How to Buy Disney Inventory Through Amazon DSP

What the DRAX integration actually connects

Disney's Real-Time Ad Exchange, DRAX, is Disney's own programmatic marketplace for its streaming and cable inventory. On 17 June 2025, Amazon Ads and Disney Advertising announced that Amazon DSP would connect to DRAX directly, making Amazon only the third demand-side platform — after Google and The Trade Desk — to get that access. The inventory on the other side of that connection spans three named properties: Disney+, ESPN and Hulu.

Rollout was staged. Disney+ inventory went live first outside the US, in France, Germany, Italy, Portugal, Spain, Switzerland, Turkey and the UK. Full access for all US Amazon DSP advertisers followed in the third quarter of 2025. If you set up a Disney line item and see thin delivery, confirm your market against the current rollout before assuming the targeting is wrong.

Magic Words and Disney Select — what they add

Two Disney-side capabilities sit on top of the raw inventory access. Magic Words is Disney's contextual targeting layer — it reads the content itself, not just the audience watching it, so a line item can target scenes or themes rather than only demographics. Disney Select is Disney's own first-party audience data, described at launch as coming rather than immediately live; check its current availability in your DSP console rather than assuming it shipped on day one.

The two combine with Amazon's own signals through Amazon Publisher Cloud, a clean-room environment that lets Disney's content and audience data sit alongside Amazon's shopping and browsing signals without either side handing over raw customer data to the other. The example Amazon uses is a pet food brand targeting Disney viewers who have separately shown interest in pet products on Amazon — a combination neither company's data could produce alone.

Setting up a first Disney line item

Inside Amazon DSP, Disney's DRAX inventory shows up as a supply source you can target directly rather than as a separate account you have to open. The practical setup steps are the same as any other DSP video line item: pick your objective, set your audience (Amazon's own segments, a Magic Words contextual segment, or both stacked), choose Disney+, ESPN or Hulu specifically if you want to isolate one from the others, and set a frequency cap that accounts for the fact that the same household may also see your Prime Video or Netflix impressions.

Amazon's streaming ad product recommends a $10,000 minimum for a self-serve DSP campaign and requires $50,000 for managed-service campaigns — Disney inventory competes for budget inside those same thresholds, it does not have its own separate minimum.

A worked example

Suppose you allocate $12,000 of a monthly budget to Disney's DRAX inventory specifically, split roughly evenly across Disney+ and Hulu. Premium ad-supported streaming inventory is reported in the industry at a roughly $20–$50 CPM range; at a representative $32 CPM blended across the two, that buys close to 375,000 impressions before frequency capping trims it to unique households. Compare that against your account's overall streaming performance rather than judging Disney inventory against a search-ads benchmark — across 27 advertisers in our own Amazon DSP book over a 31-day window this summer, all connected-TV and streaming inventory combined returned 0.77x on a straight last-click read, with the highest new-to-brand rate of any inventory type on the report at 56.3%. Disney inventory is playing the same top-of-funnel role, and should be judged against the same kind of target, not against a retargeting line item's ROAS.

What to do when it goes wrong

If Disney line items are not spending, check the market rollout first, then bid floor — a newer supply source is easy to under-bid relative to inventory your account has run for years. If it is spending and the ROAS looks weak, resist the instinct to cut it immediately: this is prospecting-shaped inventory, and a return target closer to 1.0x with completion rate and branded search as the supporting evidence is the honest way to read it, not the same bar you'd apply to a retargeting campaign. If Magic Words segments are producing almost no delivery, the contextual filter may simply be too narrow for the available inventory — broaden it before assuming the targeting concept has failed.

The common mistake

The mistake we see most is treating "Disney inventory" as one thing. Disney+ skews toward family and franchise content; ESPN is live sports; Hulu runs a broader general-entertainment slate that has historically carried more ad-supported viewing than Disney+. Buying all three as a single blended line item makes it impossible to tell which one is actually working for your product, and it's the same mistake as blending Prime Video with Fire TV Channels — the average hides the answer.

A second, quieter mistake is building creative once and assuming it will perform the same across all three properties. A pet brand's Magic Words contextual match against a nature documentary on Disney+ is a different viewing context than the same creative running mid-game on ESPN, and testing one against the other before committing a full month's budget is worth the extra week it takes. reMKTR runs 109 live Amazon DSP advertiser seats and splits Disney's three properties into separate line items from the start specifically so each one can be judged on its own delivery and its own contribution to branded search, not folded into an average that flatters the worst performer.