Intent Share: Who Creates Demand, and Who Keeps the Customer

Say you want to watch The Pitt. You open Prime Video almost by muscle memory, type the title, and find it available through HBO Max. Amazon already has your card. Two clicks and you’re watching.¹

Who made you want to watch? HBO did. You came for its show. You didn’t browse or accept a recommendation. You arrived with a title in mind and wouldn’t have settled for a different medical drama. You had intent.

Who holds the customer relationship? Amazon handles the bill, the cancellation, and the next offer. HBO created the reason to arrive. Amazon gets paid and keeps the storefront the viewer returns to.

Streaming increasingly separates the company that creates demand from the company that keeps the customer. A content owner’s leverage depends on how badly people want its titles and who controls the relationship with those viewers.

The company that creates demand does not necessarily capture all the value created by that demand.

Measuring want

Think about the last thing you watched. Did you open the app for it, or did you open Netflix with no plan and take what it handed you?

Both sessions add to Nielsen’s viewing totals.² They happened for different reasons. In the first, the show brought you to the app. In the second, the platform chose for you.

I propose calling the distinction intent share: the fraction of a title’s viewing that begins with someone seeking that title rather than accepting it from a discovery system.³

Intent share measures how viewing begins, not how much leverage a title has. A small show can have a devoted audience; a bigger show with lower intent share may bring in far more subscribers. And wanting a title is different from wanting it enough to change services.

Intent matters because it points toward a harder question: what would the viewer accept instead? If I insist on your content, the distributor needs you. If I merely want something to watch, it has alternatives. No platform can substitute a sitcom for the Super Bowl. A library movie may face dozens of acceptable substitutes. Better recommendations help the platform find them.

A typed search, a watchlist, or a deep link can signal intent, though none proves it alone. The mix changes over time. A premiere can draw people who came looking for the title; later viewing may depend more on placement. Discovery can also create demand. Someone accepts a recommendation, loves the show, and tells a friend who then searches for it by name.

An intent curve would track that changing mix. Read alongside audience size, retention, and willingness to switch, it could tell a studio more than a viewing total alone.

What happened when HBO left

Back to Amazon. Under Prime Video Channels, Amazon bills the customer and remits a share to Warner Bros. Discovery on undisclosed terms. Reporting puts Amazon’s cut near 30%, though terms vary.⁴ HBO is not blind: partners receive subscription and viewing reports, with more detailed access depending on the agreement.⁵ Amazon’s advantage is its view across the storefront and its ability to act on it. It can offer a Pitt viewer another service or a rental at the moment they are deciding what to watch next.

That can be worth paying for. At an Amazon event, Antenna presented estimates that HBO Max lost 5.1 million subscribers when it left Prime Video in 2021 and recorded 3 million signups in its first three months back. Only 16% of those signups came from people already subscribed through HBO Max directly or another distributor.⁶ Amazon was reaching beyond HBO’s existing subscriber base.

The departure exposed the other side of that reach. According to Antenna estimates reported by The New York Times, fewer than 500,000 of the departing subscribers had resubscribed to HBO Max eight months later.⁷ That is a large gap between losing a distribution channel and rebuilding the subscriber base elsewhere.

This was not a controlled experiment. Price, convenience, household access, and the programming available at the time could all have affected the outcome. The figures do not tell us that everyone else stopped watching HBO. They do suggest that a subscriber acquired through Amazon cannot simply be counted on to become a customer elsewhere when HBO leaves.

HBO can gain customers it would struggle to reach alone and still become dependent on the company that reaches them.

Originators and allocators

Two roles help explain the trade-off.

Originators create demand: a series people wait for, a league people schedule around, a franchise people will follow across a paywall.

Allocators decide where demand goes once it exists. They own the interface, the search box, the billing relationship, and the row of tiles that fills a screen when someone opens an app with no plan.

A company can do both. Netflix does; so does Amazon. The distinction is between what a company’s programming brings in and what its position in front of the customer lets it keep. Recent deals show how much companies want that position.

Fox agreed to acquire Roku in a deal valued at roughly $22 billion. It is buying a home screen, distribution technology, and direct viewer relationships alongside Roku’s content businesses.⁸ Those assets let Fox reach viewers without having to make every program they watch.

YouTube’s agreement with NBCUniversal will bring Peacock into a bundle with YouTube Premium in early 2027.⁹ Peacock gets access to YouTube’s audience. YouTube gets more reasons for that audience to pay and stay.

Netflix has integrated TF1 programming in France and reportedly discussed carrying Peacock and Fox One, though those talks were not announced deals.¹⁰ Even a company built around its own catalog sees a reason to become the place people watch someone else’s.

These are different transactions: an acquisition, a wholesale bundle, and a programming partnership. Their economics differ. What they share is an investment in being the viewer’s starting point.

