A business model is the mechanism that turns what you built into cash you can keep, and unit economics is the arithmetic that tells you whether that mechanism actually works on a single customer before you scale it to a million. This domain covers both halves: how to design a revenue model and price it, and how to read the metrics — CAC, LTV, margin, MRR, burn — that tell you whether the design is sound. Get this wrong and growth doesn't save you; it just makes the losses arrive faster. Use this guide the way you'd use a mechanic's manual: look up the concept you're stuck on, run the worked example with your own numbers, and check your instinct against the "watch out."
Revenue Model
A revenue model is the mechanism by which value turns into money — subscription, transaction fee, usage, licensing, ads — and who pays it. It answers one question: when and how does the value you deliver turn into money?
Osterwalder & Pigneur's Business Model Canvas (Business Model Generation) treats Revenue Streams as its own building block — "if customers are the heart of a business model, revenue streams are its arteries" — and lists seven shapes: asset sale, usage fee, subscription, lending/renting, licensing, brokerage (a cut of each transaction), and advertising. Monetizing Innovation (Ramanujam & Tacke) narrows this to five models proven for new products: subscription, dynamic pricing (airlines, Uber surge), auctions (eBay, AdWords), pay-as-you-go on a metric close to customer value (GE charges airlines per mile flown, not per engine sold), and freemium. Their headline claim: "how you charge trumps what you charge" — Netflix didn't out-produce Blockbuster, it replaced the per-rental fee with a flat subscription.
The pattern across all of them: charge in the shape the customer receives value. Continuous value → recurring charge. Usage-driven value → usage-based charge. A tool that delivers one burst of value (a logo, a migration) forced into a subscription just manufactures month-two churn.
Worked example: two founders sell the same $meal-planning app. Founder A charges $120 once. Founder B charges $15/month. Breakeven month for B is $120 ÷ $15 = month 8 — anyone who stays past month 8 out-earns A's customer. At 6% monthly churn, average lifetime is 1 ÷ 0.06 ≈ 16.7 months, so B earns roughly $15 × 16.7 ≈ $250 per customer, more than double A. At 15% monthly churn, lifetime drops to 1 ÷ 0.15 ≈ 6.7 months (~$100), and A's one-time sale wins. Subscription isn't inherently better — it's a bet on retention.
Watch out: comparing models on year-one cash. One-time sales front-load revenue and restart you at zero every month; subscriptions back-load it. Compare revenue per customer over their full lifetime, not the first bank statement.
→ practice this in the revenue-models mission; model the trade-off with the mrr-arr-forecast calculator.
Pricing and Value-Based Pricing
Pricing is the decision of what to charge, anchored to the value delivered and the customer's alternatives — not to your costs. Value-based pricing sets price as a fraction of the measurable value the customer receives, rather than cost-plus or copying competitors.
Monetizing Innovation is blunt about where pricing starts: have the willingness-to-pay (WTP) talk with customers before the product is finished, ideally before it's built. Their diagnosis of most product flops: companies design the product first and hope about the money later. Porsche designed the Cayenne around what customers would pay for it; Fiat Chrysler famously "kicked out the finance guys" designing the Dodge Dart, stuffing pricing in the trunk — one of those cars printed money, the other flopped. Rob Walling (Start Small, Stay Small) gives bootstrappers the practical version: quantify the value in dollars. If your tool saves a customer 5 hours a month and their time is worth $25/hour, it delivers $125/month of value — you can't charge more than that, but everything below it is playground.
Once you know the ceiling, Monetizing Innovation says pick one of three strategies on purpose: maximization (charge the optimal price now — most products, most of the time), penetration (deliberately underprice to grab share fast — rational only with network effects or lock-in, and only if you actually raise prices later; their cautionary tale is LivingSocial, which burned $900M+ without anyone tracking how it would eventually make money), or skimming (start high for eager early adopters, walk the price down — iPhones, game consoles). Copying the market rate is choosing penetration by accident, with none of its payoff.
