The Workshop: Business Model Canvas

Nine boxes on one page. The Business Model Canvas shows where revenue comes from, what it costs to serve a customer, and which assumptions hold the whole thing together: “does this business actually work?” you can read in five minutes. Worked example: Does This Actually Work?.

Business Model Canvas

The Business Model Canvas lays out how a business creates, delivers, and captures value on a single page of nine boxes, filled in customer-first order, so the team can see whether the whole thing holds together and where the most dangerous assumptions live. Universally called the Business Model Canvas or just the BMC. Sometimes confused with a business plan: a plan is a long document assuming the model works; the Canvas is a one-page hypothesis about whether it will. Invented by Alexander Osterwalder as part of his PhD, published with Yves Pigneur in Business Model Generation in 2010, and now one of the most widely used strategic tools in any discipline.

Not to be confused with Maurya’s Lean Canvas (Ash Maurya, Lean Startup author, who introduced Lean Canvas in Running Lean (2012)), which replaces four boxes (Key Partners, Key Activities, Key Resources, Customer Relationships) with Problem, Solution, Key Metrics, and Unfair Advantage. Lean Canvas is for early-stage problem validation; the question is “do we understand the problem well enough to build anything?” BMC is for model articulation; the question is “does this business actually hold together?” Use Lean Canvas before product-market fit; use BMC once there’s something to articulate.

What’s It For

“Can everyone in the room describe the business model the same way?”

That’s the question a facilitator asks at the start of a Canvas session, and it’s the question that makes the room go quiet. Not because nobody in the room knows the business model; they each know a version of it. The founder knows the pricing and the margin assumptions. The developer knows the delivery architecture and roughly what it costs to run. The ops lead knows the wastage rate on unsold perishables. The product person knows the customer acquisition story. Each version is coherent on its own. None of the versions agree with each other, and the business model isn’t any single one of them; it’s the intersection of all of them, which is a thing nobody has ever looked at as a single object.

The gap shows up quietly. A founder whose pricing assumptions haven’t met the ops lead’s wastage numbers discovers, one quiet Sunday, that each box costs $41 to source, pack, and deliver, and they’ve been selling them for $35. Every new subscriber is losing the business money. The faster they grow, the faster they go bankrupt. Nobody did anything wrong. Each person was right about their own box. The failure was that the boxes were never put on one page where somebody had to read them out loud next to each other.

The Business Model Canvas exists to be that one page. Nine boxes, filled in customer-first order, read out loud in pairs at the end so the arithmetic and the logic both have to survive contact with the rest of the model. You can’t admire the value proposition without seeing what it costs to deliver. You can’t celebrate the pricing without seeing the operational burden. You can’t forget about customer acquisition because there’s a box for it on the same page as the revenue streams it feeds.

Reach for it when:

You’re starting a new business or a major new product line
You need to explain the business model to investors, new hires, a board, or yourself
You suspect parts of the business model don’t fit together: the value proposition doesn’t match the revenue model, or the costs don’t support the pricing
You’re comparing two different business model options and need a side-by-side view
An existing business is drifting and you want to diagnose which part of the model has changed

What It’s Not For

Skip it when:

The business model is established, well-understood, and not in question. You’d be documenting, not discovering.
You’re planning features or sprints. The Canvas is strategic, not tactical.
You don’t have anyone in the room who understands the economics (revenue, costs, margins). The Canvas will have holes exactly where it needs to be sharpest.

Stop a session that’s already started if:

Four or more of the nine boxes are pure guesses, making the Canvas mostly fiction; better to pause and do research first
The founder refuses to engage with contradictions surfaced during review
The room is missing the person who owns the cost structure or the pricing, and multiple boxes depend on them

Stopping and fixing the inputs is not failure. Producing a Canvas that papers over an incoherent business is.

