the Agentic RTB Framework, lowering operational barriers for SMB advertisers. PPC Land coverage
- December 10, 2025 - WPP Media projects global advertising revenue reaching $1.14 trillion in 2025, with commerce overtaking television for the first time. PPC Land coverage
- March 31, 2026 - Criteo opens its GO platform to full self-service for SMBs, enabling AI-powered cross-channel performance advertising without managed service agreements. PPC Land coverage
- May 5, 2026 - The Advertising Association and WARC publish UK ad market data showing £46.7 billion in 2025 spend, with two thirds captured by Google, Meta, and Amazon. PPC Land coverage
- May 12, 2026 - Ian Whittaker publishes his LinkedIn post outlining the structural bifurcation of the advertising market, drawing reactions from practitioners across agency, brand, and analyst communities.
Summary
Who: Ian Whittaker, media and capital markets analyst who advises CEOs, CFOs, and investors on the economics and capital allocation shaping media, marketing, and technology. Contributors in the comment thread include Edward Papazian (Media Dynamics Inc.), James Connelly (Charlie Oscar), Andrew Jude Rajanathan (brand marketing), Alex Tait (Entropy), and Julien Boyreau (Marty TV), among others.
What: A LinkedIn post arguing that the global advertising market is structurally bifurcated into two distinct segments: a visible, brand-led, agency-mediated market serving large enterprise advertisers and growing slowly; and an invisible, performance-led, self-serve market of small and medium businesses that flows directly to platforms and is growing rapidly. The post identifies three structural implications: that holding-company decline is structural rather than cyclical; that the SMB layer represents the platforms' most durable competitive moat; and that AI agents will have opposite effects on the two segments.
When: Published on May 12, 2026, following Q1 2025 earnings reports from Meta, Alphabet, and Amazon showing advertising growth of 33 percent, 19 percent, and 24 percent respectively, alongside ongoing revenue pressure at major holding companies including WPP.
Where: Published on LinkedIn by Ian Whittaker. The underlying data referenced covers global and UK advertising markets, with specific figures drawn from Meta, Alphabet, and Amazon quarterly earnings disclosures.
Why: The post addresses a persistent confusion in how the advertising industry reads its own market data. Agency earnings, trade forecasts, and most analyst models measure only the visible segment of the market - the large-advertiser, holdco-mediated layer. The SMB segment, which Whittaker estimates at roughly half of all global ad spend, is structurally invisible to those data sources. As platform growth accelerates driven by millions of direct advertisers and as holding-company revenues decline, that measurement gap produces contradictory signals that practitioners, investors, and strategists need to understand in order to interpret market health accurately.