Ad Tech Tax Explained 2026: Every Fee, With ANA Numbers

The ad tech tax is the share of a programmatic budget kept by intermediaries before it buys an impression. The ANA's benchmark puts transaction costs at 27.2% of spend in Q2 2026, and 45.1% of each dollar reaches a viewable, fraud-free impression (ANA via Fiducia, 6 August 2026). This page maps every fee in the chain with sourced ranges.

It also explains why the studies disagree, what "working media" means in each of them, and how to measure your own path with log-level data.

Three different things are called working media, and the number changes by ten to thirty points depending on which one a study reports.

Money that reaches the sell side. This is advertiser spend minus buy-side fees (agency, DSP, data, verification) and minus the SSP fee: what the publisher's ad server receives. The ANA calls it "seller revenue" and measured it at 76.2% of spend in Q3 2025 (ANA Benchmark, Q3 2025). The CMA's 65% and ISBA's 51% are the same concept measured on different data (CMA, Appendix R, 1 July 2020; ISBA, May 2020).

Money that buys a quality impression. The ANA's headline metric, TrueAdSpend, is "the ad spending going to TrueImpressions", where a TrueImpression is one that is "non-IVT, measurable for viewability, and viewable" (ANA Benchmark, Q1 2025); the Q1 2026 release describes the index as measuring "programmatic spend delivering fraud-free, measurable, viewable, and MFA-free impressions" (ANA, 27 May 2026). That is the 45.1% figure. It is seller revenue with the wasted impressions removed.

Media versus non-media in the marketing budget. Agencies and finance teams use "working media" to mean the part of the budget that buys media at all, as opposed to fees, production and research. That definition is older and broader and is not what the programmatic studies measure.

When someone quotes a working media percentage, ask which of the three it is. The Working Media Benchmark tracks the ANA series quarter by quarter under the second definition.

Read vertically, the table says that fees (hops 1 to 8) take between 23.8% of spend in the ANA's Q3 2025 waterfall (ANA Benchmark, Q3 2025) and 49% in ISBA's 2020 study (ISBA, May 2020), and that wasted impressions (hop 10) take most of what is left. Read horizontally, every hop has a range rather than a number, because every contract is different and because the counterparties do not publish rate cards. That second fact is why the measurement method at the end of this page matters more than any single figure.

The studies disagree because they measure different things on different data. The table lists each dated public datapoint with its scope.

Scope. eMarketer, the CMA and the 2017 ANA study counted fees. ISBA counted fees plus an unknown delta. The ANA benchmark counts fees and then wasted impressions. The CMA's 35% take (CMA, Appendix R, 1 July 2020) and the ANA's 54.9% of spend not reaching a quality impression in Q2 2026 (ANA via Fiducia, 6 August 2026) can describe the same chain.

Who is in the sample. ISBA's participants were "the highest profile advertisers, publishers" and the study said its figures were "likely a best case" because only disclosed trading models were included (ISBA, May 2020). The ANA benchmark's participants are ANA members who have fought their suppliers for log access, which is itself a selection effect: in Q3 2025, 39 marketers had asked and 21 had been granted it (ANA Benchmark, Q3 2025).

CTV. When the ANA added connected TV to its waterfall in Q1 2025, TrueAdSpend fell from 43.9% to 41.0% even though transaction costs dropped 3.7 points, because CTV impressions were less measurable and less viewable by the web standard (ANA Benchmark, Q1 2025). The web-and-mobile series (47.1% in Q3 2025) and the all-environment series (39.0%) are different numbers for the same quarter (ANA Benchmark, Q3 2025).

Averages versus medians. The ANA's waterfall uses sequential averages; its detailed findings use medians per metric. The two are not comparable, and the report says so: "Detailed findings should not be compared to the Cost Waterfall" (ANA Benchmark, Q1 2025).