Worked example: 500 customers at $19/month = $9,500 MRR, with $4 variable cost per customer, so $15 contribution and $7,500/month gross profit. Raise to $29 (+52.6%): revenue-neutral point is $9,500 ÷ $29 ≈ 328 customers, so you can lose 172 (34%) and match today's revenue. Profit-neutral point uses the new contribution of $25: $7,500 ÷ $25 = 300 customers, so you can lose 200 (40%) before profit drops. Most founders overestimate how many customers a price raise will chase off.
Walling's practical tiering recipe: set your lowest tier, double it for tier two, double again for tier three ($19 / $39 / $79), make sure each step delivers more than double the benefit, and — tested but silly — end prices in 7, 8, or 9.
Watch out: cost-plus pricing. Slapping a familiar margin on cost tells you the floor of survival and nothing about the ceiling of value — Monetizing Innovation calls chronic underpricing a minivation (its opposite failure mode is feature shock: overbuilt and overpriced). Playmobil's Noah's Ark sold out and resold on eBay at 33% over list — that's not a hit, it's a mispriced hit.
→ practice this in the pricing-strategy mission; check your numbers with the pricing-breakeven calculator.
Unit Economics
Unit economics is the profit-and-loss of a single customer or transaction — does one unit of business make money after all its direct costs? Growth answers every question except this one; if a single customer loses you money, a million customers just lose it a million times faster.
Lean Analytics (Croll & Yoskovitz) calls LTV and CAC "the two essential metrics for a subscription business," built from four numbers in order: ARPU (what one customer pays per month), gross margin (the share left after direct costs of serving them), churn (with the shortcut that average customer lifetime ≈ 1 ÷ monthly churn), and CAC. Assembled: LTV ≈ (ARPU × gross margin) ÷ monthly churn, then compared to CAC.
Worked example (meal-kit subscription): customer pays $60/month, food and delivery cost $45. Gross margin = ($60 − $45) ÷ $60 = 25%, i.e. $15/month gross profit. If 20% churn monthly, average lifetime = 1 ÷ 0.20 = 5 months, so LTV = $15 × 5 = $75. Against a CAC of $90, LTV:CAC ≈ 0.83 — every new customer destroys about $15 before rent or salaries. Growing 20% a month on that ratio isn't a rocket, it's a leak with a growth rate.
Watch out: computing LTV on revenue instead of gross profit. $60 × 5 = "$300 LTV" against $90 CAC looks brilliant; it's fiction, because $225 of that revenue is food and fuel that never reaches you. Also watch for blended averages hiding disaster in one segment — Backupify (in Lean Analytics) discovered it was paying $243 to acquire consumers paying $39/year, invisible until the segments were split out.
→ practice this in the unit-economics-101 mission; the unit-economics calculator runs the full stack.
CAC
CAC (Customer Acquisition Cost) is total sales and marketing spend divided by the new customers it produced in the same period. Everything counts in the numerator — ads, discounts, referral bounties, and the salaries of the people running the campaigns, not just media spend.
Traction (Weinberg & Mares, cited alongside these frameworks) frames the acquisition side as CPA = CPC ÷ conversion rate: $1 clicks converting at 10% cost $10 per customer. The catch is that channels saturate — "tactics that once worked well will become crowded and ineffective" — so CAC is a curve, not a constant, and it rises as you press harder on any single channel or as competitors enter it. Product-Led Growth (Wes Bush) notes CACs have risen over 55% in five years industry-wide, which is exactly why PLG pushes free trials and freemium as a structural way to shrink CAC at the root: let the product sell itself instead of paying a rep or an ad platform for every signup.
Watch out: modeling payback or LTV:CAC at the CAC you paid for your last hundred customers in a fresh channel, when the number that matters is the CAC you'll pay for the next thousand in a channel that's starting to saturate.
→ the cac calculator; cross-reference with the growth domain's acquisition-channel concepts.
LTV and LTV:CAC Ratio
LTV (Lifetime Value) is the gross profit a customer generates over their entire relationship with you — driven by revenue per period, margin, and how long they stay. LTV:CAC ratio divides that by acquisition cost as a quick health check on whether customers are worth what you pay to get them.
