Streaming CAC, LTV, and Churn Model: A Cohort Guide to Unit Economics

Why a blended average hides the real business

The paper's central profitability argument is that contribution margin per user has replaced subscriber count as the governing metric. Two subscribers at the same monthly price can have different economics. One may arrive organically and remain direct; another may arrive through a paid campaign or bundle that lowers acquisition cost but gives an aggregator part of the revenue.

A blended company average erases those differences. The model should keep acquisition month, channel, product tier, price, geography, device or platform source, and bundle status attached to each cohort.

Define CAC as a cohort investment

CAC should include the spend required to create a paying or monetizable user, not only the media invoice. Include attributable marketing, incentives, agency or partner fees, and onboarding costs, then divide by qualified additions for the same cohort. Keep organic, paid, bundled, and reactivated users separate so that low-cost additions do not disguise expensive paid acquisition.

The whitepaper cites sharp CAC inflation in a saturated US market and warns against using earlier growth-era assumptions. Its strategic lesson is more important than any one range: forecast CAC by channel and scenario, compare it with gross margin payback, and avoid scaling a channel simply because it produces volume. A long payback period increases exposure to churn before acquisition spend is recovered.

Calculate lifetime contribution, not revenue-only LTV

Revenue-only LTV overstates value because streaming carries content, infrastructure, payment, platform, support, and advertising costs. Use the paper's full contribution equation instead. Model that contribution by month until the cohort becomes inactive, then discount uncertain distant cash flows if finance requires it.

For ad-supported users, include viewing hours, monetizable ad opportunities, fill, effective CPM, ad-tech leakage, and revenue share. For subscribers, include discounts, failed payments, and platform taxes. A lower-priced ad tier can produce strong lifetime contribution when fixed content costs are already covered, advertising adds revenue, and the tier expands the addressable market or improves retention.

Model churn as a curve and a cause

A single monthly churn rate is easy to use but weak for decisions. Build a survival curve that shows how many members of a cohort remain after each billing period. Separate voluntary cancellation, payment failure, bundle expiration, and reactivation. Early-tenure churn often signals acquisition or onboarding problems; later churn may reflect content gaps, price sensitivity, or ad fatigue.

Then connect churn to controllable drivers. The paper highlights bundles, content engagement, ad load, and predictive intervention as important levers. Test scenarios for a price change, a different bundle mix, improved payment recovery, a retention offer, or an ad-load increase. The model should show both added revenue and the retention damage required to erase it.

Turn the model into a monthly decision system

Review each cohort against the assumptions approved when spend was committed. Track actual CAC, first-period activation, retention, monetization, payback, and forecast lifetime contribution. When results diverge, identify whether acquisition quality, engagement, pricing, advertising yield, platform fees, or churn caused the miss. That diagnosis is more actionable than reporting that total subscribers were above or below plan.

Use explicit rules: scale cohorts that meet contribution and payback thresholds; redesign channels that acquire users who never activate; renegotiate bundles that add stability but destroy margin; and stop promotions whose discounted lifetime value cannot recover acquisition cost. A mature model links marketing, product, content, ad sales, and finance to one economic truth.

• Keep organic, paid, bundled, and reactivated cohorts separate.
• Use contribution margin after variable costs, not revenue-only LTV.
• Build retention curves rather than relying on one churn average.
• Stress-test price, bundle, ad-load, and retention scenarios.
• Reconcile approved assumptions with actual cohort behavior monthly.