Download PDF
Cite this reportIntroduction
This report is for people spending their own money on ads: online store owners, small brand operators, and the analysts who cover them. Digital advertising is growing, which surprises nobody. The useful question is where the growth went, and the answer has moved a long way from the search-plus-social model most media plans still assume. Retail media is now the fastest-growing major channel in the world and is on track to overtake linear and connected TV combined in 2026, while search, the channel most sellers still budget around first, is forecast to grow at 3.4%. Ten companies took 84.1% of every US internet advertising dollar last year. And Meta's average price per ad rose 12% year over year in each of the first two quarters of 2026, per Meta's own filings.
The report is reviewed twice a year, and the next pass is January 2027, after the December forecast round.
How big is digital ad spend in 2026, and how fast is it still growing?
Global advertising spend will reach $1.06 trillion in 2026, growing 5.0%, with digital accounting for 69% of the total. That's dentsu's May 2026 forecast across 56 markets, and it also records the year the industry actually passed the trillion-dollar mark: 2025, when spend grew 5.8%.
But no two forecasters agree, and the gap between them is not small. WPP Media put 2026 global advertising revenue at $1.3 trillion in June 2026, growing 8.9% excluding US political spend, revised up from the 7.1% it forecast six months earlier. WARC published $1.30 trillion for 2026 in December 2025, up 9.1%, with 2027 at $1.40 trillion. Statista's free-tier figure is $1.26 trillion for total advertising worldwide, up 6.6%. EMARKETER puts worldwide total media ad spending at $1.170 trillion for 2026 and says digital is growing 13.5% against traditional media's 2.5%.
The lowest and highest 2026 totals on this page are dentsu's $1.06 trillion and WARC's $1.30 trillion, and the difference is definitional rather than a disagreement about reality. dentsu forecasts advertiser spend across 56 markets. WPP Media forecasts media-owner revenue and strips out US political. WARC constitutes its global market differently again. Averaging them would produce a number no publisher stands behind. This report anchors on dentsu because it's the most recent full-vintage forecast on a free, citable page, and because it states digital's share as a published figure rather than leaving it to be inferred.
For anything that needs to be an actual rather than a forecast, there's exactly one strong option, and it's US-only. The IAB and PwC put US internet advertising revenue at $294.6 billion for full-year 2025, up 13.9%, the highest level ever recorded. It's a survey of publisher revenue with a documented methodology, published in April 2026. Anyone building a business case on a global digital number is building it on a forecast; anyone building on the IAB figure is building on measurement, but only for one market.
Which digital advertising channels are still growing in 2026?
Retail media is the fastest-growing major channel in the world, forecast to grow 12.3% in 2026, while search grows 3.4%. That's a wider gap than most media plans reflect. WARC forecast global retail media investment at $174.9 billion for 2025, rising 12.4% to a forecast $196.7 billion in 2026, which would be 16% of all advertising, not just digital advertising. WARC also expects retail media to overtake combined linear and connected TV spend during 2026. WPP Media reached the same conclusion from a different direction: it put global commerce media at $178.2 billion in 2025, the first year it surpassed total global TV ad revenue.
The clearest single proof of that shift is a company filing. Amazon's advertising services revenue was $67.635 billion in full-year 2025, up 22%, and $17.243 billion in the first quarter of 2026, up 24%. On the Q1 call Andy Jassy said "advertising grew to over $70 billion in TTM revenue." Amazon reports Q2 2026 on July 30, 2026, after this report's verification date, so that figure will move. No forecast, no methodology to argue with. Just a line in an earnings release.
In the US, the channel order changed. Social media advertising reached $117.7 billion in 2025 and overtook search at $114.2 billion for the first time, per the IAB and PwC. Digital video hit $78.0 billion, up 25.4%, and the IAB expects US digital video to pass $80 billion in 2026 and to account for more than 60% of all TV and video ad spend. Worth a caveat: the IAB's format categories overlap, so social video counts in both social and video, and the channel figures do not sum to the $294.6 billion total. They're levels, not shares.
Connected TV is where the primary disclosures got interesting this year. Netflix told shareholders in July 2026 it's "on track to deliver approximately $3 billion in ads revenue in 2026," which is a company target rather than a result. At its May 2026 upfront it said the ad tier reaches more than 250 million monthly active viewers globally. Amazon says Prime Video's ad-supported monthly reach is above 315 million. Roku reported $613 million in advertising revenue in Q1 2026 alone, up 27%. On the creator side, the IAB projected in November 2025 that US creator economy ad spend would reach $37 billion for the year, up 26% and growing roughly four times faster than total media. Anyone allocating budget on the old assumption that search plus social is the whole digital market is now missing about a sixth of global advertising.
