OpenX says a connected TV campaign that applied attention data before the bid cost 47% less per acquisition and delivered 76% more qualified traffic than a standard CTV deal run through the same DSP. The result comes from one head-to-head test for Circa Resort & Casino, run by the performance agency Causal and announced on November 19, 2026.

How the test was built

According to OpenX’s release, Causal hypothesised that reaching viewers in high-attention moments would produce more bookings and conversions for Circa than standard CTV. OpenX built a deal for Causal through OpenXSelect that applied attention data before any impression was made available for bidding, limiting the deal to environments where viewers were more likely to be watching the screen. Causal ran that deal against a standard CTV deal through the same DSP, with all other campaign settings held constant and attention as the only variable between the two.

The release reports two outcomes: a 47% improvement in cost per acquisition and 76% more qualified traffic. Jennifer Laing, SVP of operations at Causal, says in the release: “Applying attention on the supply side with OpenX meant our budget only went to environments where people were actually watching.” Erika Loberg, VP of global CTV and enterprise solutions at OpenX, adds: “Causal changed one thing, applying attention before the bid, and the cost of each acquisition fell by nearly half.”

Advertisement

MarTech Your brand belongs here. Reach the decision-makers who read MarTech every day. Premium placements across the site and newsletter. Advertise with us

What the release does not say

The release publishes percentages and no absolute figures. It gives no spend, no impression count, no number of conversions, no flight dates and no definition of “qualified traffic.” It names one advertiser in one category, a casino resort, and one agency. The comparison is between two deals, and the release does not say how many conversions sit behind the 47%. OpenX is the seller of the attention-targeted inventory and the author of the release, and the figure has not been checked by an outside party in any source this desk fetched.

The result comes from a single test, and a reader should treat it as a hypothesis worth repeating on their own campaigns.

Where the filter sits

The more durable fact in the release is where the targeting happens. OpenX says supply-side targeting grew from 9% of its CTV spend in 2022 to about half in 2025, faster than any other channel. It describes attention targeting as one of a set of supply-side capabilities, alongside co-viewing targeting, which identifies inventory most likely to reach more than one viewer at once, and content-aware targeting, which adds genre, rating and language context. Buyers apply attention and co-viewing targeting to any CTV deal through OpenXSelect.

Content signals are moving to the sell side as well. On the same day, Nexxen announced that Gracenote’s program-level data is available in its SSP, a move we covered in our brief on the Nexxen partnership. Index Exchange announced a Gracenote integration in September 2025, and Gracenote’s August 2026 integration with The Trade Desk brings show-level data into a DSP. In that release Bill Michels, chief product officer at Gracenote, says content intelligence “deserves a more prominent role in establishing targeting criteria and driving performance.”

Gracenote’s own survey, cited in the same release, asked 500 media planners, TV buyers and programmatic traders at U.S. agencies and brands about show-level data. Of the programmatic traders, 91% said the lack of show-level data limits CTV advertising, and 95% of the media planners said show-level reporting would be valuable.

Newsletter

Get the week's best tech coverage.

Free. Read by thousands of HR, tech, and business leaders.

What it means for the marketing leader

A CTV buy has traditionally been planned on audience and reach, with the viewing context judged after delivery. Two signals are now being sold before the auction: what program the ad runs in, and whether the viewer was likely to be paying attention. Each one changes which impressions a deal includes, so each one changes the price per thousand and the count of available impressions. A lower CPA from a filtered deal can come with a smaller pool of inventory, and the release does not report how much inventory the attention filter removed.

The same pattern appears in other recent coverage on this site. LinkedIn’s UK launch of CTV ads with LG and Titan OS added another supply route, and our look at four ad tech Q3 reports found connected TV as the shared growth line. More supply routes and more filters on each route make the CTV plan a set of choices a buyer has to specify.

Questions to ask before copying the test

A buyer who wants to repeat the experiment can reuse its structure and request more detail than the release gives. Ask the SSP for the attention scoring method and who supplies the underlying measurement. Ask what share of the standard deal’s impressions the attention filter would have removed. Ask for CPA in absolute terms with the number of conversions behind it, and for the definition of the qualified-traffic metric. Run the two deals through the same DSP at the same time, as Causal did, and agree on the success threshold before launch so the result is graded against a rule set in advance.

In our reading, the useful part of this announcement is the test design, which isolates one variable and holds the rest constant. The 47% is a single data point from the vendor that sold the inventory. Treat it as the reason to run your own version, and as the thing to compare your own result against.

Source: OpenX press release