Definitions & Background

The nine boxes of the Canvas, with the role each one plays:

Customer Segments: who you’re serving. Specific groups of people or organisations.
Value Propositions: what value you deliver to each segment. The benefit as the customer experiences it, not the feature you build.
Channels: how you reach and deliver to each segment, across awareness, evaluation, purchase, delivery, and after-sales.
Customer Relationships: how you acquire, retain, and grow each segment. Personal, automated, community, co-creation.
Revenue Streams: what customers pay, how, and how much. Pricing models with numbers attached.
Key Resources: the assets essential to delivering the value propositions. Physical, intellectual, human, financial.
Key Activities: the most important things the business must do well. Essential and distinctive, not every task.
Key Partners: who you depend on. Suppliers, partners, services who could break you if they disappeared.
Cost Structure: the most significant costs in the model. Fixed, variable, one-time.

Boxes are filled in customer-first order, not left-to-right. Start with the customer, trace the value outward (segments → propositions → channels → relationships → revenue), then trace the economics backward (activities → resources → partners → costs). Starting with costs produces a defensive Canvas. Starting with customers produces a strategic one.

The reading-aloud ritual. At the end, pairs of boxes are read out loud next to each other: Revenue Streams next to Cost Structure, Value Propositions next to Customer Relationships, Customer Segments next to Channels. The arithmetic and the logic both have to survive contact with the rest of the model. Most Canvas sessions produce at least one coherence break in this phase. The value of the session is finding it.

Inputs

An idea of the business or product concrete enough to test. “A weekly subscription produce box” is enough. “Something with food, maybe” is not.
People who can speak to different parts of the business. No single participant will know all nine boxes, but collectively the room should. Customer-facing voices, economic voices, operational voices.
A wall or large surface with the nine-box Canvas drawn on it (printed, taped up, or projected), sticky notes, and pens. Roughly two hours, uninterrupted.
Numbers, even rough ones, for pricing and the major costs. The Canvas can survive estimates labelled as estimates; it can’t survive nine boxes of vibes.

If the team can’t yet articulate the value proposition or the customer, run JTBD first to clarify which job the customer is hiring the product to do. JTBD output feeds directly into the Value Propositions box.

Outputs

What lands on the Canvas at the end:

A populated Canvas: nine boxes with sticky notes, photographed from directly in front and as close-ups of each box, with the notes readable.
A digital transcription: the Canvas captured in Miro, Mural, Figma, or a slide template, with every sticky note carried across.
A list of contradictions found during the review phase, each as its own line item: “Revenue $35 vs variable cost $41, losing $6 per sale”, “Personal-connection value vs automated relationship”, etc.
A list of empty or shaky boxes: the ones the room couldn’t fill confidently. These are findings, not failures.
An explicit list of the riskiest assumptions baked into the model, usually concentrated in Revenue Streams, Cost Structure, and Customer Segments.

These outputs feed straight into:

Assumption Mapping is the natural follow-up. A Canvas is nine boxes of beliefs wearing a trench coat; Assumption Mapping opens the coat. Run it specifically on Revenue Streams and Cost Structure, where incorrect assumptions are fatal.
Impact Mapping. The Canvas sets the strategy; Impact Mapping picks the deliverables to execute against it. Canvas first for a new business; Impact Mapping first for an existing business with a clear goal.
User Story Mapping. Once the Canvas is coherent, Story Mapping turns the value proposition into a user journey and a release plan.
Wardley Mapping. The Canvas shows what the business is; Wardley Mapping shows where its components sit in the evolution of the market and therefore how they should be treated strategically. Canvas answers “what”; Wardley answers “where.”
Event Storming. Once the Canvas is agreed, Event Storming maps the processes the business will actually run to deliver on it. Canvas sets the shape; Event Storming maps the operations.

