Recurring Revenue: Building a Subscription Business

Every subscription business runs on the same four numbers, and most owners can only name one of them. This guide walks the full stack: the three model types — box, SaaS and membership — then the MRR math that runs everything, churn benchmarks and the levers that move them, billing technology, dunning, annual-plan nudges and a complete unit-economics check. We use one worked example throughout: an illustrative specialty-coffee subscription with 2,400 subscribers paying $32 a month, so every formula lands on a number you can check.

The ground rules are the same ones we use with clients. Published benchmarks from Recurly, Zuora and UBS carry the macro claims, and every modelled number is labeled illustrative with its assumptions on the table. Rerun each calculation with your own subscriber count, and the priority list usually reorders within an hour.

Here is the short version before we go deep.

Key Highlights

• UBS analysts estimated the subscription economy at $650 billion in 2020, heading to $1.5 trillion by 2025; Zuora's Subscription Economy Index has repeatedly measured subscription businesses growing several times faster than product-selling peers.
• Recurly's published benchmarks put involuntary churn near 1% of subscribers monthly; total monthly churn runs roughly 5–8% for consumer products and lower for B2B SaaS.
• The coffee example holds $76,800 of MRR (2,400 × $32) and grows about 36% in twelve months on net new MRR of $6,000–$7,600 a quarter (illustrative).
• At 6.5% monthly churn, the model must replace 156 subscribers and $4,992 of MRR every month just to stand still.
• Dunning that recovers roughly half of involuntary churn leaves about 85 more subscribers at month twelve — roughly $2,700 of MRR (illustrative).
• Moving 30% of subscribers to a $320 annual plan brings in about $230,400 of cash upfront and cuts logo churn sharply in the model (illustrative).
• Unit economics: at $32 ARPU, 70% gross margin and 6.5% churn, lifetime value ≈ $345 against a $45 CAC — 7.7:1, with roughly a two-month payback.
• Let churn drift to 8% and LTV falls to $280 — a 19% swing caused by one number nobody owns.

Why Recurring Revenue Changes the Math

A transactional business wakes up at zero every month: last month's customers must be re-won one sale at a time. A subscription business wakes up already knowing most of its revenue, because monthly recurring revenue (MRR) — active subscribers multiplied by average revenue per subscriber — arrives whether or not you sell anything new today. The illustrative coffee business starts every month at $76,800 before a single new sign-up.

The macro numbers explain the pull of that predictability. UBS analysts estimated the subscription economy at $650 billion in 2020 and forecast $1.5 trillion by 2025, and Zuora's Subscription Economy Index has repeatedly measured subscription-billing businesses growing several times faster than product-selling peers over the past decade. Neither figure predicts your churn — but both explain why buyers pay higher multiples for revenue that repeats.

Day to day, the difference is operational. You can hire against next quarter instead of next week, buy green coffee on contract instead of spot, and spend on acquisition with payback math you actually trust. Transactional competitors must price all three risks into every decision they make.

Key takeaway: Recurring revenue is a forecasting instrument before it is a financing story. The moment you can predict next month within a few percent, every decision after it — hiring, inventory, acquisition — gets cheaper to make.

Model 1: Box, SaaS or Membership — Pick the Flavor That Fits

Three flavors of subscription dominate the small-business landscape. A box ships physical goods on a schedule — coffee, meals, razors — and lives or dies on logistics and product delight. SaaS sells software access, where the product improves without a single new shipment. A membership sells access to people, perks or content — gyms, communities, courses — and leans hardest on identity and habit.

The coffee example is a box, so it carries the heaviest churn and cost profile of the three. The math that follows — MRR, churn, dunning, lifetime value — works identically for all three; only the operations change. When we build subscriber apps and platforms, the model choice drives the feature list more than any technology decision, as the membership logic in our fitness training app case study shows.

One warning: hybrids multiply operations before they multiply revenue. Pick one primary model for your first year, then layer on the second.

Key takeaway: Your model sets the churn ceiling and the operations load before the first subscriber signs up. Choose the flavor whose monthly operations you can genuinely staff — the billing math is the easy part.

Model 2: MRR Math — Four Numbers Run the Whole Business

MRR moves through exactly four doors every month. New MRR arrives from first payments, expansion MRR comes from upgrades and add-ons, contraction MRR leaves through downgrades and pauses, and churned MRR leaves through cancellations. Net new MRR — the sum of the four — is the only growth number that matters at the board level.

Charted by quarter, the same mechanics build a 36% year: net new MRR of $6,000–$7,600 a quarter compounds from $76,800 to about $104,400 in twelve months (illustrative). The uncomfortable corollary: at 6.5% monthly churn, 156 subscribers and $4,992 of MRR must be replaced every month before growth even starts. Acquisition is not the growth engine — it is the fuel line feeding an engine with a leak.

Key takeaway: Track New, Expansion, Contraction and Churned MRR as four separate line items from day one. The net figure hides which problem you have; the four sub-numbers tell you exactly which lever to pull.

Model 3: Churn — Benchmarks and the Levers That Actually Move It

Churn splits into two very different problems. Voluntary churn is customers choosing to leave — poor fit, price, novelty fading. Involuntary churn is failed payments — expired and reissued cards, soft declines, stale billing details — and Recurly's published benchmarks put it near 1% of subscribers per month, roughly a sixth of the coffee example's total 6.5%.