David Skok's public essay "SaaS Metrics 2.0" popularized the benchmark most investors carry in their heads: an LTV:CAC of about 3 or better signals a healthy engine. Lean Analytics frames the same discipline differently — spend less than a third of a customer's lifetime value acquiring them, partly because the remaining two-thirds has to fund everything else (product, support, overhead), and partly because a tight CAC budget forces you to catch acquisition mistakes early rather than late.
Worked example — which lever actually moves the ratio: starting from the meal-kit numbers above (LTV $75, CAC $90, ratio 0.83): cutting CAC from $90 to $75 gets you to ratio 1.0 (barely breakeven). Halving churn from 20% to 10% doubles lifetime to 10 months, so LTV = $15 × 10 = $150, ratio 1.67 — one lever, ratio doubled. Raising price to $70 with the same $45 COGS lifts gross profit from $15 to $25/month (a 67% profit jump from a 17% price move, because price flows straight through margin); combined with 10% churn, LTV = $25 × 10 = $250, ratio 2.78 — knocking on Skok's 3. Product-Led Growth cites a finding from Teamwork.com (via Drew Sanocki): cutting churn 30%, raising ARPU 30%, and adding 30% more customers more than doubled LTV, because retention and price compound each other while adding customers only adds linearly. Pouring more into acquisition is usually the weakest lever on the board.
Watch out: treating LTV:CAC as the whole story. A ratio of 4 paid back over 5 months and a ratio of 4 paid back over 3 years are financially very different businesses — that's exactly what CAC payback measures instead.
→ practice both in the unit-economics-101 mission; ltv and ltv-cac-ratio calculators.
CAC Payback Period
CAC payback is the months of gross profit from a customer needed to recover the cost of acquiring them: Payback = CAC ÷ (monthly ARPU × gross margin). LTV:CAC tells you whether a customer is profitable; payback tells you when the cash comes back — and payroll is due monthly, not "over the customer lifetime."
Lean Analytics calls this time to customer breakeven and states the core problem plainly: there's a delay between paying to acquire customers and those customers paying you back, and your cash has to bridge that delay. Their example: a customer worth $27 over 11 months ($2.45/month) who costs $14 to acquire takes 5.7 months to break even — meaning you need 5.7 months of burn in the bank per cohort you acquire. Skok's "SaaS Metrics 2.0" pairs the LTV:CAC ≥ 3 rule with a second gate: aim for payback under roughly 12 months for SaaS. Two gates, not one — profitable and fast.
Worked example: ARPU $50/month, gross margin 80% → gross profit $40/customer/month. Churn 4%/month → lifetime 25 months → LTV = $1,000. CAC $250 → LTV:CAC = 4 (great on paper). Payback = $250 ÷ $40 = 6.25 months (fine, under 12). Now triple acquisition spend from $50k to $150k/month at the same $250 CAC: 600 new customers/month, each cohort costs $150,000 and returns 600 × $40 = $24,000/month. After one quarter you've spent $450,000 on three cohorts that have together repaid only $72k + $48k + $24k = $144,000 — a net cash outflow of $306,000 in three months, on customers who are individually profitable. Faster growth with a healthy payback doesn't produce cash, it consumes it, in proportion to how hard you press the pedal.
Watch out: computing payback on revenue instead of gross profit ($250 ÷ $50 = 5 months looks fine; the real $250 ÷ $40 = 6.25 is 25% longer — the gap that breaks payroll at scale). Also watch payback exceeding the average customer lifetime: a 6.25-month payback against a 25-month lifetime is safe, but if churn were 15% (lifetime ≈ 6.7 months) the average customer would barely repay their own acquisition cost before leaving.
→ practice in the cac-payback mission; cac-payback calculator.
