The Missing Metrics in Streaming

Ask a streaming executive about customer economics and they can answer in detail. Customer acquisition cost. Lifetime value. Churn. These metrics shape board conversations, valuation models, marketing budgets, and growth strategy. If acquisition costs rise or churn moves half a point in the wrong direction, the change is visible, an executive owns it, and the company expects an explanation.

That discipline fades when attention turns to content itself.

Content is the product customers came to see. It attracts audiences, keeps subscribers engaged, creates advertising inventory, and gives a streaming service its identity. Yet many companies struggle to explain what a title or collection truly cost, what value it created, or how long that value lasted.

The problem begins with how companies draw the boundaries. Content operations and the media supply chain describe only part of the system. The economic unit is the entire content value chain: acquisition strategy, inbound contracts, asset delivery, processing, storage, programming, curation, distribution, outbound contracts, and monetization. Every place the media travels shapes what it costs or what it can earn.

Today those stages are scattered across the org chart. The license fee sits with acquisitions. Ingestion and quality control sit with operations. Encoding and storage sit with technology. Merchandising and curation sit with programming. Revenue arrives through another system with another owner and another definition of success.

Each function holds one piece of the truth. No executive owns the whole picture.

Streaming needs a content-side version of the financial discipline it already applies to customers. CAC, LTV, and churn offer a useful starting framework, with one important caveat: content behaves very differently from a customer. Its economics are interdependent, sensitive to context, and often shared across a portfolio.

That complexity makes better measurement essential.

Measure the full cost of making content productive

Content acquisition cost usually means the license fee, production budget, or value of the deal. Those figures capture the initial commitment. The full economic cost is much broader.

Before a title can earn anything, a company may have to interpret rights, receive and inspect assets, repair files, normalize metadata, create artwork, encode versions, produce captions and audio descriptions, localize materials, clear music, configure windows, build avails, merchandise the title, deliver it to partners, and continue servicing it as platform requirements change.

Some titles move through that sequence cleanly. Others arrive with incomplete rights records, missing assets, conflicting identifiers, and metadata that must be rebuilt by hand. Every gap creates another handoff, decision, or workaround. The work lands on people whose time was budgeted for something else, and the delay spreads into teams that may never know where it started.

At Pluto TV, where I built the supply chain largely from the ground up, a title that normally moved through for $1.25 per minute of content could cost four to five times the business-as-usual rate, or more, when assets arrived outside specification or the supply chain was poorly equipped for the problems they carried. A low license fee attached to a high-friction package may be more expensive than it appears. The deal still looks good in the acquisition system. The people repairing it know better.

We could track those costs because I had visibility into acquisition contracts and input on their operational terms. I wrote the delivery specification, and we created rate cards that defined business-as-usual processing and overage work. We knew which content our supply chain handled efficiently, where it struggled, and how much those exceptions cost.

We reported significant overages back to acquisitions so the team could include processing in the content’s profit analysis. Several times, we canceled content orders because repair and delivery costs had climbed high enough that we saw no credible path to earning the money back. The operating data changed the acquisition decision. That is what a useful metric does.

Companies need a measure of total content cost: the full cost of making content capable of earning. Track it at the title level when the economics are title-specific, at the partner level when delivery quality drives the burden, and at the portfolio level when a collection is acquired and monetized together.

Define value by how the content is used

Content value is harder to calculate because a title can perform several jobs at once.

A film may generate transactional revenue, attract subscribers, reduce cancellation, create advertising inventory, strengthen a genre collection, support a bundle, or travel through secondary licensing windows. A library series may never drive a subscription, yet become the program someone watches every night for six months. A FAST title can generate attributable ad revenue while improving the schedule and increasing the value of surrounding inventory.

The value is real. Attribution is where reasonable people can build equally polished spreadsheets and reach very different answers.

Financial accounting already recognizes this challenge. Disney’s 2025 annual report separates content that derives lifetime value from title-specific revenue from content monetized as a group. Subscription streaming content generally falls into the second category because the revenue belongs to the bundle and resists clean title-level attribution. 

Leaders should resist the demand for one universal equation. A useful content value model can include direct revenue, acquisition and retention contribution, advertising yield, secondary licensing, portfolio depth, and the ability to repackage or distribute an asset in new ways.

