Lean Canvas
A one-page startup business model: nine boxes of assumptions, ranked by risk and tested cheapest-first.
On this page
The gist
- →One page, nine boxes: Problem, Segments, UVP, Solution, Channels, Revenue, Costs, Key Metrics, Unfair Advantage. Every box is a guess, not a fact.
- →One canvas per customer segment. Name early adopters narrowly, list max 3 problems and what people do today, cap the solution at 3 features.
- →Close the economics: contribution per unit and payback period turn the canvas from a poster into an argument. Interviewers spend most time here.
- →Finish by ranking boxes by risk: name the assumption that kills the business fastest and a cheap 4-week test for it (interviews, pilot, pre-orders).
The framework at a glance
When to use it
Reach for the Lean Canvas whenever the case asks you to design a business rather than fix one. Typical triggers: "Our client wants to launch a new venture in X, how should they think about it?", any B-plan, startup pitch, incubator or hackathon-style competition, an intrapreneurship or new-product-line brief, "evaluate this startup, would you invest?", or a product-management interview asking you to build something from zero. It is also the right tool when the prompt is deliberately vague and non-quantitative, because the canvas gives you nine concrete places to say something intelligent in ninety seconds. Do not use it for a classic profitability decline (use profitability), a market-entry-for-an-incumbent question (use market-entry), or a pure pricing case. And do not use it for a large established firm redesigning operations; there the discarded blocks (partners, activities, resources) are exactly where the answer lives, so the Business Model Canvas or value-chain is the better fit.
What it is
The Lean Canvas is a one-page business model, created by Ash Maurya in 2010 and popularised through his book Running Lean. He took Alexander Osterwalder's Business Model Canvas, which was built to describe companies that already work, and swapped out four blocks that only matter once you have a business (Key Partners, Key Activities, Key Resources, Customer Relationships) for four blocks that matter when you are still guessing (Problem, Solution, Key Metrics, Unfair Advantage). The result is nine boxes on a single sheet: Problem, Customer Segments, Unique Value Proposition, Solution, Channels, Revenue Streams, Cost Structure, Key Metrics and Unfair Advantage. It is deliberately small. A 40-page business plan takes weeks to write and nobody reads it; a Lean Canvas takes fifteen to twenty minutes and someone can argue with it.
Keep reading ↓Show less ↑
The important idea is not the layout, it is what the boxes contain. Every box is a guess, not a fact. Maurya's whole method treats the canvas as a hypothesis log: you write down what you currently believe, then you rank those beliefs by how badly the business dies if the belief is wrong, and you go test the riskiest one first with the cheapest possible experiment, usually customer conversations before code. A canvas is therefore never finished. You date it, you keep the old versions, and the interesting artefact is the diff between version one and version four, because that diff is what you learned. Maurya also insists the canvas is problem-focused: you do not get to describe a solution until you can describe, in the customer's own words, the pain they have today and the clumsy workaround they currently use for it.
Two structural details are easy to miss and carry a lot of weight. First, several boxes have sub-boxes. Problem has Existing Alternatives, which is how customers solve this today without you (an Excel sheet, a WhatsApp group, a local dukaan, doing nothing) rather than a list of funded competitors. Customer Segments has Early Adopters, the narrow slice who feel the pain badly enough to buy something unfinished. Unique Value Proposition has a High-Level Concept, the shorthand analogy such as "Netflix for X". Second, one canvas covers one customer segment. If you are building a marketplace with buyers and sellers, or a B2B2C play with hospitals and patients, you draw a separate canvas for each side, because their problems, channels and willingness to pay are different and a single blended canvas hides that.
How to apply it, step by step
- 1
Pick one segment and one canvas
Before you write anything, name the single customer segment this canvas is about, and inside it name the early adopters. Not 'Indian SMEs' but 'single-outlet garment retailers in Surat doing 30 to 80 lakh annual revenue who already use Tally'. If the business has two sides, say out loud that you will draw two canvases and start with the harder side. A canvas that tries to serve everyone produces mush in every other box.
- 2
Write the top three problems, plus what people do today
List at most three pains, phrased the way the customer would phrase them, and rank them. Then fill Existing Alternatives: the actual current workaround, which is usually a spreadsheet, a phone call, a local agent, or nothing at all. This sub-box is where interviewers separate serious candidates from pitch-deck candidates, because 'they currently do nothing' is a very different business from 'they currently pay a broker 2 percent'.
