How to Build a Business Model Canvas, Block by Block

The business model canvas is a single sheet laid out as nine blocks describing how a business creates, delivers and captures value. The canvas is Alexander Osterwalder and Yves Pigneur's, from Business Model Generation, 2010. Building one is a procedure: fill the blocks in a set order, start from the customer rather than the product, and read the result as a completeness check on the description.

Having a business model and being able to write one down are two different skills, and only the second one is taught here. The sheet almost everybody reaches for is the canvas, so the first job is to say whose sheet it is. The sheet itself carries no order of work, so the second job is to supply one.

Whose tool is this, and why does the name belong with it?

The business model canvas is Alexander Osterwalder and Yves Pigneur's, set out in their book Business Model Generation, published in 2010. Osterwalder's doctoral work of 2004 formalised the building blocks first, and the one sheet layout came afterwards. An attribution like that belongs with every teaching of the canvas and every copy handed to somebody.

The attribution matters more than it looks. A framework taught as though it were common knowledge has quietly taken somebody's work and dropped the receipt. Nobody sets out to do it. The dropping happens because the tool is genuinely useful, so it spreads, and by the tenth retelling the name has fallen off the front of it. The same debt sits on Porter's five forces of 1980 and Porter's value chain of 1985, both of which get drawn on whiteboards every day by people who could not name the man who built them. Naming the author costs one sentence and it costs the reader nothing at all.

There is a working benefit too, not just an ethical one. A canvas with a known author also has a known scope: it was built to do a specific job, in a specific way, and there are questions it was never designed to answer. A tool without a name feels like a law of nature. A tool with a name and a date feels like what it is, one person's good idea about how to lay out a description.

What are the nine blocks, and what does each one hold?

The list of nine is the easy part of the canvas and it is available anywhere, so it is worth taking once, quickly, leaving the effort for the two things that are genuinely hard: the order and the arithmetic. Each block is one question, and the answer is a phrase rather than a paragraph. A single sheet only works if a reader can take it in at a glance.

One feature of that table matters before anything else is done with it. Seven of the nine blocks take words as their answer. Two of them, revenue streams and cost structure, take numbers. The asymmetry between words and numbers matters a great deal later on, and it stays completely invisible while the sheet is being filled in.

In what order are they actually filled in?

Fill the five customer side blocks first, in the order customer segments, value propositions, channels, customer relationships, revenue streams, and only then turn to the four firm side blocks in the order key resources, key activities, key partnerships, cost structure. The sequence is nine steps, and on a business already known they run in about forty minutes.

Step one is who the customer is. Step two is the job the business does for that customer. Step three is how the offer travels from the firm to them. Step four is what holds them there after the first purchase. Step five is what they hand over and what each payment buys. The five are read back before anything else is written. Step six is what the firm must have to deliver it. Step seven is what it must do. Step eight is what somebody else supplies. Step nine is what all of that costs. And there it stops.

The tool carries no order, so the order given here is an addition to it rather than part of it. The layout is a grid, and a grid has no first cell. A grid with no first cell leaves every new user free to start wherever the eye lands. Where the eye lands, almost always, is the middle box marked value propositions, the box holding the thing they have built and are proud of.

Starting from the product produces a canvas that describes what the founder wants to build rather than what anyone will pay for. The bending goes like this. The product goes in the middle first. Then comes the customer segment, and because the product is already sitting there the segment written down is the one that would want that product. Then the channel that suits that product. By block nine every one of the nine boxes has been quietly bent to fit the thing in the middle, and the sheet reads beautifully, and it has said nothing the writer did not already believe on sitting down.

Run the other way, block one is a room full of people who have never heard of the firm. Everything after it has to survive contact with them. Same nine boxes, same forty minutes, different document.

A tea stall makes this concrete faster than any business does. Starting with the tea produces a sheet saying it is good tea, brewed properly, at a fair price, and the sheet will be true and useless. Starting with the customer, block one says: men and women coming off the six forty shift at the mill across the road, on foot, carrying nothing, with about four minutes. Now block two cannot say good tea. Block two has to say hot tea handed over in under a minute to somebody who does not want to sit down. Block three has to say a counter facing the road rather than tables at the back. The customer wrote the rest of the sheet.