ARPU
ARPU (Average Revenue Per User) is total revenue in a period divided by the number of active customers in that period — for SaaS, the common formula is ARPU = Total MRR ÷ Total Users. Product-Led Growth flags the definitional trap up front: "user" is ambiguous. For Netflix it means a paying subscriber; in B2B, one "account" may contain many users; in freemium, only paying users should count if you want ARPU to mean anything. Teams that need precision use ARPPU (Average Revenue Per Paying User) to be explicit that non-payers are excluded.
Worked example: $100,000 MRR ÷ 5,000 customers = $20 ARPU. If only 1,000 of those 5,000 are paying (the rest are on a free tier), ARPPU = $100,000 ÷ 1,000 = $100 — a five-fold difference depending on which denominator you pick, which is why the definition has to be stated every time the number is quoted.
ARPU feeds directly into LTV, CAC payback, and pricing decisions: it's the revenue side of every gross-profit-per-customer calculation elsewhere in this guide.
Watch out: silently switching between "all signups" and "paying customers" in the denominator between reporting periods — it makes trends look like they moved when only the definition did.
→ arpu calculator.
Gross Margin and COGS
Gross margin is revenue minus the direct cost of delivering the product, as a percentage of revenue — the slice of each dollar you keep to fund everything else. COGS (Cost of Goods Sold) is those direct costs themselves: hosting, support, payment fees, physical goods, delivery.
The Business Model Canvas makes Cost Structure its own building block because founders tend to ignore it until it bites. Osterwalder & Pigneur split it into variable costs (scale with each sale — wax, jars, shipping, hosting) and fixed costs (arrive whether you sell or not — salaries, rent, tools), and further into cost-driven models (Ryanair — strip every cost, win on price) versus value-driven models (luxury hotels — cost follows the value proposition). Lean Analytics singles out software's structural advantage: marginal costs that trend toward zero, so the nth customer costs almost nothing to serve and revenue can grow while costs stand still.
Worked example — two founders, same $50k/month revenue: A candle maker sells 2,000 candles at $25, with $17.50 of wax/jar/shipping/packaging per candle → $35k COGS. Gross margin = ($25 − $17.50) ÷ $25 = 30%, i.e. $7.50 contribution per candle. A software founder has $50k of subscription revenue and $6k of variable costs (servers, support) → 88% gross margin. Both hire a $6k/month marketer (a fixed cost, paid out of contribution, not revenue): the candle founder needs $6,000 ÷ 0.30 = $20,000/month of extra sales — forever — just to cover the hire; the software founder needs only $6,000 ÷ 0.88 = $6,818/month. Same salary, radically different bet.
Watch out: paying a fixed cost out of top-line revenue in your head instead of dividing it by gross margin first — "the marketer costs $6k and we do $50k in sales, easy" ignores that only 30 cents (or 88 cents) of every dollar actually survives contact with COGS.
→ practice in the margins-and-costs mission; unit-economics and pricing-breakeven calculators.
Contribution Margin
Contribution margin is revenue per unit minus the variable costs of delivering that unit — what each additional sale contributes toward fixed costs and profit. It's the same subtraction as gross margin, just expressed as a dollar amount per unit rather than a percentage, and it's the number that answers "how many units until I cover my fixed costs?"
Breakeven units = fixed costs ÷ contribution per unit. Using the candle example above: fixed costs of $9,000/month ÷ $7.50 contribution per candle = 1,200 candles to break even. At 2,000 candles sold, contribution is $15,000 and profit is $15,000 − $9,000 = $6,000/month. If price rises 10% to $27.50 with costs unchanged, contribution jumps to $10/candle (36.4% margin), adding $5,000/month of pure profit at the same volume and dropping the with-marketer breakeven from 2,000 to $15,000 ÷ $10 = 1,500 candles. A 10% price move did in one line what months of hustle couldn't.
Contribution margin is also the number underneath unit economics: CAC is only "worth it" if the contribution margin per customer, summed over their lifetime, clears it — which is exactly what LTV measures once you multiply contribution by expected lifetime.
Watch out: confusing "adding volume" with "improving contribution." Selling 10% more candles adds $1,500 of contribution; raising price 10% adds $5,000 and lowers your breakeven point. Volume is usually the weakest lever in the room, not the strongest.
