The Take Rate Is Becoming a Competitive Weapon

The first week of August offered a pretty clear preview of what’s coming for media budgets over the next four quarters: platform economics are becoming a competitive battleground.

The Trade Desk reported $715 million in Q2 revenue, up three percent year over year. The stock fell more than twenty percent in after-hours trading. On the call, KeyBanc’s Justin Patterson asked Jeff Green whether he had changed his pricing philosophy and would consider cutting fees to win business back. Green said TTD would always look at it, then made clear he doesn’t think he has to: “I don’t think that the net number has to change dramatically because we’re extremely confident that we’re adding more value than we cost.” Over ten years as a public company, TTD’s take rate has gone up five times and down five times and stayed within a point or two of twenty percent the whole way.

Amazon represents another model entirely. It has positioned DSP access aggressively while monetizing advertising across a much broader ecosystem. Its advertising business generated $19.8 billion in Q2 revenue, up 26 percent year over year.

For most of the last decade, fees in programmatic were something you argued about at conferences and then quietly accepted on the invoice. That’s changing. Pricing is becoming a competitive weapon, and three of the largest platforms in the market are approaching it very differently.

But the important question isn’t which platform has the lowest take rate. It’s whether advertisers can move quickly enough to take advantage when the economics change.

What’s actually being repriced

Green’s counterargument is worth taking seriously.

Amazon can price its DSP differently because Amazon makes money across a much larger advertising and commerce ecosystem. Google’s DV360 can play in a similar way through its YouTube and Adsense supply chains. Neither represents the same proposition as an independent DSP, and Green’s argument is that some of the costs are simply being shuffled somewhere less visible while the open-web inventory on offer gets worse. “These approaches look more like ad networks of 2006,” he said.

He has a point, and it will land anyway. “The fee is hidden somewhere else” is a much harder argument to win with a CFO than “the fee is ten percent.” When working media is the metric under scrutiny, a visible, lower number has an obvious appeal. But a 10 percent take rate isn’t necessarily cheap, just as a 20 percent take rate isn’t necessarily expensive. The question is what an advertiser gets in return.

Glaser landed in a similar place from the sell side: easier direct deals at low middleman margins “will put a ton of pressure on programmatic to outperform.” He also thinks programmatic can. So do we. But it has to prove it now, campaign by campaign, rather than winning by default.

The referee is consolidating too

And just as buying economics become harder to compare, the independent measurement layer advertisers rely on to judge those economics is consolidating too.

If you wanted an independent scorekeeper to settle this, the timing is awkward. Nielsen agreed on August 6 to acquire DoubleVerify for roughly $2.15 billion, or $13.60 a share, representing a roughly 30 percent premium, with the deal expected to close by Q1 2027. Less than a year earlier, Integral Ad Science went private with Novacap for $1.9 billion. Both of the industry’s large independent verification companies now sit inside something else. Nielsen repeatedly emphasized “independent” in announcing the transaction, which tells you they understand the question advertisers will ask.

The underlying reason is unglamorous. DoubleVerify’s’s Q2 revenue grew three percent to $193.8 million while its programmatic activation business declined one percent.

Verification increasingly looks like infrastructure rather than a standalone growth story. And infrastructure tends to consolidate.

Add Teads suing Google on August 4, the fifth major SSP to do so in a year, and the picture is an industry renegotiating who gets paid what, in public, in court, and on earnings calls, all at once.

The OpenAI lesson is about optionality, not AI

The most instructive thing in the market right now isn’t a platform’s roadmap. It’s how quickly the cost of accessing a new channel can change.

Early reports around ChatGPT’s advertising pilot put minimum commitments as high as $200,000. Within months, access had broadened substantially, including lower-cost partner routes and eventually a beta Ads Manager that gave advertisers a more direct way to launch and manage campaigns.

Reported pricing moved quickly too. Early CPMs were around $60, while subsequent reports put them closer to $25.

The exact numbers matter less than the direction of travel: the cost of learning about the channel fell dramatically in a matter of months.

Here’s the part that matters for a media plan:

Testing an emerging channel is always a trade-off. When the cost of entry is high, advertisers have good reason to wait for more proof before committing budget. But as those barriers fall, so does the cost of learning. An advertiser that can make a relatively small investment, measure the results, and decide whether to scale gains something valuable: real-world insight into a channel competitors may still be evaluating from the sidelines.

That’s why optionality matters. You don’t need to know which emerging channel will win. You need the flexibility to test, learn, and move budget when the economics make sense.

What we’d do with this

Four things, and none of them require predicting who’s right about fees.

1. Compare platforms on outcomes net of total cost, not on headline take rate. Walled gardens provide proprietary inventory, data and audiences at the cost of transparency and ecosystem lock in. TTD’s higher take rate comes with a different combination of reach, independence, and open-web access. Those are real trade-offs and they resolve differently by account, objective, and quarter.
2. Keep your data, creative, and measurement portable. If your audience logic and reporting only exist inside one platform’s UI, the fee conversation is already challenging, because you can’t credibly threaten to leave.
3. Build for the ability to move budget inside a quarter, not inside a renewal cycle. Pricing and product capabilities are moving faster than annual commitments do. TTD’s own results illustrate how quickly demand can shift across customer segments, geographies, and media types. Budgets are relocating faster than most contracts allow.
4. Test efficiently and often. Amazon says its Ads Agent can reduce campaign setup from hours to minutes. Every major platform is shipping some version of automation designed to lower the operational cost of buying media.. The cost of running an experiment is collapsing everywhere except, in many cases, your commercial terms.

Where we sit

For advertisers, this makes platform flexibility more valuable than it was two years ago.

Delve Deeper works across major buying and advertising platforms, helping clients reduce the operational and commercial friction involved in using multiple ecosystems. Where available, unified invoicing and flexible commercial arrangements make it easier to evaluate platforms based on what they’re delivering now rather than what made sense when an annual agreement was signed.

That structure was largely a convenience story two years ago. In a market where the largest platforms are actively repricing and repositioning against each other, it’s a strategic one.

If TTD’s neutrality, reach, and capabilities earn their higher take rate on a given account, they should be able to stay there. If a new channel suddenly becomes inexpensive enough to test, advertisers should be able to learn from it without rebuilding their entire media operation first.

The goal isn’t to pick the platform that will win the next five years. It’s to avoid being commercially locked into the answer you picked last year.

The industry spent a decade arguing about transparency. It’s about to spend the next eighteen months finding out what transparency actually costs,who was willing to pay for the flexibility to act on it, and who built enough flexibility to act on what it reveals.

Sources: AdExchanger (The Trade Desk Q2, Amazon Q2, Nielsen/DoubleVerify), Adweek (Teads v. Google), Marketecture (OpenAI ad economics).