Who’s Needed

Four to six people, around two hours:

Facilitator. Holds the box order, keeps the conversation moving, and catches when a feature has been smuggled into the Value Propositions box.
Founder or business owner. Mandatory. They’re the only person who knows (or at least believes they know) the economics, the pricing, the costs, and the margins. Without them, the Canvas will have the most important boxes filled in with guesses.
Product person. They’ll anchor the value propositions and the channels, and they’ll translate between the founder’s business framing and the team’s delivery framing.
Customer-facing people. Whoever talks to actual customers: sales, support, marketing, account managers, operations staff who handle complaints. They will contradict the optimistic assumptions in the room, which is exactly why they’re there.
Developers. One or two. They need to understand the model they’re building for. They will also catch the technical assumptions baked into the Key Resources and Key Activities boxes that nobody else will notice.
Operations / SRE. For any business where operations are a non-trivial cost or a differentiator (which is most of them) ops is a first-class participant. The Cost Structure box is often where ops has the most to say, and what they say is often unwelcome but essential.

The Canvas works by conversation between perspectives, and the conversation collapses above six. Below four, you don’t have enough perspectives to challenge each other.

Who to leave out:

Investors and board members. They see the Canvas as an output, not during the conversation. Their presence changes what the team will say out loud.
Large stakeholder groups. If ten people need to shape the model, run a pre-session to agree the goal and come to the Canvas with the group down to six.
Pure feature-thinkers. Someone who can only discuss what to build, not why or for whom or at what margin, will turn the Value Propositions box into a feature list and the whole Canvas drifts.

How To Run It

Phase	Box	Duration	Key question
1	Customer Segments	10 min	“Who are we serving?”
2	Value Propositions	15 min	“What value do we deliver?”
3	Channels	10 min	“How do we reach and deliver?”
4	Customer Relationships	10 min	“How do we acquire, retain, and grow?”
5	Revenue Streams	10 min	“What do they pay, and how?”
6	Key Activities	10 min	“What must we actually do?”
7	Key Resources	10 min	“What do we need to deliver this?”
8	Key Partners	10 min	“Who do we depend on?”
9	Cost Structure	10 min	“What does it all cost?”
10	Review for coherence	15 min	“Does the maths work?”
Total		~2 hours

Order matters. Start with Customer Segments because every other box is defined in terms of the customer. End with Cost Structure because by the time you get there, you know what you’re doing, for whom, how, and how it’s delivered. Only then can you add up what it costs.

The Canvas is a round-the-room conversation moderated by the facilitator, with notes going up in the current box only. Everyone speaks in every box, but the domain expert for the box takes the lead:

Customer Segments: the customer-facing people lead; everyone else pressure-tests specificity.
Value Propositions: the founder and product person lead; everyone else challenges whether the claimed value is actually the value the customer experiences.
Channels and Customer Relationships: marketing, sales, and support lead; the developers listen hard because these boxes define half of what they’ll need to build.
Revenue Streams: the founder leads, with support from anyone who knows the market. Numbers get written down, even rough ones.
Key Resources, Activities, Partners: operations and developers lead; the founder listens hard because this is where their optimism meets operational reality.
Cost Structure: operations and founder together. Developers add the technology costs.
Review: everyone. The reading-it-aloud ritual is where the coherence check happens.

The rhythm is customer outward, then back through to costs, then check the maths.

Phase 1: Customer Segments (10 minutes)

Point at the Customer Segments box and ask:

“Who exactly are we creating value for? I want specifics. Not ‘everyone who eats food’ or ‘health-conscious consumers.’ Specific enough that I could walk down a street and tell you whether the person next to me is one or not.”

Write each segment on a sticky note and place it in the box. Push hard for specificity:

“‘Busy families’ is closer. Which busy families? Dual-income, both parents working full-time, kids at school, lives in a city with limited supermarket access after 7pm? Now we have an actor whose behaviour we can actually influence.”

If there are multiple segments, rank them. One primary, one or two secondary. Businesses rarely serve three primary segments well in the first year.