→ unit-economics and pricing-breakeven calculators.
MRR and ARR
MRR (Monthly Recurring Revenue) is the normalized subscription revenue you can expect every month, excluding one-time payments. ARR (Annual Recurring Revenue) is MRR × 12 — the standard headline number investors and boards track for subscription businesses. Both matter because, unlike a services business, a subscription company starts each month with revenue already substantially "sold" — the number you have to hit isn't zero, it's last month's MRR minus whoever churned.
Product-Led Growth treats MRR movement as the operating pulse of a SaaS company: a single fix (removing a friction step from onboarding at Snappa) produced a 20% boost in MRR, translating to a projected six-figure ARR impact. The book also distinguishes customer churn from revenue churn: you can have customer churn under 1% and still lose 40% of MRR if the one customer you lost was your biggest — Revenue Churn = Churned MRR ÷ Total MRR.
Worked example: you have $100,000 total MRR. Your single largest account, worth $40,000/month, cancels. Customer churn might be a single logo out of hundreds (technically <1%), but revenue churn = $40,000 ÷ $100,000 = 40%. ARR before the loss: $100,000 × 12 = $1.2M; ARR run-rate after: $60,000 × 12 = $720,000 — nearly halved by one departure.
Watch out: reporting customer-count churn while revenue churn is quietly telling a different, scarier story. Concentration risk in a handful of large accounts makes the two diverge badly.
→ practice in the revenue-models mission; mrr-arr-forecast calculator.
Burn Rate and Runway
Burn rate is the net cash a company loses per month — expenses minus revenue — while it isn't yet profitable. Runway is cash in the bank divided by monthly burn: the months you have left to reach profitability or the next funding milestone.
Lean Analytics ties burn directly to the CAC-payback math covered above: the 5.7-month time-to-customer-breakeven example (a customer worth $27 over 11 months, costing $14 to acquire) comes with an explicit warning — "you now know you need 5.7 months' burn to keep the company running" per cohort of customers you acquire, because the cash goes out before it comes back. Start Small, Stay Small frames the same discipline in bootstrapper terms as breakeven — revenues exceeding costs on a regular basis — and calls it irresponsible not to think about it even if you're deliberately not optimizing for profit yet, "because if there's no way you can ever get there, you're just burning money and time."
Worked example: $300,000 in the bank, burning $25,000/month net → runway = $300,000 ÷ $25,000 = 12 months. If a new acquisition push raises burn to $40,000/month, runway on the same $300,000 collapses to $300,000 ÷ $40,000 = 7.5 months — and a push the size of the tripled-spend scenario in the CAC payback section (≈ $102,000/month of net cash outflow) would shrink it far more. Growth spend and runway are the same dollars viewed from two directions.
Watch out: treating runway as a fixed number instead of a function of your current burn decisions. Every acquisition push, hire, or pricing change moves it — recompute after each one, not once a quarter.
→ burn-runway calculator; runway decisions connect directly to the strategy-funding domain's fundraising-timing concepts.
Subscription and SaaS
A subscription is a revenue model where customers pay on a recurring schedule for ongoing access, making retention the engine of revenue rather than a one-time close. SaaS (Software as a Service) is software delivered and billed as an ongoing subscription, typically carrying high gross margins (80%+ is common, as seen in the margins example above) and economics driven almost entirely by retention.
Start Small, Stay Small makes the bootstrapper's case for subscription/SaaS directly: "it's a recurring revenue stream that can grow far beyond what you would make as a consultant, all the while creating balance sheet value" — a hosted web application aimed at businesses combines ease of support, ease of adoption, and a recurring revenue model, which the book calls a major structural advantage over one-off project work. The tradeoff, as the revenue-model section above shows with the $120-vs-$15/month example, is that subscription economics only pay off if retention holds; the whole model collapses into a bet on churn.
Because SaaS revenue compounds month over month (this month's revenue = last month's revenue minus churn plus new sales), it's the business shape where MRR, ARR, LTV, CAC payback, and gross margin all interlock most tightly — nearly every metric in this guide was built with a SaaS company as the default mental model.
