Pinterest Just Made 600M Pinners Targetable on CTV Inventory It Doesn't Own

Pinterest launched tvScientific by Pinterest on April 27, 2026, opening its 600 million monthly active users to third-party connected TV inventory through what Pinterest calls "all the major SSPs." It is the first time Pinterest audiences are retargetable off-platform. Early tests showed a 27% outcome lift per $100 spent and a 65% lift in purchases when the targeting layered onto tvScientific's CTV buys.

The signal travels. The screens belong to someone else.

This is the part I want performance buyers to notice, because it changes how to think about Pinterest's role in the CTV stack. Pinterest is not selling Pinterest CTV inventory, because Pinterest does not own any. What it bought when it picked up tvScientific late last year was the plumbing that makes a planning-intent audience portable across third-party CTV exchanges. The audience layer rides on top of inventory the buyer was probably already touching through Trade Desk or DV360.

The targeting works two ways. Deterministic via user IDs, probabilistic via modeling. Pinterest told Digiday the deterministic side covers cases where the same user logged into Pinterest also logs into a CTV app, and the probabilistic side fills in the gap with modeled lookalikes. That second bucket is doing most of the work, in my read, because deterministic match rates between social and CTV are usually rough.

So the product is essentially this: rent the Pinterest signal, point it at third-party inventory, and pay tvScientific on outcome. VideoWeek confirmed the launch is open to advertisers of any size, which matters because traditional CTV platforms tend to start with brand budgets and trickle down years later.

Planning intent is a strange signal to put on a TV screen

Pinterest's pitch is that its data tells you what someone is about to do, not what they already did. Lee Brown, Pinterest's chief business officer, framed it as discovery to buying, both on and off Pinterest. The way Jason Fairchild, the tvScientific CEO who now runs the unit inside Pinterest, said it: "Pinterest's signals are unique because they reflect what people are planning, not just what they've already done."

That distinction sounds like marketing copy, but it is meaningful for CTV in particular. Most retargeting signals plugged into streaming inventory come from purchase histories, abandoned carts, or search clicks. Those are intent signals about something the user has already started. Pinterest's strongest segments are saved pins, idea boards, and queries that index months ahead of a purchase. Wedding planning. Kitchen remodels. Holiday decor. The things people gather inspiration for before they ever land on a product page.

Whether that translates to CTV outcomes is the open question. The 27% outcome lift per $100 in spend that Pinterest cited is suggestive, not proven. It is also a vendor-published number, which means there is a decent chance it cherry-picks the categories where Pinterest already wins. I would treat it as directional and run a clean test before reallocating budget.

The LG case study is specific, which is rare

The most useful number from the launch is buried in AdExchanger's coverage: in early tests with LG, advertisers saw a 73% increase in unique household reach and a 24% lift in net new customers. Both metrics are unusual for CTV vendor announcements because they actually mean something to a buyer.

Reach lift is the most honest CTV metric. If you add a new audience source on top of your existing buy and reach climbs by 73%, you are genuinely finding households your other DSPs missed. Net new customer lift is even harder to fudge, because it requires defining "existing customer" and proving the audience surfaced people outside that list.

What I would flag is sample size. MediaPost noted the LG figures came from early test campaigns, and Fairchild called the rollout "just a start." Translation: this was probably a small set of advertisers testing in a controlled way. Wait for second-cohort numbers before assuming the lift holds at scale. The history of CTV vendor launches is paved with first-test results that decayed by the third quarter.

Where this sits next to Meta's CTV expansion

The timing is the part that interests me. Meta announced its own CTV expansion through Magnite and FreeWheel earlier this year, and the supply geometry is different. Meta is plugging directly into premium CTV inventory through SSPs to push its own ad units. We covered the implications in Meta's CTV plan and what it does to Q3 CPMs.

Pinterest's play is the inverse. It is not pushing its own ad format onto CTV screens. It is renting out the audience signal and letting tvScientific buy the inventory programmatically. Less inventory pressure on CPMs, more signal density on the targeting layer. The two strategies will probably not collide for a while, because they pitch to different buyer profiles. Meta's CTV expansion targets brand budgets currently moving from linear. Pinterest's tvScientific play targets performance buyers running e-commerce funnels who never put a dollar into TV.

The opening for shopping-intent advertisers in Q3

If you are running CTV for a category Pinterest indexes well (home, beauty, apparel, food, weddings, DIY), there is a real arbitrage window in the next 8 to 12 weeks. Vendor-published lift numbers usually overstate reality, but the LG reach figure is the kind of number that holds even when you cut it in half. A 35% reach lift on a clean A/B is still worth running a $20K test against.

Three things to ask Pinterest's team before signing on:

• What is the deterministic match rate they are seeing in your category? If it is under 15%, the audience is mostly modeled, and the lift will look more like a contextual buy than a true addressable layer.
• Which SSPs they are routing through, and whether your incumbent DSP can already buy the same inventory directly. Pinterest said "all major SSPs" but did not name them, which is the kind of phrasing that usually means Magnite, FreeWheel, and Roku.
• How outcome attribution actually works. tvScientific's pitch is pay-by-outcome, but the outcome model leans on its own measurement framework. Make sure you can independently verify lifts via your MMM or an incrementality test, not just the tvScientific dashboard.

Honestly, the bigger story is what Pinterest does next. Chip Jessopp, Pinterest's VP, called this "the first step" in their off-platform roadmap. If Pinterest's planning-intent signal works on CTV, the same data layer ends up on display, retail media, and probably YouTube before next Christmas. CTV is the proof point. The actual ambition is making Pinterest a portable audience graph, the same way Meta tried to do it through CAPI.

The thing I keep coming back to is that nobody really knows yet whether planning intent maps to a TV screen the way retargeting does. The signal is real. Whether it survives the trip from a Pinterest board to a 65-inch LG OLED is the bet you are placing if you run this in Q3. I would rather find out with a measured test than read about it in someone else's case study six months from now.

Notice Me Senpai Editorial