What to watch for:

Too broad. “People who eat food.” Push for age, geography, behaviour, pain point, or life stage.
Too many segments. More than three or four is a startup trying to be everything. Pick the one or two that matter most and park the rest.
Confusing users with customers. The person who uses the product and the person who pays may be different. If a company buys boxes for employees, the company is the customer and the employee is the user. Capture both and note which one pays.
The absent customer. Nobody in the room knows the segment concretely because they’ve never spoken to one. That’s a finding; note it, because it will shape which assumptions you test later.

Phase 2: Value Propositions (15 minutes)

For each customer segment, ask:

“What value are we delivering to this segment specifically? What problem are we solving, what pain are we relieving, what job are we helping them get done (Jobs to be Done: the framing that customers hire products to get a specific job done; see JTBD workshop for the deeper version)? I want the benefit as the customer would describe it, not the feature we’d describe.”

Write value propositions on sticky notes in the box and connect them, visually or by proximity, to the segment they serve.

Value propositions come in several flavours and a good Canvas usually has a mix:

Functional: “Fresh produce at the door every Wednesday without having to plan for it”
Problem-solving: “No more panicked supermarket trip on a Tuesday night”
Emotional: “Feel good about supporting local farms”
Economic: “Better value than buying organic at the supermarket, including the time saved”

What to watch for:

Features disguised as value. “We have a mobile app” is a feature. “Manage your subscription in thirty seconds from your phone” is a value proposition. Push for the benefit, not the mechanism.
Value that doesn’t match segment. If the segment is “busy professionals” and the value proposition is “learn about seasonal farming,” something is off. Challenge it: “Would a busy professional sign up for this reason?”
Founder passion masquerading as value. The founder may love supporting small farms; the customer may just want fresh produce at their door. Both can be true, but the Canvas should reflect the customer’s experienced value, not the founder’s internal motivation.
Too many value propositions per segment. If a segment has eight value propositions, the team doesn’t know which one is actually the reason the customer buys. Rank them and put the top two forward.

Phase 3: Channels (10 minutes)

Ask:

“How do we reach our customers? How do they hear about us, how do they decide to try us, how do we actually deliver the value to them, and how do we support them afterwards?”

Channels include awareness, evaluation, purchase, delivery, and after-sales. A good Channels box covers all five phases, not just the sexy acquisition ones.

What to watch for:

Only digital channels. For a physical product like a produce box, the delivery channel (courier, pick-up, post) is critical and often the biggest operational constraint. Don’t forget it.
Missing acquisition. The team knows how to deliver but has no plan for how customers will find them. That’s a gap worth flagging loudly.
Unreal channels. “We’ll go viral on TikTok” is not a channel strategy, it’s a wish. Push: “What specifically will we do on TikTok? Who runs the account? How do we measure whether it works?”
Every channel is owned by the founder. That’s a scaling ceiling. Worth noting now, even if you don’t solve it.

Phase 4: Customer Relationships (10 minutes)

Ask:

“What kind of relationship do we maintain with each segment? How do we acquire them, keep them, and grow what they spend with us?”

Relationships come in flavours: personal (dedicated account manager, farmer liaison), automated (emails, notifications, self-service), community (forums, social media groups, events), co-creation (customers help pick produce, vote on weekly boxes).

What to watch for:

Relationship / value proposition mismatch. If the value is personal connection to local farms but the relationship is entirely automated, something doesn’t fit. The customer signed up for connection and is getting a chatbot.
No retention strategy. Acquiring subscribers is expensive. How do you keep them? If nobody in the room has an answer, that’s a high-impact assumption to flag.
Every customer gets the same relationship. Different segments often need different relationships. A family subscriber and a corporate gift-giver behave differently and need to be managed differently.

Phase 5: Revenue Streams (10 minutes)

Ask:

“What exactly are customers paying for, how do they pay, and how much? I want numbers, even if they’re rough.”

Pricing models include: subscription fees (weekly, monthly, quarterly), per-box pricing with different tiers, add-ons, gift subscriptions, one-off purchases, referral credits.

Write each revenue stream as a sticky note with the price attached. “$35 per box, weekly” not “subscription fee.”