Why distribution is easier to defend

Selling another service’s subscriptions can earn a platform a revenue share without financing that service’s production slate. Greenlighting a series puts production capital at risk before the audience is known.¹¹ Distribution still has costs and risks, especially when a deal includes guarantees. But it need not carry the same exposure to whether a new show works.

A subscription distribution deal gives an investment committee a contractual take rate, existing customers, and comparable deals to examine.¹² An original can have useful comparables too, but its audience is less certain. That makes the distribution case easier to defend, even when the original might create more long-term value.

For the studio, the appeal is just as concrete. A partner can bring subscribers, reduce the cost of reaching them, and make the service easier to find. A deal may be worth more than the studio could earn alone. The mistake would be to count those benefits without pricing the control it gives up.

Pressure to deliver predictable earnings also shapes a greenlight meeting. A sequel or spin-off offers a known audience and a familiar comparison. That makes the forecast easier to defend. It does not, by itself, tell the studio whether the next installment will make viewers insist on its content.

The cost of the defensible greenlight

A new franchise, when it works, gives viewers something they want specifically. Another competent entry in a crowded genre may give them one more thing they are willing to accept. The difference matters more than whether the project is original or a sequel.

A sequel that deepens attachment can strengthen the studio’s bargaining position. One that merely supplies another acceptable hour of viewing gives the platform another substitute. The useful greenlight question is what the audience would miss if this project did not exist.

A studio does not have to own the app to have leverage. The NFL has NFL+, but its media-rights business spans multiple distributors.¹³ Distributors still need the games, wherever viewers watch them.

The practical test is what happens when a title is unavailable: do viewers watch something else, cancel, or go looking elsewhere?¹⁴ A title can also matter by keeping subscribers who discovered it through recommendations. Intent share should inform that assessment, not replace it.

A library title that attracts little direct demand and does little to retain subscribers may be worth more in broad distribution. A franchise that brings people to a service and makes them follow it elsewhere deserves different terms.

Those terms depend on the deal. Licensing a title can give up exclusivity; selling a service through Amazon leaves Amazon handling the storefront and billing. Neither automatically destroys value.¹⁵ The studio should ask what it gains in reach, what it retains in customer access, and whether it can use this hit to sell the next one.

Pressure to improve current earnings makes the last question easier to overlook. A familiar greenlight can add interchangeable supply. A convenient distribution deal can put the next subscription decision in someone else’s hands. Together, they risk leaving the studio responsible for making the hit while the distributor builds the repeat business.

When that happens, studios are selling future leverage to hit present margin targets.¹⁶

Viewing curves help studios account for streaming content. Intent curves should help them manage it, alongside the evidence of who stays, pays, and follows.¹⁷

A studio can make the show people came for and still leave the next sale to Amazon. The distribution deal needs to pay for that, too.

Notes

1. HBO Max can be bought and managed through Prime Video. Amazon’s guide. ↩
2. Nielsen’s overview describes viewing measurement but does not distinguish deliberate title searches from recommendations. ↩
3. Intent share is proposed here, not an industry-standard metric. Search, watchlist, and discovery events are possible signals. ↩
4. Amazon’s reported cut is roughly 30%; terms vary. Next TV; van Es and Iordache’s study. ↩
5. Partners receive subscription, billing, and viewing reports, including encrypted customer IDs; access varies by agreement. Amazon’s documentation. ↩
6. Antenna estimates presented at Amazon’s 2025 event. Its 2021 analysis reported 4.2 million lost subscribers; estimates vary by measurement window. ↩
7. Antenna estimates reported by The New York Times (reprint) and Puck, not WBD disclosures. ↩
8. Fox’s announcement. The $22 billion figure is enterprise value; the acquisition was announced, not completed. ↩
9. NBCUniversal’s announcement: Peacock Premium joins a YouTube Premium bundle in early 2027. ↩
10. Netflix/TF1 launch; reported Peacock/Fox One talks. The latter were discussions, not announced deals. ↩
11. A slate is a planned lineup of films and series. Production funding differs from subscription resale; guarantees can still put a distributor’s capital at risk. ↩
12. Take rate: the distributor’s revenue cut. Comparable: a similar deal used as a benchmark. ↩
13. The NFL has NFL+ but licenses games to competing distributors under long-term rights deals. The comparison concerns leverage, not the absence of an app. ↩
14. As note 3 explains, this is a proposed metric. Searches alone do not prove willingness to pay or switch services. Substitution tests must account for price and access; recommendation-led viewing can still retain subscribers. ↩
15. A title license and a subscription distribution deal transfer different rights. Customer access, exclusivity, guarantees, and reporting depend on the agreement. ↩
16. WBD’s letter (p. 3) describes the near-term earnings cost of supporting HBO Max. “Selling future leverage” is this essay’s inference, not proof of lost value or WBD’s stated motive. ↩
17. Netflix and WBD use viewing estimates in streaming amortization. An intent curve tracks deliberate demand over time. The proposal informs strategy, not accounting rules. ↩