Left Arrow Icon
All articles

The CTV Overcorrection: Brands Got Measurement and Forgot About Demand, Says tvScientific CCO

Ad World News Desk
Published
August 26, 2026

Heather Carver on why CTV can finally build demand and convert it, and what goes wrong when advertisers only fund one half of the funnel.

Credit: Ad World News

Make Ad World News one of your go-to sources on Google

Google Icon
Add Ad World News on Google
Quote Icon
Marketers no longer need to choose whether CTV is brand or performance. CTV can do both. It's no longer about just creating emotional connection and brand awareness and reach metrics. CTV is now measurable.

Heather Carver

Chief Customer Officer

Heather Carver

Chief Customer Officer
tvScientific by Pinterest

CTV spent years as an awareness channel with reach metrics and no receipts. Now it carries pixels, attribution windows, and holdout testing, meaning a marketer can optimize a streaming campaign the way they'd optimize paid search. That makes CTV a full-funnel channel for the first time, capable of building demand at the top and converting it at the bottom. Most advertisers haven't caught up to that. Some swung budgets entirely into lower-funnel performance the moment measurement arrived, stopped feeding the top of the funnel, and watched customer acquisition costs climb a quarter later. The channel can finally do both jobs. The question is why so few brands are letting it.

Heather Carver is Chief Customer Officer at tvScientific by Pinterest, the performance TV platform Pinterest acquired and folded into its advertising products. Carver leads customer and partner strategy and oversees the post-sale side of the business, including account management, onboarding, performance, and ad operations. Seventeen years across digital advertising and media gave her working knowledge of both the buy side and the sell side. She spends much of her time on how advertisers tie a CTV budget to business outcomes.

"Marketers no longer need to choose whether CTV is brand or performance. CTV can do both. It's no longer about just creating emotional connection and brand awareness and reach metrics. CTV is now measurable. You can optimize it. You can track back to app installs and purchases," says Carver. Impressions still have a place in her reporting as a measure of reach. She pairs them with a second question about what the exposure produced after the ad ran.

Reporting to the CFO

For Carver, the reporting line is the place where impression counts run out. A marketer presenting CTV performance to finance arrives with delivery figures and a question about what those figures bought. Search, social, and CTV each answer to their own standard, and those separate standards roll up into a single number the business cares about. "When you have to report into the CFO or the CMO, I don't think they care about delivering X millions of impressions across CTV, the biggest screen in the home," Carver says. "Okay, that's great, but how does that tie back to the actual outcomes?"

Carver runs a different test on a TV campaign, asking instead what the campaign prevented. Holdout groups and geo tests supply a baseline of comparable households that never saw the ad. The gap between exposed and unexposed households gives a number the delivery report cannot produce on its own. "You don't look at what happened because you ran the TV campaign," Carver notes. "You look at what wouldn't have happened if you didn't run TV. Would your branded searches not be as high, would your social engagement be lower?"

A team working to a one-day attribution window optimizes against the conversions visible inside that window, and the early reporting looks healthy. A one-day view can't show what the same spending is doing to demand a quarter later. Carver describes the moment a marketer finds out. "I got so obsessed with a metric for lower funnel that three months later, all of a sudden I look and I'm like, 'Oh, my customer acquisition costs are going up and I'm really reliant on branded search,'" Carver explains.

Before the brand is chosen

Pinterest completed its acquisition of tvScientific in February 2026 and began routing its audience signals into CTV delivery. Carver describes the thesis behind the deal as a supply problem in reverse. Advertisers competing for people who have already picked a brand pay more for each of them, and the pool doesn't grow. Buying against people who haven't picked anything yet puts the spend earlier in the sequence. "It gets crowded later on in the year and people are chasing the same users," Carver says. "Instead of reaching someone when they've already made a decision on what they want to buy, it's really helping to create that demand, then seed that demand and see it through the full funnel."

Pinterest supplies search behavior from a stage before brand preference forms. Users run more than 80 billion searches a month on the platform, a volume that behaves closer to a search engine than a social feed. Carver works from what those queries look like at the beginning of planning. "96% of the text searches that people do on Pinterest are unbranded," Carver notes. "People haven't even formed what they're looking for. It could just be summer vacations or wedding stuff. Their intent is there but they haven't really made a choice yet."

Frequency management follows the same logic. Carver's team watches how many exposures a household absorbs before the next one earns nothing back. Early results published after the integration went live showed a 27% increase in outcomes for every $100 spent when the platform's optimization ran on Pinterest intent data, with purchases up 65%. "Maybe we've tapped you enough and you're going to buy it and it's fine," Carver adds. "Through the different data that we have, we realize that might not be the best use of the spend. Maybe we need to hit new consumers."

Diagnosing the creative

Media mix and audience selection absorb most of the optimization attention, and Carver puts creative alongside them as a variable teams rarely test. tvScientific recently launched a tool called Creative Advisor that scores a video ad on elements including messaging, audio, logo visibility, and call to action, then returns recommendations before the spot runs. Carver applies the same diagnostic sequence when a campaign underdelivers against a well-built audience. "Maybe it's not your medium. Maybe it's your messaging in the creative itself," Carver says. "Instead of optimizing on a per user, per household basis, you can say, maybe your creative is not great."

Getting creative into the optimization loop runs into how agencies and brand teams are organized. Creative departments build the assets and hand them off, and the performance data that comes back lands with a different group. Carver heard the arrangement described firsthand at a measurement conference this summer. "I sat with a bunch of marketers and a couple of people from an agency told me, 'We're on the creative team and we sort of are siloed from everything else. So we create the creatives and then we just ship it,'" Carver says.

Platform-specific creative is the other half of the same problem. Carver points to advertisers repurposing a single asset across environments built on different viewing behavior, including radio spots pushed into audio streaming and social cuts dropped into CTV. Some of those transfers work, and Carver treats the difference as something to test before the buy. "Don't just copy and paste your strategy for Google and YouTube into other platforms," she concludes. "Every platform is unique in some way. It goes back to that partnership element of making sure that you are developing the right creative for the right person, the right environment, and the right message."