Watch out: assuming "subscription" automatically means "good business." A subscription with high churn is just a one-time sale that takes longer to collect and costs more to administer.
→ practice in the revenue-models mission; the saas metrics above all apply directly (mrr-arr-forecast, cac-payback, ltv-cac-ratio).
Freemium
Freemium is a model where a free tier acquires and activates users at scale, and a paid tier converts the fraction who hit the free tier's limits. Osterwalder & Pigneur (Business Model Generation) trace the term to Jarid Lukin, popularized by VC Fred Wilson: a large free user base, of which "usually less than 10 percent" ever convert to paid, with the paying minority subsidizing everyone else — made possible only because the marginal cost of serving an additional free user is low. Flickr is their canonical example: unlimited basic sharing for free, paid "pro" removes storage and upload caps. Skype is the extreme version — because calls route peer-to-peer instead of over owned telecom infrastructure, its cost structure let free users cost almost nothing to serve.
Product-Led Growth (Wes Bush) makes a sharper distinction than the canvas does: freemium is an acquisition model, not a revenue model — the free tier's job is to fill the funnel; something else still has to convert and collect, which is why freemium shows up as example_of revenue-model in the ontology but really lives at the intersection of acquisition and monetization. Bush quotes Rob Walling's warning that freemium "is like a samurai sword: unless you're a master at using it, you can cut your arm off" — a free tier generous enough to satisfy everyone is a charity; one that helps no one is a demo. The two metrics to watch, per BMG, are the cost of serving a free user and the free-to-paid conversion rate.
Watch out: treating a low free-to-paid conversion rate as a marketing problem instead of a packaging problem. If the free tier already does everything most users need, there's nothing left for the paid tier to sell.
→ practice freemium mechanics in the revenue-models mission; conversion-funnel calculator for the free-to-paid rate.
Marketplace
A marketplace is a business that matches supply and demand and takes a cut of each transaction, facing the chicken-and-egg problem of growing both sides at once. Osterwalder & Pigneur categorize this as a multi-sided platform: a business that brings together two or more distinct but interdependent customer segments, where the platform is valuable to one side only if the other side is also present — Visa needs both cardholders and merchants; a free newspaper needs both readers and advertisers. The defining dynamic: "a multi-sided platform grows in value to the extent that it attracts more users" — the network effect.
Lulu.com (in Business Model Generation) illustrates the pattern in publishing: it eliminated the traditional "market-worthy work" gatekeeping by letting anyone publish, and because authors become customers, the more authors it attracts the more it succeeds — a self-reinforcing loop rather than a linear sales funnel. This is why marketplaces are structurally harder to bootstrap than single-sided products: you can't just sell to one customer, you have to seed both sides simultaneously before either side sees enough value to stay.
Watch out: subsidizing one side of the marketplace indefinitely without a plan to eventually monetize it (or the other side). Multi-sided platforms need a deliberate design for which side pays and which side is subsidized to attract the paying side — that choice is a pricing decision, not an afterthought.
→ marketplace network-effect dynamics are covered further in the growth domain (cold-start problem, hard side).
Go deeper
- Osterwalder & Pigneur, Business Model Generation — the canvas, Revenue Streams and Cost Structure blocks, freemium and multi-sided platform patterns.
- Ramanujam & Tacke, Monetizing Innovation — willingness-to-pay methodology, the three pricing strategies, minivations and feature shocks.
- Croll & Yoskovitz, Lean Analytics — LTV/CAC mechanics, time-to-customer-breakeven, margins by business model.
- Wes Bush, Product-Led Growth — freemium as acquisition not revenue, ARPU/ARPPU definitions, revenue vs. customer churn.
- Rob Walling, Start Small, Stay Small — value-based pricing for bootstrappers, tiering, breakeven discipline without VC cash.
- David Skok, "SaaS Metrics 2.0" (public essay) — the LTV:CAC ≈ 3 and payback-under-12-months benchmarks referenced throughout this guide.