Programmatic transparency
Three Years of Clean-Up, Still Under 45 Cents: What the ANA Transparency Tracker Tells Us
Since 2023 the Association of National Advertisers has measured how much of the programmatic dollar actually works. The numbers have improved: waste on made-for-advertising sites has collapsed and more money reaches publishers. But even the best recent reading leaves more than half of each dollar lost before it reaches a real consumer.
Why a tracker matters
One-off studies make headlines, then fade. Tracking the same measures over time is what changes behaviour. That's what makes the Association of National Advertisers' programmatic work unusual. Starting with a landmark study in 2023, and continuing through a quarterly benchmark run with TAG TrustNet, the ANA has built a running record of how efficiently US advertisers' programmatic money is spent [3][4][6].
Read together, the releases tell a story of real progress and a stubborn ceiling.
2023: the baseline
The ANA's first look, published in June 2023, sized the US open-web programmatic market at $88 billion. It estimated that up to $20 billion, about 23% of it, was wasted [3].
The study found some striking details [3]:
- Made-for-advertising (MFA) websites, built mainly to show ads rather than serve readers, accounted for 21% of impressions and 15% of spend.
- The average campaign ran on 44,000 websites. That's far more than any brand could meaningfully check for quality.
- The sample covered $123 million of spend and 35.5 billion impressions from 21 member companies.
The complete report followed in December 2023 with the number that became the study's signature: only 36% of every dollar that enters a DSP effectively reaches the consumer [4][5]. It put the available efficiency gain at $22 billion [4].
That 36% is worth understanding. It isn't the share reaching publishers. It's what remains after fees and after waste such as MFA inventory, invalid traffic and ads that were never viewable. It measures how much of the dollar actually did its job.
2024: measurable improvement
A year later the picture had improved. In its December 2024 study, the ANA reported that 43.9% of spend now reached consumers, an improvement of 7.9 percentage points, or an extra $79 of working media per $1,000 spent [6].
The drivers were concrete [6]:
- The MFA share of spend fell from 15% to 6.2%, and the median advertiser was down to 1.1%.
- The average number of domains per campaign roughly halved, from 44,000 to about 22,600.
- The ANA put the productivity gain at around $8.2 billion across a $104 billion open-web marketplace.
In other words, when advertisers were shown where their money was going, many of them acted. They cut obvious waste, tightened site lists and moved spend towards cleaner inventory.
2025: the gains slow
By 2025 the easy wins had largely been taken.
- Q2 2025: the ANA and TAG TrustNet reported that $26.8 billion in global media value was still being lost every year to inefficiency [7]. MFA had fallen to a median of 0.8% of spend, down from 15% two years earlier.
- Q3 2025: the share of ad spend reaching publishers rose to 47.1%, up 11 points since 2023, and MFA domains had fallen to just 0.39% of spend [8].
That last number needs care. Reaching the publisher (47.1%) is a fees-only measure, similar to the 51% in the UK's ISBA/PwC study. It doesn't subtract waste. It shows fees have come down a little, not that nearly half of every dollar is working.
2026: a widening gap between leaders and the rest
The ANA's Q1 2026 benchmark shifted focus from averages to the spread between advertisers [9]:
| Measure | Top performers | Others |
|---|---|---|
| Programmatic spend converted into qualified impressions | 54% | 32% |
| Waste | 19.0% | 38.4% |
| True cost per thousand qualified impressions | $7.46 | $19.04 |
The headline "true ad spend" index for the market stood at 43.3% [9].
This is the most useful finding in the series. The best-run programmatic operations now turn just over half of each dollar into impressions that meet quality standards. Everyone else turns about a third. And the less disciplined group pays more than two and a half times as much for each qualified impression.
What the tracker proves, and what it doesn't
It proves the problem is real and measurable. Three years of data from a major trade body show that a large share of programmatic spend doesn't work, and that the amount lost is in the tens of billions of dollars a year [4][7].
It proves transparency changes behaviour. The collapse in MFA spend, from 15% to under 1%, happened because advertisers could finally see it [3][8].
It suggests a ceiling. Even top performers lose close to half of each dollar. The biggest remaining losses aren't obvious junk sites that can be blocked. They're built into how the chain is structured and paid.
It doesn't cover everything. The ANA benchmark focuses on open-web programmatic, largely in the US. It doesn't measure app install fraud, affiliate networks or reselling of direct-app inventory, which are often less transparent still.
Why the ceiling exists
Three structural features explain why clean-up alone stalls.
- Fees are charged as a percentage of spend. Removing waste lowers the bill, but it doesn't change the fact that every layer earns more when more money flows through it.
- Supply paths stay long. The ANA has repeatedly recommended consolidating towards a smaller set of trusted sellers and SSPs [4], but many chains still route spend through several hops.
- Optimisation targets the wrong thing. Impressions and clicks are easy to count and easy to bill. The customer's lifetime value is not, so most of the chain doesn't optimise for it.
Each of these is a design choice, not a law of nature.
Lessons for marketers
- Measure what reaches the consumer, not just what reaches the publisher. The two numbers differ by several points, and the gap is waste.
- Benchmark against the leaders, not the average. The ANA data shows a 22-point gap between top performers and the rest [9]. That gap is your opportunity.
- Shorten the chain. Fewer, better-vetted sellers consistently deliver more working media per dollar [4][6].
- Change what you pay for. As long as partners are paid on volume, volume is what you'll get.
The takeaway
The ANA's tracker is the best evidence available that ad-tech can improve when advertisers demand it. It's also the clearest evidence that improvement has limits. After three years of clean-up, the market as a whole still turns only about 43 cents of each programmatic dollar into real advertising [9]. That's almost exactly where an independent model of the supply chain lands: around 45 cents.
The next gains won't come from blocking more bad websites. They'll come from rebuilding how media is bought, and how the people buying it are paid.
References
[3] ANA. Programmatic Media Supply Chain Transparency Study: First Look. June 2023. https://www.ana.net/content/show/id/pr-2023-06-programmaticstudy
[4] ANA. Programmatic Media Supply Chain Transparency Study: complete report. December 2023. https://www.ana.net/content/show/id/83522
[5] ANA. Programmatic transparency hub. https://www.ana.net/content/show/id/programmatic-transparency
[6] ANA / TAG TrustNet. Programmatic Transparency Benchmark: 2024 study. December 2024. https://www.ana.net/content/show/id/pr-2024-12-programmatic
[7] ANA / TAG TrustNet. Programmatic Transparency Benchmark: Q2 2025. August 2025. https://www.ana.net/content/show/id/pr-2025-08-programmatictrans
[8] ANA / TAG TrustNet. Programmatic Transparency Benchmark: Q3 2025. November 2025. https://www.ana.net/content/show/id/pr-2025-11-transparency
[9] ANA / TAG TrustNet. Programmatic Transparency Benchmark: Q1 2026. May 2026. https://www.ana.net/content/show/id/pr-2026-05-programmatic