Age. The CMA figures are 2019 data; ISBA I is Q1 2020; Adalytics is 2022. SSP fees in the ANA series fell from 14.7% to 12.8% between Q2 and Q3 2025 alone (ANA Benchmark, Q3 2025). Old figures are still the only public ones for some hops, which is why they are kept here with their dates.

Geography. ISBA, the CMA and Plum are UK; eMarketer, Adalytics and the ANA are US. Fee structures are similar; agency models and the Google share of the chain are not.

The direction is clear even if the levels are not comparable. Buy-side fees were 28% of spend in the ANA's 2017 data and 23% in the CMA's 2019 data (CMA, Appendix R, 1 July 2020). Transaction costs in the ANA series fell from 29% in 2023 to 23.8% in Q3 2025, then rose to 27.2% in Q2 2026 (ANA Benchmark, Q3 2025; ANA via Fiducia, 6 August 2026). Working media rose from 36% to 45.1% over the same period, mostly because made-for-advertising spend collapsed from 15% of the waterfall in 2023 to under 1% (ANA Benchmark, Q1 2025) and non-viewable spend fell (ANA via Fiducia, 6 August 2026). The ANA's own reading of Q2 2026 is that "pricing efficiency, however, moved the other way" (ANA, 26 August 2026): quality improved, CPMs rose.

The ExchangeWire column that popularised the "toll road" framing in September 2026 drew the same conclusion from the same data: the chain is cleaner than in 2020 but still takes more than half of each dollar before a quality impression is bought (ExchangeWire, 16 September 2026).

Industry averages tell you the shape of the problem. Only your own log data tells you the size. The method below is the one the ANA has recommended since 2023 and the one the ISBA toolkit formalised; it is also what the Microsoft Invest exit showed to be necessary for reasons beyond fees.

1. Inventory the chain. List every counterparty between your budget and a publisher: agency, ad server, DSP, data providers, verification vendors, SSPs, curators, resellers. The ANA's 2023 participants averaged 16 SSPs and 44,000 domains per campaign (ANA, December 2023); the Q1 2025 median was 19 SSPs and 53,799 domains, with 90.3% of impressions on the top 3,000 (ANA Benchmark, Q1 2025).
2. Get direct contracts. The ANA found 52% of brands owned direct DSP contracts, 32% direct SSP contracts and 78% direct verification contracts, and recommended "direct contracts with all primary supply chain partners" (ANA, December 2023). Without a direct contract you cannot demand the logs. The DSP RFP template has the clauses.
3. Get log-level data from every hop that has it. DSP, SSP and verification logs at impression level. Expect resistance: half of the ANA's participating marketers had not been granted access in Q3 2025 (ANA Benchmark, Q3 2025), and some DSPs price the feed separately (Microsoft Learn, updated 22 October 2025).
4. Match impressions across the logs. Join DSP, SSP and verification records on impression identifiers, timestamps and placement. ISBA's match rate went from 12% to 58% between its two studies once the toolkit's data fields list was adopted (ISBA, 18 January 2023); below that, the unmatched share is your own unknown delta.
5. Build the waterfall. Gross spend, minus agency fee, minus DSP fee, minus data, minus verification and ad serving, equals money sent to the sell side; minus SSP fee and supply-side services equals seller revenue; minus spend on invalid, non-measurable, non-viewable and MFA impressions equals TrueAdSpend. Divide spend by quality impressions to get TrueCPM. The ANA's definitions are in its Q1 2025 report (ANA Benchmark, Q1 2025).
6. Compare to the benchmark. Put your TrueAdSpend and transaction cost lines next to the ANA series in the Working Media Benchmark. ANA members can join the benchmark itself; it had 86 participating marketers in Q1 2026 (ANA, 27 May 2026).
7. Act on the two biggest lines. In every published waterfall the two biggest deductions are the SSP fee and the non-viewable or non-measurable share. Supply path optimization and inclusion lists address the first; the ANA's Q3 2025 SSP fee drop from 14.7% to 12.8% was attributed to "continued progress in supply path optimization" (ANA Benchmark, Q3 2025). Viewability floors and measurable-only buying address the second. The operating levers are in 12 ad tech cost optimization strategies.
8. Repeat quarterly. Fees and mixes move; the ANA series moved four points between consecutive quarters more than once. Make the waterfall a standing report, not a project.

