Project Sunshine: TTD buys Sincera

We’ve released the agenda for MarketectureLive. We’re deep-diving on identity, AI, and more.

The Vendor Interview: VidMob

Marketecture Vendor interviews are free for a week, then require a subscription. Read more about subscription options at marketecture.tv.

Marketecture Podcast: Dissecting the Vistar / T-Mobile acquisition

We bring on our “OOH guy” Tim Rowe to help us understand T-Mobile’s acquisition of Vistar. Hint: It’s about unit economics, not data.

Why the Sincera acquisition by The Trade Desk is the most important M&A story in a while

Disclosure: I was an investor, advisor, and sometimes consultant to Sincera.

Jeff Bezos famously characterized Amazon’s strategy as “your margin is my opportunity.” I’m going to put words in the other Jeff’s mouth — Jeff Green’s to be specific — and say “Your ad tech tax is my opportunity.”

Last week’s announcement the The Trade Desk would acquire Sincera should send ripples across a wide swath of the Lumascape as the potential implications of the company’s rare acquisition unfold.

Internally it was called Project Sunshine, which is telling. The best disinfectant, after all. The Trade Desk is going to use Sincera to enhance its already strong position with supply, to offer its customers better quality data for buying, to feed its AI, and to reduce the fees customers pay on each transaction.

What does Sincera do?

Sincera markets itself as a “metadata company” which is really not very helpful in trying to understand anything. Sounds cool though, and they exited, so who am I to judge?

The easiest explanation is that Sincera operates a very sophisticated web scraper/spider, that is advertising aware. Web spiders are nothing new, and many old-timers may remember the original Alexa service from Amazon — not the voice assistant — but a directory of tens of thousands of sites with useful information. Sincera is like that, but deeply aware of how programmatic advertising works. The Sincera data set can answer questions like:

• Which companies are in the header?

• Which header wrapper(s) are implemented?

• Which alternative IDs are being passed?

• How many ad slots are on the page, and where are they?

• How often are the ad slots refreshed?

• How does the ad load and configuration change when a user has a cookie vs not?

• Etc, etc, you get the point.

I would hope if you’re reading this newsletter you could immediately appreciate the value of this data to the largest independent trading platform, but if not I’ll walk through my thoughts.

Ad quality signals

The big verification companies, IAS and DoubleVerify, make the majority of their revenue pre-bid, meaning they sell data to advertisers to use for targeting, rather than for auditing and reporting. In the most recent quarter, 57% of DV’s revenue came from “activation” vs “measurement” (10Q). The most popular activation products are:

• Brand safety

• Made for advertising (“MFA”)

• Viewability

• Fraud detection

The Sincera data can immediately be adapted to build brand safety and MFA detection products for TTD. The Sincera data already powers the MFA detection for IAS (announcement). IAS basically takes all the varied signals from Sincera, turn them into a targeting segment (e.g. “No MFA”), and take a hefty CPM fee.

As friend-of-the-pod Chris Kane has pointed out, whether a site is MFA or just aggressive at monetization is in the eye of the beholder, so arguably taking the detailed telemetry from Sincera and letting customers fine tune their definitions of MFA is a big step-up versus paying a third party to figure it out in a black box. Imagine creating “dynamic include lists” that could be continually auto-updated with targeting criteria on a per-customer basis, instead of just a binary yes/no generated without any transparency.

Brand Safety is also an area where the big verifiers are under fire. Their antiquated keyword-based approach leaves a lot to be desired, with notable bad outcomes, like marking our queen, Taylor Swift, as brand unsafe, for her Woman of the Year profile in TIME. Sincera does not currently spend a lot of effort on the content of the pages they spider, but this is a pretty obvious next step.

Viewability is a bit of a tougher nut to crack because the Sincera data does not include feedback on live user sessions and scrolling behavior. However, as a starting point they do know which ad slots are viewable on different screen dimensions and viewable panes, which may actually be more predictive than the pre-bid products currently available.*

If we were going to sum up the situation, I’d say the verification companies are in deep doo-doo. TTD is probably their number one distribution partner. I’m going to make a very wild guess that something like 30% of their Activation revenue passes through this DSP. When the ANA studied the overall supply chain of programmatic advertising they estimated that 9% of advertisers’ budgets went to “targeting and verification”, which was about 25% of the total overhead (aka “ad tech tax”). Below is a helpful, although quite old, eMarketer chart.

A lot of this revenue is available to TTD to attack, both on quality and price grounds. And the argument that you need “independent measurement” does not hold, since these are not measurement products — you can continue to use whomever you want for post-bid.

*: Last time I looked into this, IAS’s pre-bid viewability detection could not differentiate between identically-sized slots on the page. Please let me know if this has changed.

Telemetry on the open web

Another way to look at the Sincera data is that it gives TTD the best data about the open web outside of having their tags on the page. Google has the unique advantage of owning the publisher ad server (for now), giving them insights into the supply paths and ad tag behaviors prior to ever seeing an opportunity to bid. In one example that came up in the recent antitrust trial, Google’s gTrade team identified publishers who were giving “first look” to Criteo and others, and systematically downgraded their bids from AdWords and DV360 to punish this behavior.

While TTD will not have as comprehensive and real-time of a view of publisher behavior, it will be able to deeply understand the supply paths publishers utilize (or abuse) in a way that is much more detailed and meaningful than can be inferred from the bidstream alone. Some of the insights that I’m speculating could be useful for TTD:

• Understanding supply-path duplication based on header wrappers and adapters in use;

• Seeing how UID2 and other alternative IDs are being implemented to avoid “ID stuffing” or other signal corruptions;

• Understanding ad slot refreshes or other lower-quality behaviors;

• Discovering any brute-force auction manipulations, like sending in higher bids to the ad server from prebid (don’t know if people do this anymore?)

• etc.

Project Sunshine indeed.

Feed the AI beast

One of the key advantages Google and Meta have over everyone else in advertising is the scale of their data. With signal loss a lot of ad decisioning is moving to AI models like PMax, where no one really knows exactly what the model is thinking.

While not a panacea for TTD’s natural disadvantage on this issue, having terabytes of data about inventory quality is not a bad thing on this front. If I was an AI researcher working at the company I could not wait to get my hands on this mountain of data, to let my models figure out which signals were most predictive of advertiser outcomes, improved margins, etc.

OpenPath and UID2

The two most important initiatives for TTD are OpenPath (access to supply) and UID2 (access to identity). Sincera gives the business development team a goldmine of data to help both these efforts. Sincera can tell them every source of supply and every ID currently implemented on a per-site basis and help them grow their footprint. The data can also be used as an early detection signal for broken or sub-optimal integrations, thus helping the overall quality of the network.

Overall, a big deal!

Reading list

• mParticle and ROKT merge (link)

• Amazon RMN announcement from last week (link)

• DoorDash and TTD partnering on audience extension (link) 

• Mohegan Sun launches vertical ad net using Liveramp (link)

• Criteo Appoints Michael Komasinski as CEO (link)

• DV finds an AI content mill (link)

• Google search market share declines just a tiny, tiny bit (link)