Who actually takes the money in digital advertising?
Ten companies took 84.1% of all US internet advertising revenue in 2025. That's up from 80.8% in 2024, 79.8% in 2023 and 76.8% in 2022, from the IAB and PwC's April 2026 report. The line isn't straight: concentration actually fell in 2022, from 78.6% in 2021, before climbing every year since. The next fifteen companies, ranks 11 through 25, share 8.3% between them. Every other advertising business in the United States divides 7.5%.
The global picture is the same shape. WARC calculates that Alphabet, Amazon and Meta together took 56.1% of all advertising spend outside China in 2025, worth $556.6 billion, and forecasts 58.0% for 2026. WPP Media, working independently, forecasts the three largest non-China sellers at 57.6% for 2026. Two methodologies, two publishers, roughly the same answer for the same year.
The filings show why. Alphabet reported $81.629 billion in Google advertising revenue in the second quarter of 2026 alone: Search and other at $63.271 billion, up 17%, YouTube ads at $11.055 billion, up 13%, and Network at $7.303 billion. Meta reported $59.363 billion in advertising revenue for the same quarter, up 27%. Meta's advertising revenue grew 27% in the quarter and Google Search and other grew 17%, against dentsu's 5.0% forecast for the total 2026 market. Those are different measures on different bases, so the gap is not a clean comparison, but it points the same way the concentration data does.
There's a second-order effect that gets less attention. Alphabet stopped publishing paid clicks and cost-per-click after the first quarter of 2018, and hasn't reinstated either metric. Amazon does not publish an ad pricing metric. Of the three companies that take well over half of global advertising outside China, exactly one, Meta, still tells the market what its ads cost. So the industry's visibility into ad price inflation now rests on a single company's voluntary disclosure and on third-party benchmarks with self-selected samples. Nobody in the industry seems especially bothered by that, which is odd.
Is AI changing where online shoppers actually come from?
AI-referred traffic to US retail sites grew 138% year over year in May 2026, and those visitors now convert 54% better than everyone else. Adobe's figure, built on more than a trillion visits to US retail sites. The reversal is the part worth sitting with: in March 2025, traffic arriving from AI assistants converted 38% worse than other traffic. By March 2026 it converted 42% better. By May, 54%. AI visitors also spend 53% longer on site, view 23% more pages, and bounce 36% less.
Growth is decelerating from a very low prior-year base, which is normal and not a warning sign. AI-referred retail traffic was up 693% over the 2025 holiday period, 393% in the first quarter of 2026, and 138% by May. Adobe's survey of more than 5,000 US consumers puts adoption at 39%, with 85% of those users saying AI improved the shopping experience.
Meanwhile, the "AI is killing search" thesis is not showing up in the search numbers. Alphabet's AI Mode passed one billion monthly active users, the Gemini app reached 950 million, and Google Search and other revenue still grew 17% in the second quarter of 2026. Sundar Pichai's framing was that "our popular AI features are driving Search query growth." Advertising inside AI is real but early: WPP Media forecasts generative search ad revenue at $5.1 billion for 2026, rising past $100 billion by 2030. OpenAI began serving ads to logged-in adults on ChatGPT's Free and Go tiers in February 2026 and opened a self-serve Ads Manager beta in May, with dentsu, Omnicom, Publicis and WPP as launch agency partners, but has disclosed no advertiser count and no reach figures at all.
EMARKETER's forecast is the useful corrective to the hype. It expects US AI advertising spend to more than double to $68.25 billion by 2030, but says over 80% of 2026 AI ad spending sits adjacent to AI-generated content, next to AI Overviews and AI Mode, with under 20% inside standalone chatbots. One more caution about vendor claims in this space: when Google launched AI Max for Search in May 2025 it said advertisers typically saw 14% more conversions at a similar CPA. In April 2026 that became an average of 7%. The stated uplift halved in eleven months, and Dynamic Search Ads start auto-upgrading to AI Max in September 2026 regardless.
- Nov–Dec 2025 (holiday)+693%
- Q1 2026+393%
- May 2026+138%
What does eCommerce advertising actually cost, and what ROAS is realistic?
