• Marketecture
  • Posts
  • Honey, have you seen The Trade Desk remote?

Honey, have you seen The Trade Desk remote?

Happy Labor Day to those in the US, and “I’m sorry you couldn’t get Oasis tickets” day for those in the UK. We’re catching up on the news The Trade Desk might be building a TV OS.

Ad

Google won’t kill cookies—will users? Share your views on this u-turn and other key addressability topics. Take ID5's State of Digital Identity survey now!

Podcast: Adam Epstein from AdMarketplace talks Google search remedies

Adam is a lawyer and former Federal court clerk so he brings knowledge of the search business and the law.

Listen to the pod now:

Honey, have you seen The Trade Desk remote?

As if Ad Tech Twitter didn’t have enough to get worked up about, last week a blogger and reporter none of us had heard of broke a story that The Trade Desk has been secretly working on an operating system for TVs. Everyone’s first reaction was, more or less, wut?

Let’s do some idle speculation.

The fight against disintermediation

In our discussion of the Google search case we pointed out the key reason the Chrome browser and Android OS exist is to allow the company to control its distribution and avoid disintermediation. Google executives were students of the 1990’s “browser wars” and saw how your lifeblood could be cut off by a rival farther up the distribution stream with your customers.

TTD is betting its future on its ability to become the primary buying system for the future of TV. There’s no way to get to that $50 billion valuation unless you assume that a big portion of the roughly $60 billion TV market starts flowing through TTD’s pipes.

Let’s look at the pipes.

  • OEMs like Samsung, LG, Vizio all have active advertising businesses

  • OSes like Roku and Amazon Fire have active advertising businesses

  • Virtual MPVPDs like YouTubeTV and Hulu+ have active advertising businesses

  • And, of course, both FAST and SVOD channels are almost all ad supported (except AppleTV+, for now)

So, to massively oversimplify the situation, TTD’s viability is dependent on a complex set of companies, nearly every one of which would prefer to sell ads themselves and keep the margin, rather than continue letting 20% of the value go into someone else’s hands.

The risk to TTD isn’t theoretical. You can already see the distribution strategies emerging against the company’s interests:

Walled gardens: Amazon, Google (plus TikTok, Snap, Meta to be fair)

Hedged gardens (preferring channels other than programmatic): Hulu, Roku

Unknown futures: Walmart/Vizio, Netflix, Apple

What would an OS get Trade Desk?

Keeping in mind that we have no idea whether a) the story is true at all; b) the thing they are building is an OS to license to manufacturers; or c) the thing is more like a Roku/Fire stick sold to consumers; or d) something else entirely … let’s speculate freely! For the sake of readability we’re just going to call this thing the “TDOS”.

The first thing TDOS gives you is automatic content recognition (“ACR”). The closer you are to the “glass” the more of the consumer’s behavior you can see and analyze, resulting in better ad targeting and outcomes measurement. We know TTD cares a lot about this because a) its obviously valuable; and b) they have licensed this data from many sources. This data is not cheap, and independent ad tech companies might spend high single-digits to tens of millions of dollars to get the data, if they even can given exclusivity rights.

But let’s get back to disintermediation. In a world where some meaningful percentage of TVs were using TDOS, the company could break open some of the walled and hedged gardens. The standard in the industry right now is that the OEMs/OS distributors get resale rights on two minutes of ad time per hour from the FAST and MVPD services. On the TDOS, if those apps wanted placement (which surely they would) the walled garden would start cracking. You could also imagine some interesting horse trading where an app might not be required to give up resale rights on TDOS and instead give more broad programmatic access across OSes.

Side note: We asked some experts and are not confident of the answer to the question of whether Fire or YouTube have given resale rights to OSes. We are certain that you cannot buy any YT inventory programmatically, but we think Google may give some rights away but execute through their own ad servers or the convoluted integration with FreeWheel. If anyone knows, please chime in.

How much can they afford to invest?

You can’t just build TDOS and hope for the best. You need to get enough distribution with consumers to gain leverage on the TV market. How much can TTD spend to get TDOS distribution?

The break-even point to get a household on TDOS is something like:

LTV = (hours of streaming time/month) x (ads/hour) x (% of inventory unlocked) x (CPM/1000) x (TTD take rate) * (length of install in months)

Let’s do some really rough estimates:

Metric

Estimate

Notes

Hours of streaming time/month

150

Varying estimates, 40% of TV watching

Ads per hour

10

Traditional TV is ~28 ads/hour, CTV is less than half that

Percent inventory unlocked

15%

Prime is ~4% of viewership, YT is close to 10%, but also let’s assume they get better access to ads in all other apps

CPM

$15

Seems like a fair going rate, correct me if I’m wrong

TTD Take rate

15%

Conservative estimate

Length of install

60

Average TV replacements is 5-7 years

According to this extremely scientific estimate, TDOS should be willing to spend $30.38 per TDOS install (not adjusted for time-value of money):

LTV = 150 × 10 x 15% x $15/1000 x 15% * 60 = $30.38

We can do a more generous estimate, but increasing the inventory unlock to 25% and the take rate to 20%:

LTV = 150 × 10 × 25% x $15/1000 × 20% * 60 = $67.50

These LTV figures are not strictly the amount TTD would pay to break even as we haven’t considered the cost of sales, maintenance, etc along with the large amount of R&D that might be going into this project.

Let’s say after everything else, TTD is willing to pay manufacturers $20 per set to install their TDOS — is that enough? I have zero data or experience with this. So more SWAGging. Vizio, as a public company, made $502 million in revenue in Q4 2023, and sold 1.3 million smart TVs, so they grossed (including ad and data revenue) $386/TV. If these numbers are a good benchmark then TTD could subsidize maybe ~10% of the gross cost of a TV to get installed as the OS.

This analysis is sloppy

Yes, I agree. I’m literally winging this based on some Perplexity searches and a rudimentary knowledge of the TV ecosystem. I want to get the conversation started. What do you think?

Reading list

  • AdExchanger covers how much Google hates header bidding (link)

  • Perplexity’s pitch deck has everyone yelling at each other (link)

  • Criteo CEO Megan Clarken stepping away in next 12 months (link) (Digiday)

  • AdLudio goes into administration (link)

  • Tim Vanderhook on the untold story of Myspace (link)

  • Yelp sues Google for antitrust in local search (link)

  • Brian Wieser calculates Google’s network revenue (link)

Reply

or to participate.