- 3
Draft the Unique Value Proposition as one sentence
One line that says why you are different and worth paying attention to, aimed at early adopters, and containing the finished-story benefit rather than the feature. 'Get paid the same day you deliver' beats 'AI-powered invoice reconciliation platform'. Add the high-level concept underneath as a five-word analogy. If you cannot write the UVP without listing features, your problem box is still too vague.
- 4
Cap the solution at three features
For each of the top three problems write the single smallest thing that solves it. Three bullets, no more. The discipline matters because the solution box is the one every student over-fills, and because the smallest solution is what you can actually put in front of a customer in six weeks. Explicitly say what you are leaving out of version one.
- 5
Choose channels the early adopters actually sit in
Path to customers, split into how you reach them now at small scale and what could scale later. Be physical and specific for India: 4,000 kirana owners reached through a distributor's beat route, riders reached at the two busiest gig hubs in a city, doctors reached through a medical rep, students through campus ambassadors. Note whether the channel is free, paid, or owned, and roughly what a customer costs to acquire through it.
- 6
Close the economics: revenue, cost, and one unit
Fill Revenue Streams (what you charge, per what, how often) and Cost Structure (the two or three costs that dominate, fixed versus variable). Then do the thing most candidates skip: compute contribution per single unit, one customer or one vehicle or one order, and state the payback period. This turns the canvas from a poster into an argument and is where a case interviewer will spend most of the time.
- 7
Pick the two or three numbers that tell you it is working
Key Metrics should be a short funnel you would actually watch weekly, not a dashboard. Acquisition, activation, retention, revenue, referral is a serviceable spine (see aarrr). Choose one metric that proves customers want it (repeat usage, retention) over vanity metrics like signups or app downloads, and name the number that would make you kill the idea.
- 8
Name the unfair advantage, then rank every box by risk
Unfair advantage is what cannot be copied or bought: proprietary data, exclusive distribution, a founding team's regulatory relationships, network effects earned over time. 'Passion' and 'first mover' are not advantages; say so if you have none yet. Finally, close by ranking your assumptions: which box, if wrong, kills the business fastest? State that as the riskiest assumption, and name the cheap experiment (20 customer interviews, a pre-order page, one pilot pincode) that tests it in under four weeks.
Worked example
A founding team pitches VoltRide: monthly subscriptions for electric two-wheelers to gig-economy delivery riders (Swiggy, Zomato, Blinkit, Zepto) in Bengaluru, Pune and Hyderabad. One all-inclusive fee covers the scooter, unlimited battery swaps, insurance and servicing. You are asked to lay out the business model and say whether it is worth funding a pilot.
Customer segments and early adopters
Segment: full-time gig delivery riders doing 90 to 120 km a day, six days a week, earning roughly Rs 25,000 to 35,000 a month. Early adopters: riders in the top three order-density hubs of Bengaluru who currently rent a petrol bike from a local garage on a weekly cash arrangement. They are the sharpest wedge because they already accept renting, so we are not also selling them the idea of not owning. Note we are ignoring the platforms themselves for now; a B2B canvas selling fleets to Swiggy would be a separate sheet.
Problem and existing alternatives
Problem 1: petrol eats 20 to 25 percent of daily earnings, about Rs 230 a day at 100 km, 45 kmpl and Rs 105 a litre. Problem 2: a new EV costs Rs 95,000 to 1.2 lakh upfront and riders have thin credit files, so financing is either unavailable or 24 percent plus. Problem 3: a breakdown costs a full day of income and there is no service backup. Existing alternatives: renting an old petrol bike for Rs 350 a day cash from a garage, borrowing a relative's bike, or an NBFC EV loan the rider often cannot get approved for.
Unique value proposition
'Ride an electric scooter for less than what you spend on petrol, with zero downpayment and a swap point every two kilometres.' High-level concept: a Netflix subscription for a delivery rider's vehicle. Note it is framed as the rider's finished story (more take-home pay, no downtime), not as a feature list about batteries.
Solution, capped at three features
One: subscription with zero deposit, cancel with seven days notice. Two: a dense battery-swap network built inside existing petrol pumps and kiranas near the order hubs, so a swap takes under two minutes. Three: a same-day replacement-vehicle guarantee. Explicitly out of version one: an in-house app, rider credit products, and any city beyond Bengaluru.
Channels
Version one is physical and cheap: two field agents standing at the top order-density hubs at shift-change, plus a rider-referral bounty of Rs 1,000 paid after the referred rider completes 30 days. Riders cluster in WhatsApp groups, so referral is the real engine. Scale channel later: a partnership with a platform's rider-onboarding flow, but that is deliberately not the launch channel because partnering from day one with an untested product is a classic Lean Canvas mistake.