What to watch for:

Vague pricing. “They’ll pay a fair price” is not a revenue stream. Push for numbers: “If we had to set a price today, what would it be?”
Only one revenue stream. Not necessarily wrong, but fragile. Are there adjacent revenue opportunities (add-ons, gifts, upgrades) the team hasn’t considered? At least note them.
Pricing that ignores willingness to pay. “We need $50 per box to cover costs.” That’s a cost-plus position, not a market-led one. Note it; you’ll return to it in the coherence check.
Revenue shapes, not just revenue amounts. “$35 per box, weekly” is different from “$140 per month, billed on the first,” which is different again from “$1600 per year with a renewal window.” The shape of the revenue determines the shape of the cost structure you need to cover, and which box the failure mode hides in.

Phase 6: Key Activities (10 minutes)

Ask:

“What are the most important things we must do to make this business work? Not every task; the essential, distinctive activities.”

Activities might include: sourcing produce from farms, curating and packing boxes, operating delivery logistics, managing the subscriber platform, running customer acquisition marketing, handling support, navigating food safety regulations.

What to watch for:

Listing every task in the business. Key activities are essential AND distinctive. “Payroll” is an activity but not a key one unless payroll is your business.
No mention of the hard things. The activities that are difficult AND essential are the ones that matter most. If sourcing seasonal produce at consistent quality is the hardest part of the business, it should be prominent in this box.
Forgetting acquisition as an activity. Teams treat sales and marketing as “things that happen” rather than activities the business must do well. If customer acquisition is hard, it belongs here.

Phase 7: Key Resources (10 minutes)

Ask:

“What do we need in order to deliver the value propositions? What assets are essential to this business model? Physical, intellectual, human, financial.”

Resources include: physical (warehouse, refrigerated transport, packing equipment), intellectual (software, algorithms, brand, data, supplier relationships), human (team, expertise, farmer relationships), financial (capital, credit lines, working capital for perishable inventory).

What to watch for:

Forgetting people. Teams list technology and forget the farm relationships lead, the customer support agent, the on-call engineer, the person who drives the van at 5am. People are resources.
Aspirational resources. Don’t list what you wish you had; list what you actually need to make this work, and note which of those you don’t yet have.
Missing the non-obvious. “Refrigerated storage” is obvious. “A supplier network you trust enough to bet perishable inventory on” is less obvious and often more important.

Phase 8: Key Partners (10 minutes)

Ask:

“Who do we depend on to make this work? Suppliers, partners, services we can’t deliver without? Who could break us if they disappeared?”

Partners might include: farms and producers, delivery companies, payment processors, cloud providers, co-marketing partners, regulatory bodies.

What to watch for:

Single points of failure. “Our single farm partner supplies everything.” That’s a risk worth flagging. Same for a single delivery company or a single cloud provider.
Partners assumed but not secured. “We’ll partner with local farms” is an assumption, not a partnership. Is there evidence the farms want to work with you?
Hidden partners. Payment processors, email providers, SMS gateways, the cloud provider. Easy to forget, easy to break the business when they fail or change pricing.

Phase 9: Cost Structure (10 minutes)

Ask:

“What are the most significant costs in this business model? Fixed, variable, one-time. I want enough detail that when we look at the Revenue Streams box next to this one, we can tell whether the maths works.”

Categorise:

Fixed: rent, salaries, software subscriptions, insurance
Variable: produce, packaging, delivery, payment processing fees, wastage
One-time: initial equipment, software development, setup

What to watch for:

Missing costs. Teams forget customer acquisition costs, payment processing fees, wastage (unsold perishables), refunds, returns, support salaries, compliance, insurance.
Cost-per-unit vs fixed. Make sure the team separates variable from fixed. A $35 box with $25 of variable cost and $10,000 of monthly fixed cost is a very different business from one with $15 variable and $30,000 fixed.
The silent cost. The founder’s unpaid labour. At some point this becomes a real cost (a hired replacement); the Canvas should flag it even if it’s not being paid today.