The model should reflect the company’s actual strategy. Two services may own the same catalog and create different value from it because programming, curation, audience, packaging, and distribution change performance. Wall Street will understandably want comparable figures, yet these metrics will never be perfectly apples to apples across companies. The more immediate problem is that most media companies have no internal baseline at all. Leaders often cannot say what business-as-usual processing costs per minute or where an overage begins.

Some inputs belong in a dollar figure. Others should remain directional scores with confidence ranges. “High retention value and low directly attributable revenue” may be more decision-useful than an impressive but fictional $4.7 million lifetime value.

A model can be precise to the penny and still send the company in the wrong direction.

Manage content across its useful life

Customer churn is an event. Content follows a curve.

A new release may produce most of its viewing early, settle into a long tail, spike after an award, rise again when a new season launches, or become useful in a different package years later. Licensed content may leave when its rights window closes even while demand remains. News and sports can lose value quickly. A durable library title may keep earning in small increments for years.

Netflix amortizes content using historical and estimated viewing patterns, generally on an accelerated basis because it expects more viewing early in a title’s life. It expects more than 90% of a licensed or produced asset to be amortized within four years of first availability. Amazon similarly reviews viewing patterns over time and reports a weighted average remaining life of 3.2 years for capitalized video content. 

Those accounting treatments establish a useful principle: content has an economic half-life. Leaders need to understand how quickly each type loses effectiveness, what slows that decline, and which events can reactivate demand.

Measure that curve by cohort. Scripted and reality will behave differently. Film and episodic television will behave differently. FAST channels and on-demand libraries will behave differently. Regional patterns will differ as well. A single content churn rate would erase the distinctions leaders need to make good decisions.

The goal is a usable map of content behavior over time.

Use the framework to improve real decisions

Customer metrics supply the management questions: What did we spend? What did we get back? How long did the value last? Content economics require their own calculations because placement, recommendation, packaging, release timing, marketing, and the surrounding catalog all affect performance.

Start with one portfolio, partner, channel, or genre where a commercial decision is active. Map costs from contract through ongoing servicing. Separate directly attributable costs from those requiring allocation, and document the rules. A disclosed assumption can be debated and improved. A hidden one quietly becomes policy.

Define value for that use case. A FAST channel may emphasize ad revenue, fill rate, watch time, distribution, and refresh requirements. A subscription library may emphasize retention cohorts, reactivation, completion, search demand, and portfolio depth. A licensing relationship may depend primarily on window revenue and operational burden.

Segment content by business purpose, monetization model, genre, territory, partner, and lifecycle stage. A premium original and an acquired library package were built to do different jobs and deserve different benchmarks.

Establish a cross-functional value-chain review. This is the linchpin. Finance should own economic integrity. Programming and acquisitions should own strategic intent. Data teams should own measurement logic. Operations should own the cost and friction of moving content through the system. Distribution and sales should bring the economics of outbound deals. The review creates one place where those views and incentives must reconcile.

Test the model against real decisions. Should this deal renew? Is this partner still attractive after remediation costs? Does this channel need more content or better scheduling? Is a title underperforming, or was it poorly packaged? Which workflow fix returns enough margin to fund?

A useful model should answer those questions early. Eighteen months of data transformation signals that the initial scope is too broad.

The operating model is the real metric

Streaming has reached a stage where growth no longer covers every operational ambiguity. Disney reported $1.3 billion in direct-to-consumer operating income for fiscal 2025, up from $143 million the year before, and is targeting a 10% operating margin for its entertainment subscription streaming business in fiscal 2026. As margin pressure intensifies, leaders need to know which decisions created value and which merely shifted cost somewhere less visible.

The companies that succeed in the next phase will manage the full content value chain. They will connect what they buy, what it costs to make useful, how it is programmed and curated, where it travels, and what it returns to the business. That visibility will help leaders place better bets, identify deteriorating economics earlier, and invest in the capabilities that protect margin.

Content is the product. Managing it well begins with seeing its economics clearly.

---

Rebecca Avery is a Senior Streaming Operations Executive, SME of the SVTA Metadata Working Group, and writes about the operational realities of streaming media.