How should you read CTV platform claims about reach, ROAS and customers?

Vendor figures are statements, not measurements

This note uses figures from vendor pages read, including those of Vibe. The same rules apply to every vendor, and the goal is to read figures carefully, not to dispute them.

Every number on a vendor's page was chosen and calculated by the vendor. That does not make it wrong, but it means the base, the period and the method are the vendor's. On this site we quote such figures as "the vendor states" and never as our finding. Buyers can apply the same discipline.

Household reach

Reach claims look comparable and usually are not. On the pages we read:

• MNTN described "150+ premium networks reaching 99% of U.S. households".
• tvScientific described reaching "93% of U.S. households through direct and programmatic buying".
• Universal Ads stated "Reach up to 90% of US households".

These percentages may use different bases: all TV households, streaming households or addressable households. They describe potential reach across all inventory, not what a given budget will deliver. Ask: what is the base, and what reach did campaigns of my budget size achieve?

Customer and advertiser counts

Counts depend on definitions. Walmart's release described Vibe as serving "more than 10,000 advertisers", MNTN's enterprise page cited "2,082+ brand and agency customers" from its internal data for 2019 to 2025, and StackAdapt's plans page stated "Trusted by +40,000 Brands". An advertiser, a customer and a brand can mean different things, and periods differ. Where a vendor's own pages and a third-party release give different figures for the same company, we use the third-party primary source or attribute each figure.

ROAS and performance figures

Average ROAS figures on a vendor page are usually attributed ROAS across selected campaigns, over a window the vendor chose. They are not incremental and they are not your result. Customer stories that cite incremental ROAS from a named measurement partner are more informative, because the method and the measurer are stated, but they still describe one advertiser.

Speed and scale statistics

Some vendors publish platform statistics: The Trade Desk states its platform "analyzes up to 15 million ad opportunities each second". Such figures describe system capacity. They tell you little about how your campaign will perform.

Superlatives and labels

Vendor pages often carry "#1" labels or category awards. Treat them as marketing unless the page names the source, the category and the date, and you can check them on the source's own page. On this site we do not repeat superlatives about any vendor.

A short checklist

1. Who measured it, the vendor or a named third party?
2. What is the base (households, advertisers, campaigns)?
3. What period, and is it current?
4. Is the figure attributed or incremental?
5. Can I see the same metric for campaigns like mine?

Applying the checklist in a pilot

In practice the checklist turns into a few lines in a pilot brief. Ask each platform for reach delivered by campaigns with a similar budget and target, not total reach. Ask for the attribution window behind any ROAS figure in a sales deck. Ask whether a customer story's result was measured by the platform or by a named third party, and whether you can speak to that customer. And ask for your own pilot's results in a format you can check against your analytics. Vendors that publish their methods usually answer these questions quickly, and their answers are worth keeping with the pilot results.

How this applies to our scores

Our scores reward what a vendor's pages describe, such as methods, terms and partners, rather than the size of its claims. A large reach figure counts toward published inventory reach; it does not raise a measurement score.