The data asymmetry between buy-side and sell-side is humongous and that creates a power position as it relates to price discovery for the buy-side. Especially as TV comes online, because people are protecting their content, there will be more fragmentation. We will have to do more integrations. That creates more barrier to entry and that creates more asymmetry between the buy-side and sell-side.” - Jeffrey Green, Founder and CEO of the Trade Desk, Investor Day 2019

TTD is a buy-side platform for advertising on the ‘open internet’. It allows advertisers to target users on the millions of websites outside of the three main walled gardens: Google, Facebook, and Amazon.

In 2020, eMarketer estimates the US digital advertising market was ~$150bn with ~$30bn spent on the ‘open internet’. The Trade Desk competes with Google and Facebook’s ad network and currently powers over 15% of open internet ad spend. The majority of this spend today is on desktop and mobile web - although connected-TV (CTV) is estimated to double to $25bn by 2025. US linear TV ad spend was around $60bn in 2020 and is stagnating. Over time, as eyeballs move to streaming and CTV, a larger share of linear TV ad spend will transition to OTT and create a new game for media players to compete in.

We interviewed the former Chief Revenue Officer at The Trade Desk (TTD) to understand the CTV value chain and TTD’s long-term strategic positioning.

It’s hard to fully grasp the difference in the level of targeting between linear and CTV. In some cases, linear TV was measured by using a sample of a few thousand households:

“The big fundamental change is that, in linear television and my old Nielsen world, we were working off very small sample sizes because it was very expensive to get an accurate measurement. You had to literally go into a consumer's household and Nielsen would send out a service tech to go into the house, breakout a soldering gun and start tearing apart their television to install a bunch of custom hardware. Because of that cost, the samples you were working with were a relatively small number of individuals. In the United States it was 50,000 households and in smaller European countries it might only be a thousand households. That meant you were dealing with small amounts of data, which led to segmentation and targeting, all being done at very high-level demographics” - Former Chief Revenue Officer at The Trade Desk

The ability of CTV to instantaneously track and target users creates prime advertising inventory. More granular data helps drive return on ad spend and could potentially allow users to better align online spend and offline attribution:

“the big difference with CTV is that because it is delivered digitally over traditional internet protocol, you have a tremendous amount of depth of data. As an industry, we see every stream and individual ad which is delivered, and have tremendous granularity. We move from this very small scientific sample to big data which allows an ecosystem to be created around that big data, to enable very precise targeting and measurement, in theory.” - Former Chief Revenue Officer at The Trade Desk

However, in the advertising world, cooperation is crucial for an efficient market to operate. Advertisers (ad buyers) and publishers (ad sellers) need to agree on a measurement standard to define ad pricing. In linear TV, the industry settled on mediators such as Nielsen and measurement standards such as gross rating points. This leads to a commercial relationship as follows:

“Historically, in linear, the network was on the hook for delivering certain audiences. The advertiser would buy a certain number of ads delivered to a specific demographic group, and the network would deliver that number. If Nielsen, Kantar or whoever the provider in the market said that fewer than that number watched the program, the network would have to deliver more ads on additional programs to make good, they called it. The attribution and delivery model in CTV is very different.” - Former Chief Revenue Officer at The Trade Desk

In CTV, it’s possible for publishers to make all their inventory available for advertisers to programmatically bid on impressions targeting a more granular audience. In theory more effective targeting drives ROAS. Practically, advertisers require a common attribution model to compare returns across different advertising channels.

This is where CTV gets confusing. As we previously discussed, CDLX is also facing this challenge with their closed-loop model for cash-back rewards. The CTV ecosystem is just as complex. Think about the stream of data when you watch your smart-TV:

“CTV is an incredibly disjointed and confusing ecosystem which results in a rocky experience for consumers and advertisers. As an example, think how many ways you can watch HBO Max: purchased through iTunes, on ROKU, on your Samsung TV, and maybe through a Comcast subscription. There are many different players in that chain and it is not negotiated uniformly. It is different for HBO Max than it is for Showtime or Netflix. It is also different on a Samsung TV than it is on an LG. It is different if you have ROKU or Apple TV.” - Former Chief Revenue Officer at The Trade Desk

If the large media companies don’t share data, advertisers will not be able to attribute spend across their channels. The inability to accurately measure ROAS will limit ad spend on CTV. A lack of interoperability within the CTV ecosystem is arguably the biggest limitation to the growth in the channel.

So where does TTD fit into all of this? It’s interesting to consider how different ad markets are structured.

There are fundamental differences between the supply of digital inventory on the web, mobile, and CTV. On the web, there are millions of websites outside of the walled gardens that monetise inventory. Google Search effectively aggregated all of this inventory and layered DoubleClick on top to serve ads to every website globally.

Mobile app inventory is less fragmented with ~5m mobile apps on the App Store and Google Play. Google’s Admob, Facebook, and Applovin are the largest ad networks in the mobile space.

CTV is even less fragmented with hundreds of inventory owners aggregating eyeballs. TTD needs access to the data from all publishers to be an effective demand-side advertising platform (DSP). Although each individual media property may not own enough demand to create an OTT walled garden, the terminal value of TTD in connected TV relies even more on the interoperability of the CTV ecosystem than in the web.

Some media companies are actively promoting their connected-TV DSP. For example, Roku’s OneView claims to offer advertisers more targeted audiences and user level data within the Roku platform. Others could more actively follow suit.

