Bundling is the most reliably successful pricing move in modern publishing: the New York Times attributed much of its subscription growth to selling News together with Games, Cooking, Wirecutter and The Athletic, with digital-only average revenue per subscriber — ARPU, revenue per subscriber in a period — above $9 a month by 2024 per its filings, and bundle subscribers described by the company as more engaged and less likely to cancel. But the mechanism that makes bundles work has a minimum scale, and pretending otherwise produces "bundles" of two barely distinct newsletters that confuse readers and convert nothing. This guide covers the mechanics, the minimum conditions, and the small-publisher variant.
Licht Journal publishes information, not pricing advice; figures are company disclosures and published platform data.
Why do bundles raise revenue at all?
Three separate effects, each measurable. First, perceived value: a subscriber evaluating five products anchors on total utility, not on the marginal worth of the one product they hesitated over. Second, churn reduction — the dominant financial effect. A subscriber who uses two products has two reasons to stay; the Times repeatedly disclosed that multi-product subscribers canceled less. A subscriber retained twelve extra months at $9 is worth more than any price increase you could have asked for. Third, cost dilution: each product added to a bundle is sold with near-zero incremental marginal cost, so bundle revenue converts to margin at a rate single products cannot match.
What is the minimum viable bundle?
Honest answer: three genuinely distinct use cases, or don't call it a bundle. Distinct means a reader could want one without the others — news, a game, a recipes app, a reviews vertical. Two products that serve the same reader at the same moment (a daily newsletter and a weekly digest of the same coverage) is one product described twice, and readers price it that way. The failure signature of premature bundling is visible in publishers' public retrospectives: a "bundle" launch that converts at the same rate as the single product, followed by quiet unwinding, because the second product never changed the calculus.
Related stories: How to cut subscription churn: the levers that actually move retention · The Athletic stopped losing money — the inside math of a sports-news turnaround.
How should the bundle be priced against single products?
The counterintuitive rule with strong support in operators' published experiments: make the bundle barely more expensive than the flagship single product — commonly 15-30 percent above it. The Times bundle followed exactly this shape, which is why most of its subscribers took the bundle: the marginal cost of adding four products was small enough that the choice felt non-economic. The bundle price is not the sum of parts; it is the flagship price plus a token increment. If your bundle costs the sum of its parts, readers will do that arithmetic and buy only what they came for.
| Pricing shape | Effect | When to use |
|---|---|---|
| Bundle ≈ flagship +15-30% | Most take the bundle; ARPU and retention rise | 3+ distinct products exist |
| Bundle = sum of parts | Readers cherry-pick; bundle exists in name | Almost never |
| Bundle below flagship price | Cannibalizes flagship revenue; margin leak | Never knowingly |
What is the small-publisher version?
Most outlets cannot build five products. The scaled-down mechanics that transfer: pair your subscription with one habit product — a daily game, a tool, a database, a calendar — built or licensed cheaply, that gives subscribers a reason to open the app on days without news. Judge the pair on retention delta: measure 90-day churn of paired-product users against single-product users, and if the gap is real, you have proved the bundle effect at your scale. The second scaled-down mechanic is partner bundling: two non-competing publications with adjacent audiences selling a joint subscription and splitting revenue — an old magazine-industry move that newsletter operators revived through 2024-2025 with cross-recommendation infrastructure doing the marketing.
What generalizes — and what doesn't?
What generalizes: retention is the payoff, price construction drives adoption, and distinctness of use cases is the entry ticket. What does not: the Times' ARPU figures, which reflect a product portfolio built over a decade with acquisitions; and the assumption that any second product helps — a weak second product drags onboarding, support and perception while adding nothing to retention. Bundle when the parts exist; until then, ship a better single product and bank the discipline.