What has a filled canvas actually established?

A filled canvas is a checklist of completeness, not an analysis: it establishes that the model has been described and that the obvious holes in the description are now visible. A description with its holes made visible is a real service, and it is worth forty minutes of anybody's time. The service is also much smaller than most people think they have bought.

The rest of the argument turns on the distinction between describing and checking, so hold on to it. Describing means putting down what is there. Checking means asking whether what is there hangs together. A canvas does the first job thoroughly and does not attempt the second one at all. Nothing in the tool is broken when it declines to check. The canvas was never built to check.

The strongest thing a finished canvas gives is the list of things that could not be written. The writer sat down believing the business was understood. Forty minutes later there are two boxes with nothing in them, and those two boxes are worth more than the seven that were filled. The two blanks are the parts of that understanding that were running on assumption. The blanks are the service: they convert a vague sense of knowing into a specific list of not knowing.

Why does the canvas never make the two money blocks agree?

Two of the nine blocks carry money, revenue streams and cost structure, and the tool never once requires the two to reconcile. There is no line between them, no arrow, no total, no space for a difference. The two money blocks sit on the same sheet of paper, at opposite ends of the bottom band, and the canvas treats them exactly as it treats the seven word blocks: as things to be described.

The missing line has a consequence once the sheet is finished. A set of revenue streams has been written. A cost structure has been written. Both are honest. Both are complete. At no point did the tool ask anyone to subtract one from the other, and nobody has. The sheet is marked done, and the arithmetic that decides whether the business survives the year has simply not been performed.

The gap is not a defect the authors overlooked. A canvas is a description tool and a subtraction is an assessment, and the moment a description tool starts producing assessments it stops being safe to fill in honestly. The problem is not with the tool. The problem is with what the reader believes a finished sheet has proved.

A household sits beside it. Every rupee that comes into a home in a month, and every commitment going out of it, can be written out accurately and completely. School fees, rent, the loan instalment, food, the phone. Complete on both sides, nothing hidden. And all of that can be done without ever putting the two columns next to each other. A household quietly short every month goes a whole year that way without anyone in it saying the word short. Completeness on both sides. No subtraction anywhere.

What does the canvas of Setu Bazaar look like with every block filled?

Setu Bazaar is an invented marketplace: buyers on one side, small sellers on the other, and the business sitting between them taking a slice of what passes through. Here is its canvas, filled in the nine step order.

Nine blocks, nine answers, nothing left blank. Blocks five and nine are the two that carry numbers, and every component of each is set out below, so the arithmetic can be redone rather than a total taken on trust.

Rs 20,00,00,000/- was written into one block, Rs 22,50,00,000/- was written into another, and Setu Bazaar is short by Rs 2,50,00,000/- for the year, which is Rs 2.5 crore that nothing on the canvas objected to. Every block is filled. The description is complete and accurate. Subtracting is not among the nine things the tool does, so it raised no flag, drew no line and produced no total.

The subtraction takes four seconds: Rs 20,00,00,000/- less Rs 22,50,00,000/- leaves a shortfall of Rs 2,50,00,000/-. Four seconds, and the canvas has been sitting on that answer the whole time without producing it. None of that is a criticism of a good tool. The missing subtraction simply marks where the tool stops and the reader starts.

Does the canvas of Anjani Stationers look any less complete?

Anjani Stationers Private Limited is an invented manufacturer of paper and stationery selling to a few dozen business buyers. Its figures were built elsewhere in these notes and are quoted here rather than recalculated: revenue of Rs 2,70,00,000/-, operating profit of Rs 41,50,000/-, and an operating margin of 15.37 per cent. Its canvas fills in nine blocks exactly like Setu Bazaar's does.