The biggest advantage for TTD is that advertisers typically demand an open market so they can easily compare ad returns. TTD is the perfect partner in this regard: it’s a completely independent DSP that is built for agencies. The company owns no inventory or supply-side platform that could distort the buy-side incentives.

In the long run, unless you really own demand like Google, Facebook, or Amazon, all publishers need to serve the ad buyer. This could lead to an interoperable CTV ecosystem like:

“As an ad agency, if you load your budget into the Trade Desk and set up a campaign which works beautifully when you deliver an ad to Peacock directly via a direct transaction with NBC, then your next ad goes into the Samsung ads business via the Samsung TV operating system and the next ad fires to an advertising in the ROKU marketplace and the next one is with HULU, the Trade Desk is able to see all that, buy all those and reconcile between them. That’s the dream scenario for Trade Desk.” - Former Chief Revenue Officer at The Trade Desk

If the market does behave rationally and cooperates, TTD is in a great position in the value chain to add value to both advertisers and publishers. This reminds us of an interesting comment by TTD CEO in 2019 about the power of a scaled, independent DSP in the value chain:

The buy-side is forever in the power position. I look at what today is 10 million QPS (queries per second), but on the sell-side, they look at a fragment of it. And so the data asymmetry between buy-side and sell-side is humongous and that creates a power position as it relates to price discovery for the buy-side. And especially as TV comes online, because people are protecting their content, there will be more fragmentation. We will have to do more integrations. That creates more barrier to entry and that creates more asymmetry between the buy-side and sell-side.” - The Trade Desk CEO, Investor Day, 2019

The fact TTD can see the vast majority of open internet bidding creates high barriers to entry. The more bidding that flows through TTD, the more efficient the price discovery. This enables TTD to price impressions more effectively than other DSP’s which drives higher ROAS for agencies, the ultimate goal in the industry.

At the other end of the adtech value chain, the SSP seems to have relatively less bargaining power:

"SSPs have some challenges which the DSPs don't have, because DSPs have the advantage where the brand gets a lot of benefit from consolidating with a small number or even a single DSP, because managing your workflow across multiple DSPs is a big pain. It is difficult from a human capital standpoint and you do not get the advantages of scale of concentration. On the SSP side, there is not much advantage consolidating with one SSP for a publisher. Publishers typically work with multiple SSPs and get some advantage from having their inventory available in multiple places. The SSPs have not been able to establish that same type of client consolidation value proposition, but the DSPs have. Therefore, it is easier for the publisher to view them as commoditized, and when you are in that position, the only reason the publisher would consolidate their business with you is price. That results in SSPs having some price pressure which the DSPs do not." - Former Chief Revenue Officer at The Trade Desk

In CTV, the one advantage for SSPs is that the inventory is premium with CPMs at $30-40 rather than ~$10 for open web. The unique nature of premium video programming attracts the highest quality brands. Such premium inventory could require a standalone SSP to ensure the brand gets the protection and inventory quality they wish.

So if TTD does win the buy-side of CTV, how much of the value created does a DSP deserve?

This framing of the economics is particularly interesting:

“In linear television, the data and supporting tech takes 5% of the total spend, once you add in all the different software and data partners. In open web, it is 50%, but CTV works harder than supporting tech in linear television, and the question is where between 50 and 5 it ends up. Nobody knows, but 20% is not an unreasonable number to assume.” - Former Chief Revenue Officer at The Trade Desk

In 2020, TTD’s take was ~20%. This comment by The Trade Desk's CEO at the 2019 investor day also suggests why a 20% take rate is sustainable over time:

“You may say why will [take rates] be even that high? I mean, if they're at 50%, why do you think they'll be as high as 20% to 30%? And it's a lot of triangulation on the lack of fungibility and the uniqueness in the asset that is advertising, where there's no one thing that's exactly the same as another, that makes it so -- the volatility in our industry is so much higher than what it is in almost any other financial or commodities market that we think makes it so that the data differences in those groups make it so that those take rates will continue for a long time.” - The Trade Desk CEO, Investor Day, 2019

Given the relatively consolidated CTV supply side, there could still be pressure on the take rate. However, if TTD does become the scaled, independent DSP, the size of the opportunity will likely outweigh any potential take rate compression.

TTD has an owner operator on a mission to monetise the open internet. Green’s earnings calls are mini-masterclasses in the ad tech landscape. There seems to be a relentless competitive streak within him that enjoys the battle against the walled gardens. This leadership plus the combination of buy-side independence and effective pricing discovery puts TTD in a powerful position within the CTV adtech value chain.

As we discussed, the only potential obstacle is the lack of interoperability of CTV inventory owners. We believe the market will eventually cooperate given no one media owner captures enough demand to become a walled garden.

Well, excluding Netflix. Could this be a huge tail risk? In 10 years time, could Netflix run an AVOD offering? It's unlikely that Netflix becomes the next advertising walled garden, but not impossible.

One final observation that piqued our interest in this space was the fact that the two largest, independent ad networks are run by owner-operators with huge skin-in-the-game.

Applovin, the largest independent mobile ad network, with a CEO that owns 10% of shares out and 25% of the Class B voting shares, and TTD where Jeffrey Green has 50% voting power and is the largest single shareholder. It seems that both CEO’s are on a mission in their own but somewhat similar way: TTD in the open internet and APP in mobile. We will be following both companies closely.