The case for owning the data layer that makes steps 3 to 5 routine, rather than negotiating for it each contract, is made in The real cost of cloud DSPs and data ownership and why brands are bringing programmatic in-house. The in-house programmatic hub collects the series; the build option is in How to Build a DSP in 2026.

What is the ad tech tax? The share of an advertiser's programmatic spend retained by intermediaries between advertiser and publisher: agency, ad server, DSP, data, verification, exchange or SSP, supply-side services and publisher ad server. The UK CMA estimated it at about 35p in every pound on 2019 data, with publishers receiving about 65% (CMA, Appendix R, 1 July 2020). The ANA's benchmark measured transaction costs at 27.2% of spend in Q2 2026 (ANA via Fiducia, 6 August 2026).

What percentage of programmatic ad spend reaches publishers? Between half and three quarters, depending on the study. ISBA and PwC found 51% in 2020 with a 15% unknown delta (ISBA, May 2020); the CMA found about 65% on 2019 data (CMA, Appendix R, 1 July 2020); the ANA's Q3 2025 waterfall put seller revenue at 76.2% (ANA Benchmark, Q3 2025). The ANA figure excludes agency fees, which is one reason it is highest.

What does working media mean in programmatic? In the ANA benchmark it means spend that bought a TrueImpression: one that is "non-IVT, measurable for viewability, and viewable" (ANA Benchmark, Q1 2025), with made-for-advertising inventory also excluded in the 2026 definition (ANA, 27 May 2026). That was 45.1% of spend in Q2 2026. Other studies use the term for money reaching the publisher, which is a higher number. Agencies use it for the media line of a marketing budget as opposed to fees.

What did the ANA programmatic transparency study find? The 2023 study found that only 36 cents of each dollar entering a DSP reached a consumer, that transaction costs took 29 cents and media productivity losses 35 cents, and that about $22 billion of an $88 billion open-web market was wasted, across 21 advertisers, $123 million and 35.5 billion impressions (ANA, December 2023). The quarterly benchmark that followed reported 43.9% in Q4 2024 (ANA, 12 December 2024) and 45.1% in Q2 2026 (ANA via Fiducia, 6 August 2026).

How do I measure my own ad tech tax? Get direct contracts with your DSP, SSPs and verification vendor, obtain impression-level logs from each, match the records, and build the waterfall from gross spend to quality impressions. ISBA's second study raised its match rate to 58% using a standard data fields list (ISBA, 18 January 2023); the ANA benchmark offers members a continuous version of the same reconciliation (TAG TrustNet, 17 June 2024). The eight steps above are the full method.

