Business model diagram, tested.
Also called: Business model canvas · Commercial logic map · One-page business model · Model-on-a-page
Mapping the whole commercial logic on one page so customer, value, channel, revenue and costs are visible together, and every box is testable.
One page that shows who buys, what they get, how it reaches them, what it earns and what it costs. A box without evidence is a guess wearing a label. The job here is to mark which boxes are proven and which are still wishful thinking.
What a business model diagram is
A business model diagram is the whole commercial logic of the product drawn on a single page: who the customer is, what they get, how it reaches them, what it earns, what it costs, and who you depend on to make it. The Business Model Canvas is the common form. Done well, it lets anyone read the economics of the idea in two minutes without a spreadsheet.
Here is where I get wary. A canvas is the easiest document in product development to make look finished. Nine tidy boxes, a confident font, and the thing reads like a plan even when every box is an untested assumption. The pretty diagram is not the danger; the danger is mistaking the diagram for the evidence. A filled box and a proven box look identical on the page, and only one of them is true.
So the discipline is not drawing the canvas. It is interrogating it. For each box I ask one blunt question: what do we actually know, and how do we know it? A price band is not proven because it feels right; it is proven when someone has paid it, or refused to. A channel is not proven because it exists; it is proven when it has delivered a customer at a cost you can afford. Until then the box is a hypothesis, and it should be labelled as one.
Why this gets faked
- The format rewards confidence. The canvas template hands you nine boxes and a quiet instruction to fill them all. Empty boxes feel like failure, so people write something, and “something” hardens into “fact” by the next meeting.
- Internal logic masquerades as evidence. “The maths works” is a statement about the diagram, not about the market. A model can be perfectly consistent and completely wrong, and consistency is far easier to achieve than truth.
- Nobody marks the unknowns. A canvas with no flagged assumptions is not a strong canvas. It is a canvas that has stopped asking questions, which is the most dangerous state it can be in.
A worked business model canvas
Here is the proofing box’s model on one page, the one we ran through in the pilot, so you can see the shape of a good answer rather than a generic template. Read it as claims to be tested, not facts already won.
Five boxes, one page, the whole logic legible at a glance. Useful, but only half the work. The other half is admitting which of those boxes had any evidence behind them and which were still hope.
- Nine boxes filled, none flagged as a guess.
- “£149 holds a 30% margin” with no costed BOM behind it.
- “DTC channel” named, never tested for cost-per-customer.
- Partners listed who have never quoted a real run.
- Every box tagged: proven, partial, or unknown.
- Margin checked against a real Stoke-on-Trent ceramic quote.
- Channel tested with a small pre-order to read demand.
- The riskiest box named, with the next test against it.
The left-hand canvas and the right-hand canvas can hold the exact same words. The difference is invisible on the page and total in reality: one has been interrogated, the other has only been drawn.
How it fits the bigger picture
Business model diagram is activity 04.10.08 in the framework: the eighth and last activity of Stage 04 Evaluate. It builds on the MVP scope and the market and competitor work before it, gathering the commercial logic into one view. It closes Evaluate and leads into Stage 05 Define, where the design brief turns that proven model into something to build.
What it can do
It puts the entire commercial logic on one page, so a gap or contradiction between, say, the price and the cost base becomes obvious instead of buried. Used as a hypothesis map, it tells the team which assumption to go and test next, and which the whole model rests on.
What it can’t do
It can’t prove anything by itself. A canvas is a record of beliefs, not evidence; filling a box does not make the box true. The proving happens out in the market, in pricing tests, supplier quotes and pre-orders. The diagram only tells you where to point that work, and where you are still guessing.
See the full 10-stage process →
Try it yourself
Draw the five boxes for your own product: customer, value proposition, channel, revenue and price, costs and partners. Fill each in one short line. Then do the part most people skip: against every box write proven, partial, or guess, and one line on how you know. Count the guesses. The box you most need to be true and least have evidence for is your next test.
Want a structured first pass? Start the Free Sprint → and the GPT will help you draft the model and flag the weak boxes.
Your business-model checklist
Project notes: the box that hadn’t earned its ink
▸ From the notebook · optional reading
Dan and Anna’s canvas looked finished in Stockport. Tagging every box proven, partial or guess showed two of the five were pure hope.
3 min read · click to open
Dan brought the canvas to the table already drawn, and it was a handsome thing. Five boxes, clean lines, the £149 price sitting confidently in the revenue box. “That’s the model,” he said. It looked done. That is exactly what worried me.
So I did the unglamorous bit. I drew a second column beside each box and asked one question: proven, partial, or guess? And one line on how we knew.
What the tagging exposed
Customer. Partial. We had the Sourdough School audience and real conversations, but no one had handed over money yet.
Value proposition. Proven, near enough. The no-app, hold-the-temperature promise had come straight out of bakers’ own words in the persona work.
Revenue and price. Guess. The £149 felt right. Nobody had been asked to pay it. The 30% margin rested on a bill of materials we had estimated, not quoted.
Costs and partners. Guess. We had names for a Stoke-on-Trent ceramic maker and a Manchester PCB assembler. Neither had quoted a real production run.
Two of five boxes were pure hope, and they happened to be the two the whole business stood on. The pretty diagram had hidden that completely. I asked Dan how he would feel if the ceramic quote came back at double our estimate. He went quiet.
What we did about it
We stopped admiring the canvas and went and tested the two guesses. The ceramic maker quoted; the BOM landed inside the £38–55 band, just. A small pre-order to a slice of the Sourdough School list read the £149 price against real intent rather than polite enthusiasm. Only then did those two boxes earn the ink they were already drawn in.
The canvas did not change one word in the process. What changed was that we now knew which words were true. That is the entire difference between a diagram and a business model.
— Evaluate stage, project notes, 2026
— Next stage → Stage 05 · Define