Revenue streams and cost structure
Revenue: Rs 5,999 per vehicle per month, all-inclusive. Costs per vehicle-month: vehicle capex Rs 95,000 amortised over 48 months is about Rs 2,000; energy for 2,600 km at roughly Rs 0.35 per km is about Rs 900; servicing and spares Rs 400; insurance Rs 250; swap-network operating cost allocated Rs 500. That is about Rs 4,050, leaving roughly Rs 1,950 contribution per vehicle-month before a 12 percent haircut for idle and defaulting vehicles, so call it Rs 1,700. Against Rs 95,000 of capex per vehicle, payback is around 56 months, which is longer than the asset's useful life. That single calculation is the finding.
Key metrics
Utilisation days per vehicle per month (target 26 or higher), monthly rider churn (target under 6 percent), contribution per vehicle-month, swap-point uptime, and cost to acquire a rider. The kill metric: if utilisation sits below 22 days or churn runs above 10 percent, the unit economics never close and the pilot should stop.
Unfair advantage
Not the scooter, which anyone can buy. The defensible things are exclusive multi-year swap-point agreements at the pumps closest to the highest order-density hubs, and the repayment and usage data on riders who have no formal credit history, which over two years becomes an underwriting asset no new entrant has. Honest note for the pitch: on day one there is no unfair advantage, and pretending otherwise is worse than admitting it.
Riskiest assumption and the test
Rank the risks. Customer risk (will riders pay Rs 5,999 a month) is real but partly answered, since they already pay Rs 350 a day for petrol bikes. The genuinely fatal assumption is economic: the model only works if the effective price rises to about Rs 7,500 or vehicle cost drops to about Rs 60,000, or utilisation is far higher than assumed. Test: a four-week, 40-vehicle pilot in two Bengaluru hubs at two price points, Rs 5,999 and Rs 7,499, measuring signups, 30-day retention and actual daily utilisation. Cost: roughly Rs 40 lakh of vehicles, largely recoverable on resale. That is the recommendation, not a full launch.
Takeaway: The canvas surfaces that VoltRide's binding constraint is not demand, it is asset payback: at Rs 5,999 the contribution never repays the vehicle within its life. So the riskiest assumption to test is price and utilisation, not product desirability. Fund a 40-vehicle, two-price pilot for four weeks, not a three-city launch.
More worked examples
Worked example: Dropbox in 2007, before a single paying user+
Drew Houston has just been accepted to Y Combinator with a file-sync idea, and the loudest objection from experienced people is that the space is already crowded with Xdrive, Mozy, Carbonite, box.net and Microsoft's own sync tools. You are asked to lay out Dropbox's business model on a Lean Canvas as it would have looked before the product shipped, and say which single assumption you would test first. Figures below are approximate and illustrative, drawn from what Houston has since discussed publicly.
Free tier
2 GB
Paid plan (early)
~$99/yr for 50 GB (approx)
Paid-search CAC
~$230-390 per user (approx)
Referral bonus
500 MB to each side
Beta waitlist after demo video
~5k to ~75k overnight (approx)
Customer segments and early adopters
Segment: individual knowledge workers who use two or more computers, typically a work desktop and a home laptop, and who create files all day. Early adopters, named narrowly: technically confident people who already tried a sync tool and abandoned it because it broke, plus students moving coursework between a lab machine and a personal laptop. This matters because the canvas rule is one segment per sheet, and the enterprise IT buyer is a completely different canvas with a different problem (compliance, admin control), a different channel (sales reps) and a different price. Choosing the individual first is what lets every later box stay sharp.
Problem and top existing alternatives
Problem 1: your file is on the other machine and you are not at that machine. Problem 2: emailing a file to yourself creates five versions and you edit the wrong one. Problem 3: an external drive can be forgotten, lost or corrupted, which is literally the founding anecdote of a USB stick left behind on a bus. Existing alternatives is the box that reframes the crowded-market objection: the real incumbent is not Mozy, it is USB drives, self-addressed email and corporate network drives, all of which are free, universal and genuinely awful. That distinction says the market is not crowded, it is badly served, and it changes the whole investment case.
