Cost Structure

Origin & evolution

The block belongs to the canvas Alexander Osterwalder and Yves Pigneur published in Business Model Generation (2010), one of nine components that describe how an organisation creates, delivers, and captures value. Cost Structure sits on the efficiency side of the canvas, directly under Key Resources🏭Key ResourceBusiness ModelA key resource required by the businessView reference → and Key Activities🛠️Key ActivityBusiness ModelA key activity the business performsView reference →, because those two blocks are what generate the cost. The model's logic is causal: name the resources and activities a model depends on, and the major costs follow.

Osterwalder and Pigneur drew a distinction that still organises the conversation. A cost-driven model minimises expense wherever it can, through low-price value propositions, heavy automation, and outsourcing; a no-frills airline is the standard example. A value-driven model worries less about the cost of producing and more about the value created, accepting a premium structure to fund high-touch service or luxury positioning🎪PositioningGo-To-MarketProduct positioning statementView reference →. Most real businesses sit somewhere between the two poles, and the block forces a team to say where.

The accounting vocabulary underneath is older than the canvas. Fixed costs stay constant regardless of volume, such as rent and salaries. Variable costs scale with output. Economies of scale lower the per-unit cost as volume rises, and economies of scope lower it when one resource or activity serves several products. The canvas did not invent these terms; it put them where a founder could see them sitting next to the revenue they are meant to justify.

How it works in practice

A subscription🔂SubscriptionSales & RevenueA recurring subscriptionView reference → analytics company charges £40 per seat per month. Its Cost Structure has three large lines. Cloud infrastructure is variable: roughly £6 per active seat, rising in step with usage. Engineering and product salaries are fixed at £180,000 a month, independent of how many customers sign up that week. A managed-onboarding team adds £55,000 a month, a cost the founders chose deliberately to keep churn low.

Reading the block top to bottom changes the questions. The fixed engineering line means the model needs🫀NeedUserA user need, pain, desire, or constraintView reference → scale to work: at 500 seats the company loses money, at 5,000 it is comfortably ahead, because the £180,000 spreads thinner per customer with every signup. That is economies of scale made concrete. The onboarding line is the value-driven choice, a cost accepted because retention pays for it. When the founders later move onboarding into self-service tooling, they are not cutting a line item at random; they are shifting the model one notch towards cost-driven, and the canvas shows exactly which block moved.

Cost Structure vs. its neighbours

• Revenue Stream is the income side of the same ledger. Cost Structure is what the model spends to earn it. A model can have a healthy revenue stream and an unsustainable cost structure at the same time, which is how well-funded companies fail; comparing the two blocks is the first test of whether a model holds together.
• Key Resource and Key Activity are the causes; Cost Structure is the effect. The canvas reads these blocks as the drivers of cost, which is why this entry connects to them directly. You do not reduce a cost structure by editing it; you reduce it by changing the resources you hold or the activities you run.
• Unit Economics💹Unit EconomicsBusiness ModelPer-unit economic metrics (CAC, LTV, etc.)View reference → zooms into a single customer or transaction and asks whether that one unit is profitable after its share of variable and allocated fixed cost. Cost Structure is the whole-model view that unit economics decomposes. One is the aggregate picture; the other is the per-customer slice, and a model needs both to know whether scale helps or hurts.

For subscription businesses, Skok's SaaS Metrics🔢MetricStrategyA unified metric that measures progress, health, or behaviour across the productView reference → 2.0 reframes what the per-customer slice has to measure. Acquisition cost is paid upfront, but the revenue arrives in small monthly instalments, so the test is not whether a single transaction is profitable but whether lifetime value recovers acquisition cost — Skok sets the bar with two numbers, an LTV-to-CAC ratio and a payback period (months to recover CAC), and a model has to clear both. On that test a deliberate onboarding line, like the managed-onboarding spend in the example above, is not waste: it is a wager that lower churn lengthens customer lifetime enough to raise LTV. Skok also notes that faster growth deepens the loss before it closes, because each new cohort👥CohortGrowthA group of users sharing a common characteristicView reference →'s upfront cost lands before its revenue does — so a model that looks cost-driven at scale can still run deep losses while expanding, which Cost Structure alone does not show.

In the graph

In the Unified Product Graph, Cost Structure lives in the business-model region, attached to its model through business_model_costs_via_cost_structureBusiness Modelcosts viaCost Structurehierarchy. Its causal neighbours are wired explicitly: cost_structure_driven_by_key_activityCost Structuredriven byKey Activitycausal and cost_structure_driven_by_key_resourceCost Structuredriven byKey Resourcecausal encode Osterwalder and Pigneur's claim that costs flow from what a business does and what it holds. The block also carries cost_structure_measured_by_metricCost Structuremeasured byMetricsemantic, which ties the abstract category to the numbers that govern it, such as gross margin or cost per acquisition. The structure matters because it makes cost queryable as an effect: change a key resource and the graph can trace which cost lines should move, a link that otherwise lives only in a founder's head.