PayPal Ads Pitches Transaction Data as Cookie-Free Audience Signal

Full analysis

PayPal Ads stood up at Advertising Week in New York and told advertisers that its 293 million U.S. users, plus 100 million-plus Venmo users, plus browsing from the Honey extension, add up to a cookie-free read on who is about to buy. Asaf Davidov, PayPal's head of global ad insights, pitched actual spending as a stronger signal than someone reading an article about running shoes. The real decision for a buyer: do you treat PayPal as a serious new data source for Q4 and 2027 planning, or as another card network promising to turn transactions into media and never delivering?

This is easy to undo. Nobody is signing a multi-year exclusive off a keynote. A test budget is a test budget. The deadline, such as it is, is the Q4 commitment cycle and 2027 upfront planning, when buyers decide which alternative signals get a line item.

The Skeptic. Three things all have to be true at once, and none is confirmed. PayPal's match rate to inventory buyers can actually address has to be high. The attribution window has to line up with how campaigns are judged. And the Honey data has to survive regulatory heat, which is not a given with a class-action cloud over affiliate manipulation already hanging on it. Venmo "social payments" is the weakest leg: splitting a dinner tab is not purchase intent, it is a Tuesday. And PayPal's ad product is still subscale in the programmatic pipes where budgets actually move. Davidov is selling a signal buyers can't yet buy where they spend. For the non-specialist: great data you can't plug into your buying tools is a slide, not a product.

The Strategist. This is the pivot Visa and Mastercard circled for years and never ran. Purchase-confirmed intent at scale is genuinely better than guessing from browsing, and no demand-side platform (the software buyers use to purchase ads) can rebuild it from scratch. If PayPal becomes the identity anchor for commerce campaigns, the middle-layer identity vendors, LiveRamp and ID5, lose a reason to exist at the edges. The move that proves PayPal is building a real stack rather than licensing a data feed: a clean-room tie-up with Databricks and a connection into Roku's TV inventory. Those two together mean full-funnel. Absent them, this is a data-licensing line with a big user number attached.

The Operator. The pitch dies or lives on two unglamorous things. First, latency. If a transaction takes 24 to 48 hours to show up in the signal, it is close to useless for mid-funnel retargeting, which is where a lot of the money is. Second, permissioning. After the FTC scrutiny on Honey, one consent gap poisons the whole stack, and your trafficking team owns that risk, not PayPal's keynote. Before anyone commits Q4 dollars, run PayPal's match rates against your own clean-room output and see if they hold. For the non-specialist: a buyer has to confirm PayPal's audiences actually overlap with the people they can reach, and that the data is clean enough to touch.

The Market Analyst. On PayPal's P&L, ad revenue is a rounding error, and the stock is not priced as a media company. The bull case needs PayPal to pull measurable retail-media budget away from Amazon Ads and Walmart Connect, both of which already have checkout data AND their own inventory to sell against it. PayPal has the data and no inventory. That is the gap. The signal that consensus is underrating this story would be a top-five holding company naming PayPal a preferred partner before year-end. Short of that, analysts are right to file this under payments optionality and move on. For the non-specialist: PayPal is trying to sell the ingredient when the winners sell the finished meal.

Where they split

The real disagreement is whether data without inventory is a business. The Strategist says purchase-confirmed intent is so much better than browsing guesses that buyers will route to it regardless of where PayPal sits in the pipes. The Market Analyst and the Skeptic say the opposite: Amazon and Walmart win precisely because they own the shelf and the ad slot, and PayPal owns neither. A data feed that has to ride someone else's rails is a commodity, and commodities get repriced to the floor.

The second split is on the data itself. Davidov's line, "there's a person behind them, the identity match is incredibly strong," is true and beside the point. A strong identity match on a noisy signal (Venmo check-splitting) is still a noisy signal. The Operator's latency problem compounds it.

What it hinges on

Two things. Can PayPal's transaction signal be bought inside the tools agencies already use, fast enough to matter? And does it survive the Honey regulatory question clean? If both land, the Strategist's moat is real and LiveRamp and ID5 should be nervous. If either wobbles, this is Visa and Mastercard again: a beautiful data asset that never became media. The council leans skeptical. Distribution and inventory are the hard problems, and PayPal showed neither at the keynote.

What to check before spending: match rate against your own clean room, propagation latency, and the Honey consent chain.

Prediction: PayPal Ads will not be named a preferred or primary data partner by any of the five largest advertising holding companies (Omnicom, WPP, Publicis, Dentsu, Interpublic) before the end of Q1 2027 earnings season in February 2027.

Confidence: Medium. Data without inventory rarely wins a holdco mandate fast.

Why: PayPal is pitching a data signal but has no owned ad inventory and remains subscale in the programmatic pipes where holdco budgets move, so a buyer who likes the data still has nowhere native to spend against it. Amazon Ads and Walmart Connect won retail-media mandates because they pair checkout data with their own ad slots; PayPal has the first half and not the second. A holdco preferred-partner deal requires provable match rates, a buying path inside existing tools, and clean permissioning on Honey data that is under a class-action cloud, and none of those was demonstrated at Advertising Week. The opposite outcome, a fast holdco mandate, would require a holding company to bet its clients' Q4 and 2027 budgets on a signal it cannot yet buy where it spends, which is not how these commitments get made.

Revisit by 2027-03-01: We're right if no top-five holding company has publicly named PayPal Ads a preferred or primary data partner by the close of Q1 2027 earnings season. We're wrong if any of Omnicom, WPP, Publicis, Dentsu, or Interpublic announces such a partnership before then.

A holdco signing PayPal would tell you the transaction-data pitch finally cleared the distribution problem that sank the card networks. Silence tells you it didn't.