The Trade Desk vs Publicis: Principal Media Trading & Agency Fee Transparency
Executive Profile Hidden
- "It is a huge escalation… the beginning of a crack in the system" of agency programmatic buying; the expert sees "ongoing tension… between ad tech platforms and the holding agencies" and argues "the agencies are a little bit more at risk than The Trade Desk" because of how agencies "make margins from the programmatic media that runs through them."
- "Agencies have this thing called principal media… a trading model": they "buy a bunch of media from publishers… at a bulk pricing discount, and then they will resell it to their advertiser clients at a markup" via "internal trading desk teams" (e.g., GroupM’s "Xaxis").
- Operationally: "all programmatic media goes through Xaxis" and "Xaxis uses the principal media that GroupM purchased at a bulk discount… run it through The Trade Desk or whatever platform… and then take a huge margin out of it"; agencies "broker deals with media publishers to buy a bunch of their ad space at a discount."
- Economics/fees: "It's very opaque… agencies don't want to tell the clients what percentage margin they're taking"; the expert "used to work in programmatic" and estimates principal markups "can be 40% to 50%" but "at least 20%" plus "tacking on a management fee."
- Incentives stack: agencies can "make double money"—principal markups plus "rebates" (expert: "Yes, there is some aspect of that too," but "I do not know that number").
- Why OpenPath is resisted: WPP/Dentsu "were not recommending OpenPath… because it would create more transparency in the margins"; the expert frames Publicis’ audit move as "a negotiating tactic" but also "more meaningful than just a normal cat fight."
- Stakes for holdcos: "GroupM… make $1 billion" from principal; "Publicis makes more than 50% of their costs" from it; "Omnicom earns $4 billion"—"that's the reason for this public dispute."
- Counterweight/moat-like driver for TTD: JBPs—"over 50%" of advertisers have them; if a client commits multi-year spend, it "de-risks" agency pushback; without JBP, budgets are "more susceptible" to agencies steering to "Google… Amazon."
- Core demand drivers: "Trade Desk… give them good results," with "a lot of CTV inventory" and retail media connections ("Walmart, Target, and CVS"); cited "95% retention rate."
- Risks/inflection: "advertisers are going to ask agencies for more transparent fee structures," potentially weakening agency leverage and accelerating more JBPs for transparency.
- Near-term outcome view: "50% to 60%… with the JBPs… is secure"; overall "90% to 95% of the billings are going to stay with Trade Desk" and "things will stabilize"; other agencies "not… like Publicis" publicly, but will "just not… recommend OpenPath."
Interview Transcript
I was wondering, with this entire saga going on, how does the relationship between agencies and The Trade Desk work? Who has more say—is it the client or the agencies? Can you help me explain that?
It depends. What is my take on it? I think it is a huge escalation. It is an ongoing tension that has been going on between ad tech platforms and the holding agencies. I think this is the beginning of a crack in the system of how agencies have been purchasing programmatic media for almost a decade now. Overall, I think the agencies are a little bit more at risk than The Trade Desk because of how it is structured and how the agency makes margins from the programmatic media that runs through them.
Can you help me explain this margin structure? For example, let's say I am a client spending $100 million. From what I understand, The Trade Desk takes 20%, but agencies might have a negotiated lower take rate. How much margin do they charge? For example, if a client is spending $100 million, how much do I have to pay to you as an agency versus The Trade Desk?
I do not know how much the agency management fees and all of that are, but I know how agencies have been buying programmatic media and how that works. There is this thing called principal media. Have you heard of that?
No. Can you help me explain that?
Agencies have this thing called principal media. It is a trading model. Agencies will buy a bunch of media from publishers and get it at a bulk pricing discount, and then they will resell it to their advertiser clients at a markup and run it through their internal trading desk teams. For example, WPP's GroupM has a trading arm called Xaxis. All of our clients are told they have to run programmatic through Xaxis because they are the programmatic experts and have more technology, capabilities, and expertise to run the programmatic media. The team at Xaxis will run on The Trade Desk, Amazon, Google, and then run it on that principal media that they purchased.
Can you help me give an example? Let's say I am P&G or Unilever, and you are WPP and you are buying in bulk. What I have understood is you are buying from publishers at a discounted rate, and then you are selling it through Xaxis. Am I understanding that correctly?
Let me clarify. Let's say P&G is the advertising client. They want to run programmatic media. Me at the agency will say all programmatic media goes through Xaxis, which is part of WPP. Xaxis uses the principal media that GroupM purchased at a bulk discount. They run it through The Trade Desk or whatever platform, but on the bulk media—the websites—and then take a huge margin out of it.
Can you help me clarify that a little bit? I am not able to understand from whom you buy at a discount rate?
The agency—WPP, Publicis, Omnicom—will broker deals with media publishers to buy a bunch of their ad space at a discount.
How is it different from a client directly spending with you versus you buying from these publishers? How is that split if you have an understanding?
The advertisers cannot buy the media directly. They have to go through either The Trade Desk or through the agency's trading desk and then through the media.