Phase 10: Review for coherence (15 minutes)

Step back from the Canvas. This is the phase where the session earns its cost.

Read each pair of boxes out loud, looking for contradictions:

“Revenue Streams says $35 per box. Cost Structure says variable cost per box is $41. This model loses $6 every time we make a sale. Is that right?”

“Value Propositions says ‘personal connection to farms.’ Customer Relationships says ‘automated self-service.’ Are those consistent?”

“Customer Segments says ‘busy professionals.’ Channels says ‘farmers’ market stall.’ Do busy professionals go to farmers’ markets?”

Most Canvas sessions produce at least one coherence break. The value of the session is finding it.

Once you’ve found the breaks, list them explicitly. Each one becomes an assumption worth testing or a strategic decision worth making. Add a sticky note in the margin of the Canvas for each break, so the photograph captures them.

What to watch for:

The optimism spiral. Every box looks rosy. Force the question: “What’s the weakest part of this Canvas? Which box are we least confident about?”
No contradictions found. Either the team has done excellent work, or they’re avoiding the hard look. Challenge them to read the specific numbers aloud: revenue minus variable cost, for example. The contradictions often hide in the arithmetic.
The empty box. If a box stayed mostly empty, that’s a signal. Either the team doesn’t know (valuable finding) or the model has a gap (also valuable finding). Don’t leave an empty box unflagged.

See Business Model Canvas: Does This Actually Work? for the Greenbox team’s first Canvas session, including the moment the founder does the arithmetic between Revenue Streams and Cost Structure out loud and the room goes very quiet.

What Can Go Wrong

The feature session. The team keeps listing product features in the Value Propositions box. Recovery: “Features go in Key Resources or Key Activities. Value Propositions is what the customer gets, not what we build.” Stop if: The team can’t hold the distinction. The Canvas isn’t the right session yet; they need to finish Impact Mapping or Story Mapping first.

The optimism spiral. Every box looks rosy. Recovery: “What’s the weakest part of this Canvas? Which box are we least confident about? Which assumption, if wrong, kills the business?” Stop if: The team refuses to identify a weak box. They’re not ready to be honest with themselves; the Canvas will be decorative.

Analysis paralysis. Twenty minutes debating whether something is a Key Activity or a Key Resource. Recovery: “It doesn’t matter. Canvas is a thinking tool, not a taxonomy exercise. Best-fit box, move on.” Stop if: The argument happens on a second box. The team is using classification to avoid the real conversation.

The absent economics. Nobody in the room knows the actual costs or pricing. Recovery: Fill those boxes with labelled estimates and flag them explicitly: “These boxes are guesses. They go on the assumption list.” Continue the session. Stop if: Four or more of the nine boxes are pure guesses. The Canvas is then mostly fiction; better to pause and do research first.

The contradiction denial. The facilitator names a contradiction (cost exceeds revenue, channel doesn’t match segment) and the room brushes it off. Recovery: Make the contradiction concrete: “Let’s write the arithmetic on the wall. $35 minus $41 is minus $6 per box. Is that what we believe?” Numbers on the wall are harder to dismiss than numbers in the head. Stop if: The room refuses to engage with the arithmetic. The session has produced its finding even if the team won’t accept it: record the contradiction and end.

The wrong room. Halfway through, you realise the person who knows the cost structure isn’t in the room and nobody in the room can speak to it. Recovery: Flag the box as unfinished, capture it as a to-do for a follow-up. Continue with the boxes the room can actually fill. Stop if: Multiple boxes depend on absent people. Reschedule with the right invite list.

The dominant founder. One person (usually the founder) talks every box, and the Canvas becomes their mental model rather than a shared one. Recovery: Round-robin the next box. “Let’s hear from the ops lead first on this one. Founder, hold your view until we’ve heard from everyone else.” Stop if: The pattern survives a second redirect. The Canvas will reflect one person’s beliefs and won’t deliver shared literacy; better to address the dynamic outside the session.