1. Fiducia, "ANA Programmatic Transparency Benchmark: Q2 2026 Working-Media Efficiency Reaches a Record 45%" (ANA release of 6 August 2026). https://www.fiducia.eco/post/ana-programmatic-transparency-benchmark-q2-2026-working-media-efficiency-reaches-a-record-45
2. ANA, "Programmatic Transparency Benchmark: Q2 2026 Benchmark Findings" (26 August 2026, gated). https://www.ana.net/miccontent/show/id/rr-2026-08-programmatic-transparency-benchmark-q2
3. ANA, "ANA Programmatic Transparency Benchmark Reveals Widening Gap Between Top and Bottom Advertisers" (27 May 2026). https://www.ana.net/content/show/id/pr-2026-05-programmatic
4. ANA, "ANA Introduces Brand Safety & Suitability Metrics, Achieves 99.1% Low-Risk Programmatic Spend" (5 November 2025). https://www.ana.net/content/show/id/pr-2025-11-transparency
5. ANA and TAG TrustNet, "Programmatic Transparency Benchmark: Q3 2025 Benchmark Findings" (November 2025). https://s3.amazonaws.com/media.mediapost.com/uploads/Q32025ProgrammaticTransparencyBenchmark.pdf
6. ANA and TAG TrustNet, "Programmatic Transparency Benchmark: Q1 2025 Benchmark Findings" (data to 31 March 2025). https://s3.amazonaws.com/media.mediapost.com/uploads/ANAQ12025Programmatic.pdf
7. Measured, "ANA Finds $26.8B in Programmatic Waste: What Marketers Can Do Now" (15 August 2025). https://www.measured.com/blog/ana-finds-26-8b-in-programmatic-waste-what-marketers-can-do-now/
8. ANA, "ANA Releases 2024 Programmatic Benchmark Study" (12 December 2024). https://www.ana.net/content/show/id/pr-2024-12-programmatic
9. ANA, "Programmatic Media Supply Chain Transparency Study: Complete Report" (December 2023). https://www.adslot.com/wp-content/uploads/2023/12/ana-programmatic-media-supply-chain-transparency-study.pdf
10. ANA, Media Practice, "Programmatic Transparency" page (accessed 17 October 2026). https://www.ana.net/content/show/id/media-programmatic-transparency
11. TAG TrustNet, "ANA, TAG TrustNet Launch Programmatic Transparency Benchmark" (17 June 2024). https://www.tagtoday.net/pressreleases/ana-tag-trustnet-launch-programmatic-transparency-benchmark
12. ExchangeWire, Shirley Marschall, "The Ad Tech Toll Road" (16 September 2026). https://www.exchangewire.com/blog/2026/09/16/the-ad-tech-toll-road/
13. ISBA and PwC, "Programmatic Supply Chain Transparency Study, executive summary" (May 2020). https://www.isba.org.uk/system/files/media/documents/2020-12/executive-summary-programmatic-supply-chain-transparency-study.pdf
14. ISBA, "Second programmatic supply chain transparency study" (18 January 2023). https://www.isba.org.uk/knowledge/second-programmatic-supply-chain-transparency-study
15. UK Competition and Markets Authority, "Online platforms and digital advertising market study, Appendix R: fees in the adtech stack" (1 July 2020). https://assets.publishing.service.gov.uk/media/5fe49625e90e071207e10eff/Appendix_R_-_fees_in_the_adtech_stack_WEB.pdf
16. Adalytics, "How much in fees are ad tech companies charging publishers & advertisers?" (April 2022). https://adalytics.io/blog/adtech-supply-fees
17. AdExchanger, "The Takeaways On Crazy Take Rates" (daily roundup, 14 April 2022). https://www.adexchanger.com/ad-exchange-news/thursday-14042022/
18. eMarketer, "eMarketer's New Ad Tech Tax Estimates Show One-Third of Spending Goes to Intermediaries" (August 2019). https://www.emarketer.com/content/emarketer-s-new-ad-tech-tax-estimates-show-one-third-of-spending-goes-to-intermediaries
19. AdExchanger, "eMarketer: Ad Tech Gets A Tax Cut As Programmatic Fees Decline" (5 August 2019). https://www.adexchanger.com/analysts/emarketer-ad-tech-gets-a-tax-cut-as-programmatic-fees-decline/
20. The Trade Desk, Inc., Form 10-K for fiscal year 2025 (filed February 2026). https://www.sec.gov/Archives/edgar/data/1671933/000167193326000014/ttd-20251231.htm
21. PPC Land, "The Trade Desk launches data marketplace overhaul with AI scoring" (29 September 2025). https://ppc.land/the-trade-desk-launches-data-marketplace-overhaul-with-ai-scoring/
22. Google, Display & Video 360 Help, "Tips for using third-party audience lists" (accessed 17 October 2026). https://support.google.com/displayvideo/answer/6212219?hl=en
23. Microsoft Learn, "Log Level Data Feeds" (updated 22 October 2025). https://learn.microsoft.com/en-us/xandr/log-level-data/log-level-data-feeds