Benchmarks vary by model, and published ranges make that visible: consumer boxes commonly run 5–8% total monthly churn, membership products a point or two lower, B2B SaaS on monthly billing around 2–4%, and annual contracts lower still. Treat any benchmark as a range rather than a target — mix, price point and onboarding move the number more than industry labels do. What matters is knowing which end of your range you are on and which direction it is drifting.

Measure churn by cohort — the month people joined — rather than as one blended number. Blended churn mixes your onboarding failures with your vintage successes and tells you nothing about either. A simple cohort grid in a spreadsheet exposes both within a quarter.

Key takeaway: Churn is not one number; it is voluntary versus involuntary, measured by cohort, attacked by lever. Start with onboarding — it is the cheapest lever, and it compounds across every subscriber you will ever acquire.

Model 4: Billing Tech and Dunning — Recover the Revenue You Already Sold

Billing technology is the plumbing under everything above, and it comes in three tiers. Platform-native tools — Stripe Billing, Shopify subscriptions, WooCommerce plugins — cover simple needs at near-zero software cost. Dedicated subscription platforms (Recurly, Chargebee, Maxio) add smart retries, card updaters, proration and dunning analytics. Custom billing buys exactly what you specify, at the price of owning compliance and every edge case forever — we recommend it only when billing is the product, as in our fintech wallet case study.

The dedicated tier exists because of dunning — the automated craft of recovering failed renewals. At the coffee example's scale, roughly 1% involuntary churn means about 24 failed payments and $768 at risk every month; recovering around half puts $384 a month back immediately (illustrative). The dunning ladder runs: smart retry schedules, network account updaters that refresh changed card details automatically, pre-dunning emails that ask customers to fix a card before cancellation, and a grace period before access actually stops.

Compounded over a year, the model leaves about 85 more active subscribers — roughly $2,700 of MRR — at month twelve than letting failed payments lapse (illustrative). Dunning tooling costs a small fraction of that, which makes it one of the highest-ROI hours in subscription operations. The email half of the ladder is retention craft we detail in our coffee subscription email case study, and the payment rails underneath it are covered in our payment integration guide.

Key takeaway: Dunning recovers revenue you have already earned, with tooling that pays for itself inside the first month of recoveries. Turn it on before you spend another dollar on acquisition.

Model 5: Annual Plans and Unit Economics — The Margin Check

Annual plans are the cheapest churn lever most subscription businesses never pull. The standard shapes: two months free on the annual price (an effective ~17% discount), a small surcharge on monthly billing, and upgrade prompts timed to engagement peaks. ProfitWell's published research and billing-platform data consistently show annual-plan subscribers churning well below monthly-plan peers; in our model we assume about one-third the monthly rate (illustrative).

The cash effect is immediate. Moving 30% of the coffee base — 720 subscribers — onto a $320 annual plan brings in about $230,400 upfront, enough to fund roughly 5,100 new subscribers at a $45 CAC (illustrative). Prepaid revenue also strengthens your hand with suppliers, which is why consumer brands push annual hardest ahead of peak seasons.

Two heuristics make the table actionable. A 3:1 LTV:CAC floor with payback under three months is the commonly used line for consumer subscriptions; at 7.7:1 and two months, the coffee example can fund growth from its own cash flow. Pricing also deserves its own experiment calendar — Price Intelligently's published analysis puts the profit impact of a 1% pricing improvement at several times that of a 1% acquisition improvement.

Key takeaway: Annual plans convert a churn problem into a cash-flow asset, and unit economics decide whether growth is worth funding at all. If payback stretches past three months, fix retention or pricing before you buy another ad.

What Your Business Should Do

Days 1–30: Instrument the model.

• Compute MRR, ARPU, logo churn and revenue churn from the last 90 days of billing data, splitting voluntary from involuntary churn.
• Map your billing stack against the three tiers and list which dunning pieces — retries, account updaters, pre-dunning email — are missing.
• Name the model you are really running (box, SaaS or membership) and write down the three monthly operations it demands.

Days 31–60: Stop the leaks.

• Switch on smart retries and account updaters, and write a three-email pre-dunning sequence with a grace period before cancellation.
• Rebuild onboarding around the first 30 days and add a pause-or-skip option one screen before the cancel button.
• Price the annual offer — two months free is the standard shape — and schedule the upgrade prompt for your happiest cohort.

Days 61–90: Tune the economics.

• Recompute LTV, CAC and payback with the new churn numbers; set a 3:1 LTV:CAC floor as the go/no-go line for acquisition spend.
• Run one pricing experiment — annual default or monthly surcharge — and judge it on net revenue retention, not raw sign-ups.
• Stand up the monthly MRR board: New, Expansion, Contraction, Churned, Net New — one owner, one page, reviewed on the first.

The Bottom Line

Subscription businesses do not win on the size of their launch; they win on the four MRR numbers, the churn levers and the plumbing that recovers what was already sold. The coffee example's entire edge — a 36% growth year, 85 recovered subscribers, a 7.7:1 LTV:CAC — came from math a two-person team can run in a spreadsheet. Start with the dunning ladder and the cohort grid this week; the compounding starts immediately.

If you want a second pair of eyes on your model, Book a free strategy call and we will walk your MRR, churn and unit economics with you — or send us a message with your questions and we will answer within one business day.