Business Model Canvas
Summary: A nine-block visual language for describing, designing, and testing a business model, created by Alexander Osterwalder, built on the idea that words alone can’t align a team on what “business model” means, but a shared canvas can.
Sources: BUSINESS MODEL CANVAS INTRODUCTION.md (transcript of a Stanford talk by Alexander Osterwalder, joined by Steve Blank); 64c7c201a7f4a5e05072cbe1_A3-Business Model Canvas-2023.png (the official canvas template, Strategyzer AG)
Last updated: 2026-07-08
Osterwalder’s starting observation: everyone uses the word “business model” constantly, but almost nobody means the same thing by it. Put a room of smart people together and the discussion becomes what he calls “blah blah blah,” a lot of words, no shared understanding (source: BUSINESS MODEL CANVAS INTRODUCTION.md). His answer, developed in his PhD (“The Business Model Ontology”) and later the bestselling book Business Model Generation, was to replace words with a shared visual language: the Business Model Canvas.
The nine building blocks
The canvas has nine blocks (source: 64c7c201a7f4a5e05072cbe1_A3-Business Model Canvas-2023.png; BUSINESS MODEL CANVAS INTRODUCTION.md):
- Customer Segments. Who you’re creating value for.
- Value Propositions. The products, services, and features that solve a problem or satisfy a need.
- Channels. How you reach and deliver to customers.
- Customer Relationships. The type of relationship each segment expects.
- Revenue Streams. How the business earns money from each segment.
- Key Resources. The assets required to make the model work: patents, brand, capital, people.
- Key Activities. The most important things the business must do.
- Key Partnerships. The network of suppliers and partners that make the model work.
- Cost Structure. The costs incurred to operate the model.
Osterwalder draws a “front stage / back stage” split across the canvas: the right-hand blocks (Customer Segments, Relationships, Channels, Value Propositions) are what the customer sees; the left-hand blocks (Key Partnerships, Activities, Resources, plus Cost Structure) are what makes that front stage possible. He calls it a mirror: read the right side and you can usually infer what the left side needs to contain (source: BUSINESS MODEL CANVAS INTRODUCTION.md).
Worked example: Nespresso
The single illustration used throughout the talk was Nespresso, chosen because it’s a commodities business (coffee) that found an unusual business model. This is entirely Osterwalder’s own teaching example, not connected to any wiki source beyond this talk (source: BUSINESS MODEL CANVAS INTRODUCTION.md).
Swiss households pay 600-800% more for coffee at home than they did before Nespresso, in a commodities business. The story, once mapped onto the canvas: sell the machines cheap and wide (retail, 85% household / 15% business, thin margin, built by contracted manufacturers rather than Nespresso itself), but sell the pods exclusively through Nespresso’s own channels (mail order, call centers, nespresso.com, and flagship stores on streets like the Champs-Élysées and Fifth Avenue). Once a customer owns the machine, only Nespresso’s own pods fit it, so distribution reach stops mattering for the pods; the customer has no choice left. Recurring pod revenue at a very high margin is the entire business. Key resources: patents protecting the pod format from copycats, high-end coffee, an international parent (Nestlé), and brand, expensive to build, but without it “you’re dead” in a consumer business (source: BUSINESS MODEL CANVAS INTRODUCTION.md).
Nespresso’s first business model nearly bankrupted the company: a joint venture with machine manufacturers selling to offices through the manufacturers’ own sales force. Offices weren’t interested, and that sales force didn’t want to sell small machines. The company had already built out inventory before discovering this didn’t work, and only pivoted because the board balked at full warehouses. Osterwalder’s point: they could have tested this far more cheaply before building it (source: BUSINESS MODEL CANVAS INTRODUCTION.md). This is the same instinct behind candles-not-lightbulbs and first-wrong-answer: get something real in front of people before committing the capacity to build it out fully.
Four levels of using the canvas
Osterwalder and Steve Blank (creator of the customer-development methodology and coauthor of The Startup Owner’s Manual) laid out four levels of maturity in using the canvas (source: BUSINESS MODEL CANVAS INTRODUCTION.md):
- Checklist. Fill in the nine boxes. Useful as a starting discipline, but Blank calls it “cognitive murder” when that’s where you stop: information without a story.
- Story / connections. Understand how the blocks connect to each other, and walk someone through the model one block at a time rather than dumping all nine at once. This is where the Nespresso “machines cheap, pods exclusive” logic lives: not any single block, but the relationship between blocks.
- Patterns. Recognize that the same underlying dynamics recur across unrelated businesses (a “double-sided market” like Google, a freemium model, and so on), and can be pulled off the shelf and applied to a new problem, the way design patterns work in architecture and software engineering.
- Evolving over time (Blank’s addition). The canvas isn’t a static document. Snapshot it over time, Blank calls these “scorecards”, and you get a record of what an entrepreneur actually learned and when, rather than the after-the-fact story where “I knew that all along.”
Search versus execution
Blank’s sharpest distinction: searching for a business model and executing one are different jobs, requiring different people and a different tolerance for failure (source: BUSINESS MODEL CANVAS INTRODUCTION.md).
- In execution, the business model is already known. People are hired for known, repeatable work, governed by plans and budgets. Failure here is a failure of personal competence.
- In search, the business model is still a set of untested guesses (some strategy professors call these WAGs, wild-ass gambles). Failure isn’t punished, it’s the mechanism: you go from experiment to experiment, expecting most to fail, and use what you learn to run the next one.
Large companies tend to take their best executors and hand them a search problem, then evaluate them on execution metrics. Blank calls this “career suicide” for the person, and it’s a structural reason big companies struggle to find genuinely new business models even when individual employees are talented.
This maps directly onto other planning ideas already in the wiki. the-7-budgets names the same split in organizational terms: Operating Expense and the already-committed Investment Capital roadmap are the execution budgets, while the mid-year “Refund,” the single highest point of discretion in that model, is where search-stage, shovel-ready ideas actually get funded. my-leadership-principles’s “Practical and evolutionary” principle (“start small, and expand through feedback and realized business value”) and “Good Enough is Good Enough” are the same search-stage instinct, stated as personal habits rather than an organizational chart.
The First Wrong Answer is the same discipline compressed to the scale of a single stuck decision, rather than a whole business model.
Three threads connect them directly:
- Being wrong is the mechanism, not the failure. Blank’s search-stage guesses exist to be tested and mostly discarded. Paul’s First Wrong Answer is offered “almost guaranteed to be wrong in places” on purpose, since that’s exactly what invites the critique that moves a team forward. Both treat a wrong answer as productive, not embarrassing, as long as it’s offered honestly and early.
- A concrete, testable artifact replaces abstract debate. A search-stage business model is “a set of untested hypotheses”; a First Wrong Answer is explicitly “gotten out of one’s head and onto paper as a concrete starting point.” Neither is a claim of correctness. Both are something specific enough for other people to react to, which is the whole point.
- It takes a search-stage tolerance for failure that execution doesn’t train. Blank’s observation about why large companies struggle here, their best people are hired and evaluated for execution, where being wrong is a personal failure, explains why volunteering a First Wrong Answer takes real courage. Someone optimized to avoid being wrong has to unlearn that instinct to do it.