Disney’s Jamie Power explains why advertisers are buying audiences, not shows

Full analysis

Disney SVP of Addressable Sales Jamie Power says the quiet part out loud: on Disney's streaming inventory, advertisers have almost stopped buying shows and now buy audiences — 95% of streaming campaigns arrive with an audience segment attached. The consolidation of Disney+, Hulu, and ESPN onto one proprietary ad server is what makes this possible, and it's turning a $5 billion streaming ad business into something that behaves more like programmatic plumbing than a traditional TV upfront.

What's actually being decided: Not a decision for Disney — that's already made. The live question for everyone else is whether the big streamers become open data exporters into the programmatic pipes (as Disney's Compass integration with The Trade Desk suggests) or stay walled gardens. That shapes where identity, measurement, and yield value accrues over the next three years.

Reversibility: Type 1 for Disney's own stack investment (years of migration, hard to undo). Type 2 for how the rest of the ecosystem responds — partnerships and integrations can be added or dropped quickly.

Forcing function: The 2026–2027 upfront cycles, where "biddable" commitments Power calls the fastest-growing transaction method either keep eating fixed deals or plateau.

The Market Analyst — The headline number is the tell: "way more than half" of a $5 billion streaming ad business now goes through programmatic pipes. That's a premium content owner voluntarily routing its most valuable inventory through auction dynamics — something the old TV world resisted for a decade. For a generalist: Disney is treating its shows less like scarce reserved seats and more like a live stock exchange for eyeballs. Winners are the demand-side platforms (the software agencies use to buy ads) plugged into that pipe — The Trade Desk being named is not accidental. The risk for Disney is that programmatic transparency compresses the pricing power that premium content historically commanded.

The Skeptic — This episode is a Disney sales pitch, and the load-bearing assumption is that "audience-based buying" is buyer-driven demand rather than seller-driven margin engineering. Selling audiences instead of shows lets Disney monetize low-rated inventory by attaching a valuable segment to it — the advertiser pays for the "auto-intenders," not the mediocre program they run against. The 95% figure could mean buyers demanded it, or that Disney's tooling makes show-level buying inconvenient. And "biddable upfronts" — negotiate a floor, spend flexibly later — sounds like Disney converting firm commitments into soft ones. Whether that's good for buyers depends entirely on who holds pricing leverage when the floor gets set.

The Operator — The real-time bid-stream signal during the NBA Finals spike is the most operationally interesting thing here, and the hardest to actually use. Disney flags a viewership surge to programmatic buyers mid-event; a trading desk then has minutes to react. Who's watching that signal at 9pm on a Sunday? Most buy-side optimization runs on dayparts and daily pacing, not live-event triggers. In plain terms: Disney built a fire alarm, but most buyers don't have anyone in the building to answer it. The SMB self-serve on-ramp is the opposite problem — it works, but it floods measurement and brand-safety vendors with thousands of tiny, messy campaigns.

The Customer / End User (the agency buyer) — For a trading desk, Compass is genuinely useful: Disney sending first-party data out for deduplicated reach — counting a viewer once across Disney+, Hulu, and ESPN instead of three times — solves a real measurement headache. But the "holistic buy beyond Disney's inventory" framing glosses over the friction. Matching Disney's identity to my other data still runs through clean rooms and identity resolution that break at the edges. And biddable's flexibility cuts both ways: I get to chase cultural moments, but I've also given up guaranteed price certainty that made upfront planning safe. For a generalist: I'm trading a fixed-price dinner reservation for a table I can grab anytime — great until the restaurant's full on the night I need it.

The CFO — Follow the margin. Disney spent years and real capital migrating ESPN off a third-party ad server onto its own stack. That's not a feature — it's Disney refusing to pay a toll to outside ad-tech vendors and capturing that spread itself. For any publisher CFO watching, the lesson is blunt: owning the ad server is now a P&L line, not an IT decision. The counter-cost is that programmatic-heavy revenue is lower-margin per impression than negotiated direct deals. Disney is betting volume and flexibility beat unit price. That math only works at Disney's scale — a mid-size publisher copying this playbook eats the migration cost without the demand to fill the pipe.

Where they disagree:

1. Is audience-buying buyer demand or seller leverage? The Market Analyst reads 95% as the market speaking. The Skeptic reads it as Disney's tooling steering buyers toward what monetizes Disney's weak inventory best. Both can't be fully right — and the answer determines whether other publishers should rush to copy it.

2. Does biddable help or hurt buyers? The Customer likes the flexibility to chase moments; the Skeptic sees firm commitments quietly converted into soft ones, with Disney setting the floors. The leverage question — who needs whom more when the floor gets negotiated — is unresolved.

3. Is the real-time signal a breakthrough or a demo? The Market Analyst sees dynamic yield management arriving in CTV. The Operator sees a capability most buyers structurally can't act on yet.

What this hinges on: Whether premium streamers exporting first-party data into the programmatic ecosystem is a durable strategy or a phase. Disney is betting that being a data exporter — feeding DSPs and agencies for cross-platform measurement — wins more total spend than hoarding data behind a wall. If that bet pays, the walled-garden model erodes and independent identity/measurement vendors get a seat at the premium-CTV table. If premium pricing craters under programmatic transparency, Disney pulls back toward the wall and the door closes.

Which way the council leans: Toward this being real and structural, not vaporware — the ESPN migration and the programmatic share of a $5B business are committed, sunk investments, not roadmap slides. But the Skeptic's caution on pricing leverage is the thing to watch, not the tech.

What to verify: For agencies — model whether biddable floors actually beat your negotiated upfront rates over a full year, not in the moment. For DSPs/SSPs not named — treat the Trade Desk integration as a signal to get in line for Compass-style data access before it becomes table stakes. For measurement vendors — build the capability to verify real-time bid-stream signals, because if Disney's model spreads to other sports-rights holders, unverified "attention spikes" become a manipulation vector.

Prediction: At least one other major streaming/sports rights holder (Netflix, NBCUniversal/Peacock, Paramount, or WBD) will publicly announce a first-party-data-to-DSP data-collaboration integration comparable to Disney's Compass — naming a specific DSP partner — by the close of the 2026 fall upfront/NewFronts cycle.

Confidence: Medium — Peers face identical pressure to prove deduplicated cross-platform reach to agencies.

Why: Disney has now publicly demonstrated that exporting first-party data to DSPs like The Trade Desk wins programmatic share rather than cannibalizing direct sales — and it's disclosing scale ($5B, majority programmatic) specifically to recruit buy-side partners. Every rival streamer is chasing the same agency budgets and the same measurement demand for counting a viewer once across platforms; once one large player proves data-collaboration attracts spend, the others follow to avoid looking like the closed option on the media plan. The opposite outcome — peers staying fully walled — is less likely because agencies are actively rewarding interoperability, and no major streamer can afford to be the one buyers can't measure against everyone else.

Revisit by 2026-11-30: We're right if a second major streamer announces a named DSP data-collaboration/clean-room integration for cross-platform reach by the end of the fall upfront cycle. We're wrong if no comparable named integration is announced and peers publicly reaffirm keeping first-party data inside their own walls.