Business Model Canvas

What this tool does

A business model canvas is a single page describing how an organisation creates value, delivers it and captures some of it back as money, set out in nine building blocks — a format popularised by Alexander Osterwalder and Yves Pigneur.

CM8-252 makes that page behave like a register rather than a poster. Each row is one entry under one block, with the segment it serves, your confidence in it, the risk if it is wrong, an owner and a date; revenue and cost lines carry a figure. The tool draws the shape of the canvas, prices the model and produces the list of assumptions you have not tested — all inside this single file, with no account and no network request.

The nine building blocks

What belongs in each, and the mistake that spoils it.

Customer segments. The distinct groups you serve, with different needs or worth different amounts. Mistake: one segment called "customers"; if they do not all want the same thing, that is two segments pretending to be one.

Value propositions. The job you do for a segment that makes them choose you over the alternative, including doing nothing. Mistake: your own adjectives — quality, service, reliability — instead of what a customer would say.

Channels. How a segment finds you, buys, receives the thing and gets support. Mistake: naming the selling channel and forgetting delivery and after-sales, where the cost and the complaints live.

Customer relationships. What each segment expects of the relationship, and what keeping it takes. Mistake: confusing it with the channel — the salesperson is a channel, "one engineer who knows their parts" is a relationship.

Revenue streams. The distinct ways money arrives: one-off sales, repeat supply, subscription, licensing, service contracts. Mistake: a single row called "sales", hiding that a fifth of revenue is a different business.

Key resources. The assets the model cannot run without — equipment, capital, people, certifications, data, brand. Mistake: listing everything you own rather than the few whose loss would stop the model.

Key activities. What you must do well yourself for the value propositions to exist. Mistake: a list of departments; the activity that differentiates you is usually buried inside one.

Key partners. Who does the parts you chose not to do, and what you depend on them for. Mistake: naming friendly suppliers rather than the dependencies whose disappearance would hurt.

Cost structure. What running the model costs, in the lines that matter, noting which are fixed and which move with volume. Mistake: costing what you enjoy thinking about and omitting what quietly consumes the year.

Right side, left side

The right-hand side — segments, value propositions, channels, relationships — is everything the customer can see. The left-hand side — key resources, key activities, key partners — is the machinery that makes it possible, and none of it exists for its own sake. Underneath sit the two money blocks: revenue streams hang off the right, what the customer pays for the value; cost structure off the left, what the machinery costs to run. So one view answers one question — does what customers pay exceed what it costs to produce?

Linking entries to segments

The most common failure of a canvas is not a missing block. It is nine independent lists: every box filled in, nothing connected to anything, a page describing a company rather than a model. It happens because filling boxes feels like progress and nobody asks the follow-up question, which is "for whom?". Every entry here carries a serves which customer segment field, and the tool counts the ones left blank. A value proposition that serves no named segment is a hobby — something the business is good at, or proud of, that nobody has been shown to want. A channel with no segment is a habit; a resource with no segment is an asset you happen to own. Each may be real, but none is part of a model until it is attached to somebody who pays. The discipline pays back the other way too: once every entry names a segment you can read the canvas one segment at a time and check the chain — this segment, this value proposition, through this channel, paid for by this revenue stream, made possible by these resources and partners — and the gaps surface at once. Where an entry serves everybody, write "all segments". That is an answer; blank is not.

Confidence, and the point of the exercise

A canvas is a set of hypotheses. On a new venture almost every box is a guess; in an established business a surprising number turn out to be things everybody believes and nobody has checked for years. Every entry therefore carries a confidence — assumption, some evidence, or validated — and a judgement of how badly the model breaks if it is wrong. Together they produce the number this tool exists to show you: the count of critical assumptions, untested entries where a wrong answer takes the model down with it.