I understand that. But let's say, for example, you are WPP. I am an auto client, let's say Hyundai or BMW. I have to spend $100 million. I come to you and ask you to help me. I have to spend this $100 million, and you figure it out. Let's say you allocate 50% of that—$50 million—to programmatic. How does that work? Am I going to dictate that I want to spend it on The Trade Desk, or is it the case that you would decide how it will go?
It depends on the client. The Trade Desk has been making these things called joint business plans directly with clients like Hyundai, for example. They will say, "Hey, Hyundai, let's partner together. If you agree to spend this much money on our Trade Desk platform over several years, then you get beta product testing and other benefits." I have read that over 50% of The Trade Desk's advertisers have joint business plans. That de-risks them in terms of having the agency say, "No, don't run with The Trade Desk," because they already have these contracts in place where they need to spend a certain amount with The Trade Desk in order to fulfill the contract.
So let's assume they have a 50% plus joint business venture. Let's say it's 60%, and 40% is with agencies. Out of that 40%, how much depends on the client versus the agency? For example, if an agency has to spend $10-20 million on Trade Desk in a given month - I know they spend more than this, but this is just an example - who is deciding that? Is it the agencies or the client?
It can come from either of them. If they have a joint business plan already and they have a target of spending $15 million on Trade Desk, then the client will say, "We need to spend $15 million this year on Trade Desk." That's very secure. But for the remaining 40% who don't have that JBP in place, they might be more susceptible to the agency saying, "Why don't we try spending on Google instead? Or why don't we try spending on Amazon instead?" Or some other tactics.
What is in your mind regarding the second question?
Some people think that Publicis is using this whole spat as a negotiating tactic for their annual contracts with Trade Desk. It was unusual because usually one of the big four accounting firms does the audits and the billing audits, but this time Publicis chose FirmDecisions, which has never done a DSP audit before. It was very out of the norm. I think it's something more meaningful than just a normal cat fight. Agencies, ad tech vendors, and everyone is trying to take a cut of the ad dollars, but now it's a very public spat that everyone is digging into the details of. I think Trade Desk is less at risk to lose from this public spat. It's more the agencies that are about to lose out. There's a lot at stake for GroupM - they make $1 billion from principal media trading. Publicis makes more than 50% of their costs from principal media. Omnicom earns $4 billion from this principal media margin-cutting method. That's the reason for this public dispute.
In principal media buying, what is your rough estimate of a markup fee? For example, what I'm not able to understand is, as a client using Trade Desk, I have to give Trade Desk 20%. Let's say it's a $60 CPM. I have to give Trade Desk 20%. The first question is, in that 20%, is the agency markup included? If not, how much would I have to pay on that $60 CPM to agencies?
It's very opaque. That's the whole reason that agencies don't want people to use Trade Desk, because Trade Desk is trying to make things more transparent. Agencies don't want to tell the clients what percentage margin they're taking from every single ad impression being served. I don't actually know what the markup is, but it can be substantial. Let's say a client runs a $30 CPM. 20% of that is going to Trade Desk, but we don't know what other percentages were tacked on by the media agencies.
What is it for WPP? I think it would be standardized for most of these agencies. How much is it for WPP from your rough estimates? How much do you make from this trading?
I used to work in programmatic. It can be 40% to 50%, but I'm going to say at least 20% because they're getting the bulk discount of the media that they're running on. They're also tacking on a management fee because you're running through Xaxis, who are experts. At least 20% is going to the agency.
So this entire conversation between Publicis and Trade Desk - Trade Desk is more transparent regarding fees, and agencies don't want the clients to know how much they are making out of it because they are at more threat?
Yes, that's right. I think I said 20% of Trade Desk clients are direct on Trade Desk. The reason that Dentsu and WPP were not recommending OpenPath is because it would create more transparency in the margins that the agency is taking.
That's surprising for me. I had some understanding about this, but never thought that agencies are making so much money from Trade Desk?
I was thinking about it, and in the financial industry, there's a fiduciary duty to be transparent and make money for the client. All the fees are very disclosed. But in the ad tech world, everything is opaque and everyone's trying to take a cut of it. When you have a $100 million budget, who knows what huge percentage of it is going to fees and ad tech vendors and verification vendors and agencies. There's tons of fraud in the marketplace too. Trade Desk is revealing all of these things if they continue with OpenPath and what they're trying to do. That's the main alarm that's happening.
Coming to the third question, how much do you think that this holding company relationship actually matters in practice? Is there something that holding companies have that's special, like a special tech rate or something?
The holding companies are the gatekeepers to billions of dollars of ad spend across thousands of advertisers. They're the reseller for Trade Desk. The reason is because Trade Desk can't make direct partnerships with thousands of advertisers - they would need a huge sales force to do that. It really does matter in terms of working with the agencies. But as I said, they're reaching out to all these advertisers to make those joint business plans. I think that helps de-risk it and helps them control the narrative versus just relying on agencies to do whatever they want.