Block one is a few dozen firms buying in bulk on credit terms rather than fifty thousand people paying at checkout. Block three is a sales representative and a delivery van rather than an app. Block six is a factory, a reel store and a bank limit rather than matching software. Block five records what the buyers pay across the year and block nine records what it costs to serve them.

Check that yourself against the two published figures rather than trusting the middle line: Rs 2,70,00,000/- less the operating profit of Rs 41,50,000/- gives the cost structure total of Rs 2,28,50,000/-, and the margin of 15.37 per cent is Rs 41,50,000/- divided by Rs 2,70,00,000/-. The cost figure is derived, so it is shown with both of the figures it came from beside it.

Both canvases have nine blocks filled and both are honest. One describes a year that earned Rs 41,50,000/- and the other a year that lost Rs 2,50,00,000/-, so completeness said nothing about the difference. Neither business is being ranked against the other. One sells to a few dozen firms, the other to fifty thousand people, and one year of arithmetic is nowhere near enough to judge either.

When is the canvas the wrong tool to pick up?

A tool built to describe a business cannot also sit in judgment on it, so the canvas is the wrong thing to reach for the moment the question stops being what is going on here. Three questions in particular get put to a canvas by people who then wonder why the sheet went quiet.

The first is how much to charge. Block five asks what customers pay, and it accepts whatever is written there. Block five has no opinion on whether the number should be higher, and how much to charge is set out under revenue and pricing. The second is whether the field lets anybody earn at all. Whether a field lets anybody earn is a question about the arena rather than the business, and it belongs to Porter's five forces and the rest of industry structure. The third is the price a business would fetch. Putting a price on a business is a separate discipline, and a description tool is the last place to look for one.

There is a fourth misuse worth naming because it is so common. People fill in a canvas and then use its neatness as evidence. The sheet is tidy, so the thinking must be sound. Tidiness is a property of a grid with nine boxes on it. Any nine sentences fit in it neatly, including nine wrong ones.

What happens to the one block that could not be filled?

An empty block is the most useful thing on a finished canvas. The blank is either a genuine gap in the business or a gap in what the reader happens to know, and telling those two apart is the first move worth making. An empty block is where an analyst, a lender or a buyer actually earns the forty minutes back.

A credit officer reading a small manufacturer works exactly this way. The canvas comes back with key partnerships blank. The blank means one of two things. Either the business has no settled supply arrangement, and that is a real risk sitting in front of the officer, or a settled arrangement exists and nobody thought to mention it, leaving a hole in the file. The two readings lead to completely different conversations, and both begin with the same short question: is this empty because there is nothing there, or empty because I was not told?

An equity analyst does the same with the revenue streams block of an unfamiliar business. Blank because the business genuinely does not know which of its lines makes the money, or blank because the disclosure never separated them. The first is a management problem. The second is a reading problem. Both are worth writing down and only one is worth worrying about.

A household version makes the move obvious. Ask someone what they spend in a month and the food and rent come instantly, and then there is a pause. The pause is not usually money that is not being spent. The pause is money being spent that nobody has looked at. The empty box is the interesting one in a household budget for exactly the reason it is interesting on a canvas.

What should actually be done with the sheet next week?

Treat a canvas as a first draft that gets redrawn, not as a document that gets filed, and pair every finished sheet with one line of arithmetic underneath it. The authors always meant it to be worked over. Each iterationOne pass through a job of work, done knowing it will be done again. An iteration earns its place when the next one is better, not when this one is right. tightens the customer segment, and a tighter customer segment tightens everything downstream of it. A canvas redrawn three times is a much better description than a canvas filled in once and framed.

Then comes the line the tool does not have. Under block five and block nine go what came in, what went out, and the difference. The question after that one line is what a single customer brings in against what a single customer costs to serve, a contributionWhat is left of one sale after the costs that only happen because of that sale. Everything the business would still pay if the sale never happened sits outside it. question, set out under unit economics. But the one line comes first. The subtraction takes four seconds, and it is the single most useful addition to a tool that will never supply it.

Anjani Stationers Private Limited and Setu Bazaar are invented.
Educational material. Not advice on any investment, tax, budget or market position.