Business Model Canvas: Nine Boxes

In the enterprise investment class, a diagram was introduced: the Business Model Canvas. It is simply a canvas, with each of its nine boxes holding one element. The line the teacher kept emphasizing was — the nine boxes are not nine optional items, but nine parts of the same machine; only when all of them turn and mesh with each other does the project hold up. A more precise way to put it: the nine boxes must be able to explain each other and be internally consistent; being thin is not the same as being missing (some boxes are naturally meant to be light — for example, a self-serve SaaS keeping customer relationships light is an advantage, not a hole), but anything that cannot be explained will definitely leak. (This is the end-state of the mature stage; not all nine boxes are necessarily complete mid-journey — see “Viewed by stage” below.)

First, let me write down the nine boxes as a record. Each box is accompanied by the question it must answer and one line of my own understanding.

The nine boxes

1. Customer Segments — For whom do we create value? Who are the most important customers? One line: first figure out “who you’re selling to”; don’t build a business “for everyone.”

2. Value Propositions — What problem do we help customers solve, what need do we satisfy? One line: why would others choose you instead of someone else, or instead of doing nothing.

3. Channels — Through which channels do we reach customers? Which is most effective and cheapest? One line: how to make customers know you, be able to buy from you, and use it smoothly.

4. Customer Relationships — What kind of relationship do we want with customers? How do we acquire, retain, and upsell? One line: one-off deal or long-term lock-in — this is a choice that costs something.

5. Revenue Streams — Why are customers willing to pay? How much, how, and what share does each contribute? One line: which routes the money actually comes in through, and which route is the main one.

6. Key Resources — Which resources (people, money, technology, qualifications, channels) are indispensable for delivering the value proposition? One line: the things without which the business cannot be done.

7. Key Activities — What must be done well for the value proposition to land? One line: the few things you are actually busy with every day and cannot outsource.

8. Key Partnerships — Who are the key partners and suppliers? What do we get from them, what do they do on our behalf? One line: which links you don’t do yourself but rely on others for, and why.

9. Cost Structure — What are the most important inherent costs? Which resources or activities burn the most money? One line: where the money mainly goes, and whether it is fixed or variable cost.

How the nine boxes connect

In the middle of the canvas is the Value Proposition (2), connecting the left and right sides. It is not a 3×3 grid, but four boxes on each side with one box sandwiched in the middle:

• The right side is the customer side: Customer Relationships (4), Channels (3), Customer Segments (1), Revenue Streams (5) — answering “how do we make money.”
• The left side is the internal side: Key Partnerships (8), Key Activities (7), Key Resources (6), Cost Structure (9) — answering “how do we get things done, and where does the money go.”

The real use of this framework is not in “filling all nine boxes,” but in checking whether the nine boxes are consistent with each other. A few points I most often check against:

• Whether the customers you chose (1) will actually be moved by your value proposition (2);
• Whether revenue streams (5) can cover cost structure (9) over the long run;
• Whether key activities (7) and key resources (6) are actually supporting the value proposition (2), or each off doing its own thing.

This is what the class meant by “a project can only be done well when all these conditions are satisfied”: it is not checking off the nine boxes one by one, but the nine boxes connecting into a self-consistent whole. The canvas helps you see whether “the business logic holds” — whether any box is empty, or which two boxes are fighting each other; but “whether it is worth investing in” depends on things outside the canvas (team, market, competition, valuation). The canvas is necessary but not sufficient.

Viewed by stage: it doesn’t only hold when “all nine boxes are full”

The opening line about “meshing with each other, being self-consistent” describes the end-state of the mature stage, not a uniform requirement for every stage. Each stage has its questions that must be answered clearly and are the most falsifiable; the remaining boxes may be left as assumptions for now, but must be honestly managed as assumptions, not presented as facts. All boxes must be watched continuously — it’s just that the principal contradiction differs at each stage; it’s not that “once a stage is past, no one looks at it.”

(The transformation phase / second curve is not the fifth stop that comes after this; it often has to be started in parallel or ahead of time before the mature stage. The question to ask: can the old value proposition and key resources be migrated, or must you start from scratch.)

So investment judgment, on the canvas, comes down to two things:

1. For the questions the current stage must answer clearly, is the evidence sufficient?
2. For boxes not yet validated, are they being managed as “assumptions/risks,” or presented as “already established”?

The second one is especially fatal — many projects fail not because boxes are empty, but because assumptions were spoken of as facts.

Business model archetypes and examples

With the same canvas, the combination of boxes you fill in usually falls into one particular archetype. Archetype = which two boxes mesh the tightest. Recognize the archetype, and you’ll know which boxes to focus on.

Fill in the nine boxes with Meituan to see what the “platform type” looks like:

Meituan’s key mesh is that “revenue (commissions) and costs (riders) must expand in sync, and users on each side reinforce one another” — whether the flywheel is spinning is the core of whether it can be invested in.

Variants and easily confused frameworks

The Business Model Canvas has several derived versions, each modified for a different purpose:

There are also several frameworks with similar names but which are actually a different system, easy to confuse (I nearly confused them at first):

• Porter’s Value Chain: not “four on each side with one in the middle,” but a horizontal chain of 5 basic links (inbound logistics → operations → outbound logistics → marketing & sales → after-sales service), overlaid with 4 support activities (firm infrastructure, human resource management, technology development, procurement); it does not have the term “value proposition”. The value chain itself also has a string of variants (smiling curve, value shop / value network, shared value, etc.), to be discussed separately.

• GE–McKinsey nine-box / talent nine-box: it is a 3×3 matrix (industry attractiveness × competitive position; performance × potential), used to rank “whom to keep and whom to cut,” and does not describe how a business runs.

Extension: this canvas and “startup pitch, investor due diligence” are three different perspectives; see Three Perspectives on a Project: Canvas, Pitch, Due Diligence.