Is this just a negotiating tactic, but at the end, agencies have to rely on Trade Desk and come back to them because their clients have already allocated the budget for this year? They cannot simply move away from that? For example, in this dynamic, I know you mentioned it is a negotiating tactic, but because at the start of the year most of these clients have allocated their budgets on Trade Desk, is it the case that even whatever Publicis, WPP, or Omnicom says, they have to come back to Trade Desk because the budget is already planned?
Yes, because they have that contract in place, they are going to spend to that minimum amount that they agreed with Trade Desk. Trade Desk is not bad. Trade Desk does give them good results. There is a lot of CTV inventory, there are many retail network connections, things like that. Advertisers want to spend on Trade Desk.
Agencies also earn from rebates. If you spend a lot on Trade Desk, you get that rebate. So agencies are making double money. You are buying at a discount and making that cut, and also getting rebates from Trade Desk?
Yes, there is some aspect of that too, where agencies are also incentivized by Trade Desk to run more on their platforms.
What is a typical rebate, is it 2% or 3% from your understanding?
I do not know that number.
Coming to the fourth question, what share of clients' spend across Publicis do you think would stay with Trade Desk?
I think for the 50% to 60% or so that have the JBPs, that money is secure. That will not change and it is fixed. But for the remaining, it is mostly based on performance. If Trade Desk is showing better results than DV360 or Amazon, they are going to keep with them because, as I said, Trade Desk has a lot of the premium CTV inventory. They also did a lot of newer retail media connections with Walmart, Target, and CVS. That is very valuable and that is kind of to combat Amazon and their retail data, shopper data pitch. I think there is a ton of value on Trade Desk. It is said somewhere that Trade Desk has a 95% retention rate of their clients. I think that speaks to how valuable it is to the clients.
Out of those people who do not have those JBPs, the 30%, do you also think that out of that 30%, 90% to 95% of clients would stay intact?
Yes, I think so. Trade Desk is very valuable to advertisers, and just because an agency says do not use it, they are still going to want to use it. I think that is going to be true for all of the bigger national brands. But if a client is more green in the whole advertising and ad tech space, maybe they might be swayed to go a different route, like DV360 or something. But I think for the most part, big brands know the value of Trade Desk.
How much share do you think Publicis spend has on Trade Desk? Is it 10% of their billing or more than that? And how much is WPP if you have to take a guess?
I think Publicis is obviously one of their large clients. I have read that it is around 13% or so.
And what about WPP and the other ones?
They are probably around 10%, 9% or so.
What discussions are you having in your firm, like WPP and the related sister subsidiaries you have, how you view Trade Desk?
I think for now the conversation is to keep business as usual, but from the leadership, we are getting more push to keep feeding into our trading desk, our internal trading desk. There is an incentive program for even the most junior associates to push more dollars towards the internal trading teams.
But at the end, whatever you spend on your internal trading desk, that person can also spend on Trade Desk, right?
They can, yes. The internal trading desk runs on Trade Desk or DV360 or Amazon.
Do you also think, like Publicis, that you are going to negotiate the contract, or are you content with the existing relationship?
I think there will continue to be a relationship. I do not think they can go without each other, but I think what will happen is that advertisers are going to ask agencies for more transparent fee structures. I do not think they are going to stop using Trade Desk outright. That sounds insane and it sounds dumb, but it is going to make agencies have to be more transparent. It might even propel some advertisers to create joint business plans with Trade Desk so that they can have more transparency.
Do you think we have missed something or we have covered most of it, or is there something that you want to share that I might have not asked and you feel is very relevant for this topic?
No, I do not think so. I think overall, I would say that 90% to 95% of the billings are going to stay with Trade Desk. I do not think it is going to be as bad as the stock price was showing. I think things will stabilize.
Are you buying stock at this level then?
No, well, I am not involved in that, but I think Trade Desk is well positioned still. I know that was a huge drawdown in its stock price, but I think Trade Desk is well positioned, it has good footing and it has good performance. So I do not think it is going to disappear just because Publicis said something like that.
Do you think other agencies would do the same as Publicis? Would they come and charge at Trade Desk, try to renegotiate, and do something similar to what Publicis has done?
I don't think so. Publicis did what it did, but the other agencies are going to take a more tactful approach. They're probably just not going to recommend OpenPath, but they're not going to say they don't recommend The Trade Desk at all. That was absurd.
Because OpenPath is more transparent, so they don't want to get exposed to the public?
Exactly.
Related Content

The Trade Desk vs Amazon DSP: Open Web Budget Shifts & CPG Advertiser Testing
Media Director at WPP

HBM, DRAM & NAND: Memory Hierarchy for AI Workloads
Former Vice President at Samsung Semiconductor

Dick's Sporting Goods: Merchandising Strategy & Execution Culture
Former Vice President & General Merchandise Manager at Dicks Sporting Goods

Atlassian: AI, Rovo, & the Teamwork Knowledge Graph Advantage
Former Head of Product Management, Enterprise Solutions at Atlassian
© 2024 In Practise. All rights reserved. This material is for informational purposes only and should not be considered as investment advice.