Next Steps

The session ends; the work begins.

Same day, the facilitator:

Photographs the Canvas from directly in front, and close-ups of each box. Make sure the notes are readable.
Transcribes the Canvas into a digital template (Miro, Mural, Figma, or a simple slide) with every sticky note captured.
Lists the contradictions and empty boxes found during the review phase, each as its own line item.
Sends the transcribed Canvas and the contradiction list to participants and relevant stakeholders.

This week, the founder:

This is where the pattern earns its cost, and the work is mostly the founder’s. The Canvas is worthless if the contradictions aren’t resolved.

Fix the arithmetic. If the revenue and cost numbers don’t work, they have to be made to work: by raising prices, cutting costs, changing the operational model, or abandoning the business. Sitting on a broken model is the most expensive option. The founder owns this call.
Run Assumption Mapping on the shaky boxes. Any box that was filled with guesses, or that was the source of a contradiction, needs its assumptions pulled apart. Book the Assumption Mapping session for the next week.
Test the riskiest beliefs fast. Pricing, willingness to pay, cost per unit, churn rate, and customer acquisition cost are the five numbers that kill businesses quietly. If any of them are guesses, they’re the first things to validate in the real world, not in a spreadsheet.
Walk the Canvas to absent stakeholders. Anyone who should have been in the room but wasn’t gets a walk-through. Their challenges will either strengthen the Canvas or reveal problems the original group missed.
Use the Canvas to say no. Any new feature, initiative, or hire that doesn’t improve a box on the Canvas, or worse, makes a box harder, gets parked. The Canvas is the strategic filter.

Ongoing, the team:

Revisits the Canvas quarterly, or when the business model changes significantly. New segments, new pricing, new partners, new costs: each is a reason to update.
Keeps the photographed Canvas visible where strategic conversations happen. It’s the reference that prevents the slow drift back into feature-thinking.
When someone proposes a new initiative, asks them to point to the box it changes on the Canvas. If they can’t, the initiative is probably cost without coherence.

Variants

Standard BMC (default). Nine boxes, four to six people, around two hours, customer-first order. Output: a populated Canvas, a contradiction list, an assumption list. This is what most teams need, and the rest of this post describes it.

Lean Canvas. Ash Maurya’s variant for early-stage problem validation. Replaces Key Partners, Key Activities, Key Resources, and Customer Relationships with Problem, Solution, Key Metrics, and Unfair Advantage. Reach for it when you’re earlier than BMC, when the question is “do we understand the problem well enough to build anything?” rather than “does this business hold together?” Lean Canvas before product-market fit; BMC once there’s something to articulate.

Comparative Canvas. Fill two Canvases side by side for two business model options – “subscription with weekly delivery” vs “on-demand single-box purchase” – and read each pair of boxes across both Canvases. The contradictions surface faster because the alternative is right next to the option, not a hypothetical. Useful when the team is genuinely undecided between two strategic directions.

Diagnostic Canvas. For an existing business that’s drifting, fill the Canvas as it actually is today, then a second Canvas as the team believed it was a year ago. The deltas, which boxes have quietly changed without anyone noticing, are usually where the drift lives. Pricing held while costs crept up. The original segment quietly shifted. The acquisition channel that worked at launch stopped working but wasn’t replaced.

Remote. A Miro or Mural board with the nine-box template pinned, video call for the conversation. Slightly slower (the rhythm of “write a sticky, place a sticky” is faster in person), but the structure transfers cleanly. Use one shared cursor: only the facilitator places stickies, prompted by the team, to keep the layout legible.

Scaled (multi-business or multi-product). A company with several products or business lines runs one Canvas per line, then a master Canvas for the parent. Tensions between Canvases (shared resources, conflicting segments, channel cannibalisation) become visible at the parent level. Six hours total, ideally split across two days so the team can sleep on the first pass.