Unique value proposition and a solution capped at three features
UVP: your files, everywhere, in a folder you already know how to use, with nothing to learn and nothing to remember. High-level concept: a magic folder. Solution, three items only: a folder that syncs silently in the background, versioning so an overwrite is recoverable, and a shared folder so two people work on the same files. Deliberately out of version one: mobile, enterprise admin, encryption marketing, storage quotas and any web-app-first workflow. The discipline here explains why Dropbox beat better-funded rivals, since competitors sold a backup product with its own interface while Dropbox sold zero new behaviour.
Channels, and the box where the model nearly broke
The default assumption was paid search. It failed on arithmetic: a keyword-bought user cost somewhere in the range of a few hundred dollars against a roughly $99-a-year product, so the channel could never pay back, and Houston has cited that gap publicly as the moment the plan had to change. The channel that worked was a two-sided referral loop, roughly 500 MB of extra space to the referrer and 500 MB to the friend, capped, which raised signups sharply and eventually accounted for a large share of daily signups. Notice the loop is paid for in marginal storage rather than cash, which is why it works for a software product and would not work for the physical-asset business in the guide's VoltRide example. Lesson for the canvas: Channels is not a list of ideas, it is a constraint that has to reconcile with Revenue and Cost, and if it does not, one of the three boxes is wrong.
Revenue streams and cost structure
Revenue: freemium, roughly 2 GB free forever, paid tiers around $9.99 a month or $99 a year, with a low single-digit to mid single-digit share of free users ever converting. Costs: storage and bandwidth, both falling every year and both roughly variable per gigabyte; engineering, which is fixed and heavy because the sync engine has to be near-perfect; and near-zero marginal cost to serve a free user who stores a few hundred megabytes. Run the unit: if a free user costs a few cents a month in storage and roughly one in twenty-five eventually pays about $99 a year, then 25 free users cost a couple of dollars a month to carry and generate about $99 a year, which pays back easily so long as acquisition is close to free. That single line is why the referral loop was not a growth-hack garnish, it was the business model, and why paid search would have killed the company.
Key metrics and unfair advantage
Metrics worth watching weekly: percentage of signups that install the desktop client and put a file in the folder within 24 hours (activation, the number that really predicts survival), files synced per active user per week, share of signups arriving through referral, free-to-paid conversion by cohort age, and paid churn. Signups alone are a vanity metric here, because a user who never installs the client never experiences the product at all. Unfair advantage on day one: honestly none, and the canvas is more useful when you say so. What accrues later is real, though: a sync engine whose reliability is hard to copy because the hard part is invisible edge cases, shared folders that pull in colleagues and create a mild network effect, and the switching cost of being the folder where a person's working life already lives.
Riskiest assumption and the cheapest possible test
Rank the assumptions. Cost is not fatal, storage prices were falling; competition is not fatal, since the alternative is a USB stick. The two candidates are, first, do ordinary people want this enough to sign up before it exists, and second, can invisible cross-platform sync actually be built reliably. Houston tested the first with what is now the canonical example of a cheap canvas experiment: a roughly three-minute screencast of the product working, posted to a technical audience, with a waitlist behind it. The waitlist went from a few thousand to tens of thousands within about a day, at a cost of one weekend and no infrastructure, and that single result converted the demand box from a guess into evidence and refocused the whole team on the engineering risk, which was the one that remained.
Takeaway: The canvas relocates the argument. The crowded-competitor objection dissolves once Existing Alternatives is filled honestly, because the real incumbent is a USB stick; and the box that actually decides the business is Channels, where a paid-search cost of a few hundred dollars against a roughly $99 product makes the model impossible until acquisition is made almost free. The right first move is therefore not to build more product but to spend a weekend on a demo video plus waitlist to price demand, and to treat referral-driven acquisition as a core model assumption to be tested, not a marketing afterthought.
Worked example: an Indian case-competition brief on at-home diagnostics in tier-2 cities+
You are pitching Nidaan, an at-home blood-sample collection service in Nashik and Aurangabad. The company owns no laboratory: a salaried phlebotomist collects samples from homes each morning and hands them to a partner pathology lab for processing, with reports returned on WhatsApp the same evening. The panel asks you to lay out the model on one page and tell them whether to fund a pilot. All figures are illustrative and stated so you can see the logic, not audited market data.
