Business Model Canvas Generator

Where the Business Model Canvas came from

The canvas began as research. In his 2004 PhD thesis at the University of Lausanne, Alexander Osterwalder set out to build a shared language for describing how any company makes money. He grouped a business model into four pillars: the product, the customer interface, the infrastructure, and the finances¹. His later paper with Yves Pigneur and Christopher Tucci set out what a business model is and how the idea had been used until then².

The one-page canvas most people know was made popular by Business Model Generation, the book Osterwalder and Pigneur published with Wiley in 2010³. A common claim online is that 470 people "created the canvas." The publisher says something narrower: the book was co-created by 470 canvas practitioners from 45 countries³. The model underneath it came from the thesis.

The Business Model Canvas is designed by Strategyzer AG (strategyzer.com) and licensed under the Creative Commons Attribution-Share Alike 3.0 Unported License⁴. You may copy and adapt it, as long as you credit Strategyzer and share your version under the same license.

Treat every block as a guess to test

A filled canvas looks finished. It is not. Steve Blank, writing in Harvard Business Review in 2013, describes how lean startups use the canvas: on day one, founders have only "a series of untested hypotheses," so they sum them up on a canvas instead of writing a long business plan, then go out and test them with real customers, partners and buyers⁵.

That changes how you read the draft this tool gives you. Every sticky note is a claim that could be wrong. A useful habit:

• Mark what you know and what you assume. A block backed by sales, contracts or customer interviews is evidence. A block you wrote because it sounded right is a guess.
• Find the riskiest guess. Ask which block, if wrong, would sink the whole model. That is usually a customer segment, a value proposition or a revenue stream, not a key resource.
• Write the test next to it. Say what you will do, who you will talk to, and what result would prove the guess wrong.
• Redraw after each test. Date each version. The canvas is a record of what you have learned, not a poster.

Make the nine blocks fit together

The nine blocks are not nine separate lists. The right side of the canvas (customer segments, value propositions, channels, customer relationships, revenue streams) describes value for the customer. The left side (key partners, key activities, key resources, cost structure) describes what it takes to deliver it³. A good canvas reads across: each item on one side should explain an item on the other.

Three checks catch most gaps:

• Every value proposition has a customer. If a promise does not match a named segment, cut it or name who wants it.
• Every channel reaches a segment you listed. A channel that reaches people you do not serve is a cost with no return.
• Every key activity and resource serves a value proposition. Then check that revenue from the segments can pay for the cost structure those activities create.

A loose canvas and a tighter one

This is the QA-testing marketplace shown in the example above, rewritten. The business and its numbers are invented for illustration.

In the tighter version, each block points at the others. The 72-hour promise needs the genre-sorted tester pool, and the price has to cover the payout. The lines that are still guesses say so, which tells you what to test first.

Common mistakes

• One canvas for every customer. Two segments with different needs usually mean two value propositions, and often two models. Draw a canvas for each, or color-code the notes so you can see which items belong together.
• Features as value propositions. "Matching algorithm" is a feature. The value is what it does for the customer, such as a build tested by someone who knows the genre.
• Filling every box to look complete. An empty block is honest. A vague entry like "partnerships" hides the question you still need to answer.
• Drawing it once. A canvas that never changes was never tested.

What the canvas does not answer, and what to use next

The canvas describes a model. It leaves some big questions to other tools:

• Whether customers want the promise. The value proposition block is one box. Osterwalder and his co-authors wrote Value Proposition Design in 2014 to zoom into it, with a separate canvas that matches customer jobs, pains and gains to what you offer⁶. Draft yours with the value proposition generator, then bring the best line back into the canvas.
• What is changing around you. The canvas has no box for the market or rivals. A PESTLE analysis covers outside change, and Porter's Five Forces shows where profit in your industry is under pressure.
• The effect on people and the planet. The canvas tracks money. Alexandre Joyce and Raymond Paquin, writing in the Journal of Cleaner Production in 2016, added an environmental layer and a social layer to it for that reason⁷.
• Which direction to take. The canvas can show two models side by side, but it does not choose. Playing to Win is a good way to make that choice on purpose, and a SWOT analysis helps you weigh it.

For a very early startup, some founders prefer the Lean Canvas, which Ash Maurya adapted from the Business Model Canvas to put problems, solutions and key metrics up front⁸.