Meta's average price per ad rose 12% year over year in both the first and second quarters of 2026, up from 9% across full-year 2025. That's from Meta's own filings, and it's the cleanest available evidence that ads are getting more expensive for the people buying them. Ad impressions grew 14% in Q2 2026 while price grew 12%, so the platform is selling more inventory and charging more for it.
Benchmark data from advertisers themselves lines up. Triple Whale, measuring across roughly 35,000 brands for full-year 2025, puts the median eCommerce CPM on Meta at $14.19, up 20.03% year over year, with a 2.19% click-through rate, a $38.19 cost per acquisition, and a median return on ad spend of 1.86, itself up 1.29%. TikTok came in cheaper and, in 2025, more efficient: a $13.26 CPM, up 16%, and a median ROAS of 2.21, though TikTok's ROAS fell 5.7% over the year. Triple Whale does not state a geography for either sample, which is a real limitation on both figures. On the search side, where most eCommerce PPC budgets end up, WordStream's study of 13,474 US Google Ads campaigns between April 2025 and March 2026 found a median all-industry CPC of $5.42, with Apparel, Fashion and Jewelry at $4.44 and a 4.50% conversion rate, and Shopping, Collectibles and Gifts at $4.14 and 4.01%.
Which brings us to the number most sellers are measuring themselves against. The "4:1 is an acceptable ROAS" rule that circulates everywhere traces to a Shopify blog post from October 2023 that cites no study, no sample and no source. The measured median across 35,000 brands is 1.86. That gap has consequences. A store judging a 2.5x campaign against an invented 4:1 benchmark will read it as a failure, when it sits above the measured median for its channel. Common Thread Collective's portfolio of 200 to 300 DTC stores showed the same pressure in Q1 2026, with Meta spend up 25.28% while Meta ROAS fell 3%, and Google spend up just 3.65% while Google ROAS rose 12%.
There's a real gap in this data and it's worth naming rather than filling. No named, dated, sampled benchmark for blended customer acquisition cost in eCommerce exists in the public record. The widely-quoted "CAC rose 222%" claim comes from two 2022 pages on a vendor's own site that attach the same percentage to two incompatible metrics, with no methodology on either. Platform-reported CPA is the only defensible proxy, and it measures a narrower thing. On budget intensity the best available figure is Gartner's, from 401 CMOs across North America, the UK and Europe surveyed in early 2026: marketing budgets at 7.8% of company revenue with paid media taking 31.4% of that. Read it cautiously, though. The vast majority of Gartner's respondents work at companies above $1 billion in revenue, and nobody publishes an equivalent with a small-seller sample.
Two structural facts round out the cost picture. The ANA's programmatic transparency work, using log-level data from 86 marketers, found that only 43.3% of programmatic investment in Q1 2026 delivered impressions that were fraud-free, measurable, viewable and not made-for-advertising. Top-performing buyers got an effective $7.46 TrueCPM; the worst performers got $19.04. And the signal-loss story most sellers were told to prepare for did not happen: Google confirmed in April 2025 that third-party cookies stay in Chrome, then retired most of Privacy Sandbox in October 2025, including Topics and Protected Audience. Regulation is still real, with EUR 7.1 billion in cumulative GDPR fines and EUR 1.2 billion in the year to January 2026, but the specific technical migration the industry spent five years planning was canceled.
Conclusion
Three things the 2026 data says without ambiguity.
The advertising market passed $1 trillion in 2025, and dentsu forecasts digital at 69% of the 2026 total. But no two forecasters agree on that total: the range on this page runs from dentsu's $1.06 trillion to WARC's and WPP Media's $1.30 trillion, on different definitions. Anyone quoting a single global digital figure without naming the publisher and the vintage is quoting a guess.
The money moved. dentsu forecasts retail media growing 12.3% in 2026 against search's 3.4%, and WARC forecasts retail media passing combined linear and connected TV spend during the year. Amazon's advertising business alone cleared $70 billion on a trailing-twelve-month basis, per Amazon.
Ad prices are climbing and the benchmark most sellers judge themselves against has no study behind it. Meta's own filings show average price per ad up 12% year over year in each of the first two quarters of 2026. The measured median eCommerce ROAS on Meta is 1.86 across roughly 35,000 brands, not the widely repeated 4:1, which traces to a 2023 blog post with no sample.