Which kinds of measurement proof do CTV platforms describe on their pages?

What a Proof Card shows

Every platform on this site has a Proof Card: four tiles that are filled when the platform's own pages describe that kind of proof, and hatched when the pages we read do not. The four kinds are:

1. Outcome attribution: linking ad exposure to visits, conversions, revenue or installs.
2. Incrementality or lift: holdout, geo or lift measurement of what the ads added.
3. Named measurement partners: third-party measurement or attribution partners named on the vendor's pages.
4. Data access: an API, log-level export or assistant access to campaign data.

A filled tile is not a quality rating. It means the capability is described, with a source link.

The pattern across eight platforms

One of eight describes all four; three describe none on the pages we read.

What the filled tiles say

• Outcome attribution is described by Vibe (pixel, 1 hour to 7 day window), MNTN (Verified Visits), tvScientific (deterministic household matching, online and offline), Tatari (CPV, CPI, CAC and ROAS with multi-touch attribution) and StackAdapt (cross-device attribution).
• Incrementality or lift is described by Vibe ("incrementality testing built in"), tvScientific ("incremental lift analysis"), Tatari ("Measure the true incremental impact") and StackAdapt (brand lift and sales lift).
• Named measurement partners appear on Vibe's integrations page (Haus, Northbeam, Prescient, Triple Whale and others) and MNTN's integrations list (AppsFlyer, CallRail, DoubleVerify and others).
• Data access is described by Vibe (API and MCP included with every account), MNTN (Flexible APIs for BI), and Tatari (log-level exports, reporting APIs, server-to-server sharing).

What the hatched tiles mean

A hatched tile means the pages we read did not describe that kind of proof. For Amazon Ads, Disney Campaign Manager and Roku Ads Manager, the pages we read focused on access, pricing, inventory or creative rather than measurement. These companies may offer measurement through other pages, sales teams or partners. A buyer should ask, and we will update the cards when a vendor publishes more.

Limits of this mapping

The cards record whether a capability is described, not how well it works. Two filled tiles can hide very different methods: a pixel with a short window and a deterministic household graph both count as outcome attribution. The cards also depend on which pages we read. We read each vendor's main product, measurement, inventory and pricing pages, not every help article or sales document. A capability described elsewhere would not appear. Finally, the cards say nothing about cost: an incrementality study may be included for one vendor and priced separately for another. Read the card as a list of questions answered in public, and the hatched tiles as questions still to ask.

Why we built the cards

Our ranking answers one question: which platform lets a brand or agency launch and prove CTV results with the least friction. Proof carries 40% of the weight. The Proof Card makes that part visible at a glance, so a reader can see why a platform with large inventory or simple pricing may still rank lower on our weights.

How to use them

Before a pilot, look at the hatched tiles of every platform on your shortlist and turn each into a question in your brief. If the answer is a published page, send it to us through the contact form and we will update the card with the source and date.