Cost Structure in the Business Model Canvas (With Examples)

Cost Structure is the last block of the Business Model Canvas, and it's where honesty gets enforced. Every promise made elsewhere on the canvas — the relationships you'll maintain, the activities you'll master, the partners you'll pay — lands here as a number.

What belongs in the block

Not your full budget — your defining costs: the handful of lines that determine whether the model works. For Netflix, one line dwarfs everything: $15B+ a year on content. For Uber, it's driver incentives and insurance. For Spotify, roughly two-thirds of every dollar leaves immediately as royalties — a ceiling on margin that no product excellence can lift, which is precisely the kind of truth this block exists to surface.

Fixed vs. variable

Fixed costs stay flat regardless of volume: salaries, rent, that content budget. Variable costs scale with each unit: payment fees, shipping, cloud compute per customer, AI inference per request.

The mix shapes your risk profile. Heavy fixed costs mean losses until scale, then expanding margins. Heavy variable costs mean you survive small but must guard per-unit margin forever. Software used to be the ultimate fixed-cost business; AI products have quietly reintroduced variable economics — every user interaction costs real compute, a shift many 2026-era canvases haven't caught up with.

Cost-driven vs. value-driven models

Cost-driven models compete by being structurally cheaper: lean operations, automation, self-service everywhere. Budget airlines are the canonical case.

Value-driven models accept higher costs to deliver premium value: dedicated service, craftsmanship, brand. Most companies blend both — but your canvas should declare a bias, because it decides arguments. When support quality and cost cutting collide, which wins? A canvas that answers that question in advance is doing its job.

Canvas-level unit economics

You don't need a financial model at this stage. You need one sanity check:

Does serving one more customer cost meaningfully less than that customer pays?

Estimate, roughly: what one customer costs to acquire (from your channels), to serve (variable costs), and what they're worth (from your revenue streams). If the arithmetic fails at the sketch level, it will fail harder with real numbers. Some models — Amazon's retail arm famously — only close at enormous volume; knowing that in advance is the difference between a strategy and a surprise.

Economies of scale and scope

Two forces that improve the block over time: scale (per-unit costs fall as volume grows — bulk purchasing, amortized fixed costs) and scope (shared resources serve multiple lines — Amazon's one logistics network carrying retail, marketplace, and subscription businesses). If your model depends on either, write down how much volume "enough" is.

Common mistakes

• Listing only the obvious. The costs that kill are the ones implied elsewhere on the canvas but never written here: support headcount, content moderation, compliance, churn replacement.


• Ignoring acquisition cost because it feels like marketing's problem. It's the model's problem.


• Precision theater. Five significant digits at the assumption stage is fiction wearing a spreadsheet.

Study the cost sides of all 12 example canvases, then pressure-test your own — the AI coach totals the left side of your canvas against your revenue block and asks the uncomfortable question early.