That count is the headline tile, and the Assumptions to test table is the real output of a canvas session. A canvas is not finished when the nine boxes are full; it is finished when you know which three things you must find out before committing money, and who is finding them out. So the tool refuses a critical assumption with no test recorded: "we assume our only heat treatment supplier will keep taking our volume" is not a plan, "ask for a written capacity commitment and approve two alternatives this quarter" is. What counts as critical is yours to set — high risk only, or medium and above when the money is unrecoverable.

Putting numbers on it

Most canvases are innumerate: the revenue box says "product sales", the cost box says "staff, rent, marketing", and the page is admired for months before anyone notices it does not add up. Two of the nine blocks are money blocks, so this tool insists they carry money — a revenue or cost line needs an amount and a basis, and the other seven refuse one.

Enter each line on the basis you know — a monthly rent, an annual contract value — and the tool puts everything on a common footing before adding it up. Mixing monthly and annual figures in one total is the commonest arithmetic error on a costed canvas.

Now the honest part. The modelled margin is a sanity check on the shape of the model, not an accounting statement. It is the sum of the lines you typed and nothing else — it knows nothing of seasonality, growth, the timing of cash, working capital, tax, depreciation, or any cost you left out. Its job is to catch the model that cannot work. A plausible margin says the shape holds, not that the business is profitable. Accounts do that.

Current and planned canvases in one file

The canvas name field lets one file hold more than one version of the model, and every entry carries a status of current, planned or retired. Draw the model as it is, honestly, then the model you are moving to — the difference is the strategy. One arithmetic warning: if a file holds both, adding up every revenue line counts the same money twice, so the totals cover today's model by default. Widen them only when you have filtered to one version.

Where this sits next to other tools

The canvas describes one business model as it stands. Ansoff is about direction of growth — existing or new products against existing or new markets — and each move implies changes to particular blocks here. BCG is about portfolio: which product lines earn investment and which fund the others.

A 90-minute canvas workshop

Long enough to produce something real. Have someone type straight into the register as you go.

• 0–10 — segments only. Who do we actually serve? Split anything where the answers to "what do they want" differ. Nothing else is discussed until this list is agreed.
• 10–30 — value propositions, segment by segment. What job do we do that makes each segment choose us? Write the customer's words, not yours.
• 30–45 — channels and relationships. How does each segment find us, buy, receive and get looked after? Anything nobody can describe concretely is an assumption.
• 45–60 — the left side. Working back from the value propositions, what resources, activities and partners are required? If nothing requires it, ask why you pay for it.
• 60–75 — the money. Revenue and cost lines with figures. Rough is fine.
• 75–90 — confidence, risk and owners. Rate every row, then read the critical-assumption list aloud and put a name and a date on each.

Saving and printing

Entries and settings are written to this browser's local storage as you type. That storage belongs to one browser on one computer, so treat Export .json as the real save, restored anywhere by Import .json. Reset asks twice, then erases everything. Print Report builds the pack from whatever the filter shows.

FAQ

How many entries should a block have? Two to four. One usually means the block has not been thought about; ten means the entries describe tasks rather than model components.

Can one entry sit in two blocks? Usually it means the entry is written too loosely. "The sales engineer" is a channel; "customers can reach one person who knows their parts" is a relationship; "an experienced technical salesperson" is a key resource. Split it.

Should the cost lines match my accounts? They should be recognisable from them, but need not reconcile. The canvas splits costs by what the model needs, not by the chart of accounts.

What if we have no numbers yet? Enter the revenue and cost lines as assumptions with your best estimate and a basis, and let the test list say which figure to find first.

Accuracy & disclaimer

This tool records and arranges what you enter. It cannot tell whether an entry is in the right block, whether a segment is real, whether a value proposition is one anybody would pay for, or whether an entry marked validated ever was. A canvas built on wishful thinking produces the same nine tidy groups as one built on customer interviews and a costed ledger. The canvas is a description and a set of hypotheses — not a plan, not a budget and not a forecast. The modelled margin is the arithmetic of the lines you entered and omits everything you did not. It is an analysis aid, not advice.