Average order value
~Rs 950 (illustrative)
Contribution per order
~Rs 250, about 26% (illustrative)
Blended CAC, v1
~Rs 300 (illustrative)
Monthly breakeven
~480 orders, about 17 a day
Kill metric
90-day repeat below 20%
One canvas per side, and the segment named narrowly
This is a two-sided business, so say out loud that there are two canvases: the patient side and the partner-lab side. Start with the patient canvas because it is the harder one. Segment: households in Nashik and Aurangabad with at least one member managing a chronic condition, diabetes, thyroid or hypertension, requiring three to six test panels a year. Early adopters, narrowed properly: adult children aged 25 to 40 who live in the same city as their parents aged 55-plus, are already UPI-native, and currently pay Rs 800 to 1,500 per panel at a neighbourhood lab. They are the wedge because they are the payer, they are reachable digitally, and the patient is not the one who has to be convinced.
Problem and existing alternatives, which is where the case is won
Problem 1: the patient must fast overnight, then travel four to eight kilometres and queue 45 to 90 minutes at the lab, which for a 62-year-old means the test gets postponed. Problem 2: reports arrive as paper or a WhatsApp photo with no history, so the doctor re-orders tests already done six months ago. Problem 3: price opacity, where the same lipid-plus-HbA1c panel is quoted around Rs 700 by a standalone lab and roughly double by a branded chain, with no way to compare. Existing alternatives are not the funded startups: they are the neighbourhood pathology lab that most tier-2 testing already runs through, the lab's own collection boy who visits regulars for cash, and simply skipping the test. That last one, doing nothing, is the biggest competitor, and it tells you the pitch is about adherence and convenience rather than price.
UVP and a solution capped at three features
UVP: your parents' full panel collected at home before 9 am, report on WhatsApp by evening, at what the neighbourhood lab charges. High-level concept: Urban Company for blood tests. Solution, three items: booking entirely over WhatsApp with no app to download, since the payer will install nothing for a service used four times a year; one salaried phlebotomist on an electric scooter running a fixed four-to-six-house morning route inside a three-kilometre cluster; and a single PDF report with a running trend line for repeated markers, shareable to a doctor by link. Explicitly excluded from version one: owning a lab, radiology, insurance tie-ups, a mobile app and any third city.
Channels the early adopters actually sit in
Version one is physical and narrow: sign 25 to 30 general physicians, diabetologists and pharmacies inside two pin codes, with a referral share on the prescription pad, plus paid society-level agents in about 40 apartment complexes where residents already share a WhatsApp group. Both channels put the offer at the exact moment a test is prescribed, which is the only moment demand exists. At roughly Rs 250 to 350 blended acquisition cost per first order, the channel is affordable but not free. The scale channel later is becoming the outsourced collection arm for one regional lab chain and running corporate wellness camps, deliberately not the launch channel, because signing a big partner before the operating model works is the classic canvas mistake.
Revenue streams, cost structure and one unit
Revenue: Rs 950 average order value, collected from the patient at roughly the local lab's list price. Costs per order: processing bought from the partner lab at about 45% off list, so around Rs 520; phlebotomist cost of roughly Rs 25,000 a month all-in over 26 days at eight collections a day, about Rs 120 per order; consumables, vacutainers, gloves and an ice pack, about Rs 40; payment gateway and WhatsApp messaging about Rs 20. Contribution is therefore roughly Rs 250 per order, about 26%. Fixed city cost of a city manager, a sample runner and a small office is about Rs 1.2 lakh a month, so breakeven is roughly 480 orders a month, about 17 a day, which is two fully loaded phlebotomists. Critically, at a CAC near Rs 300 the first order loses money, so the model only works on repeat: a chronic patient testing three to four times a year is worth around Rs 1,750 of contribution over two years, and payback lands on the second order.
Key metrics and an honest unfair advantage
Weekly numbers: collections per phlebotomist per day, target eight and kill below five, since route density is the whole cost structure; share of orders that are repeat, target 45% by month nine; before-9 am on-time rate; report turnaround under eight hours; and contribution per order. Vanity metrics to refuse: WhatsApp enquiries and total registered users. Unfair advantage on day one is nil, and say that. What accrues is threefold: a volume-locked discounted rate card with the largest lab in each cluster, a longitudinal test history for chronic patients that creates both switching cost and a future doctor-facing product, and route density itself, because once you own the dense houses in a pin code the second entrant is stuck at four collections a day and therefore at double your per-order labour cost.
Riskiest assumption and the four-week test
Rank the boxes by what kills you fastest. Consumer demand is not the top risk, since home collection is already proven in metros. The fatal assumption sits in Cost Structure: it presumes a standalone tier-2 lab will hand a third party a 45% discount off list while that third party stands between the lab and its own patient. Run the sensitivity out loud: at only a 25% discount, processing costs about Rs 712, contribution collapses to roughly Rs 60 an order, and the business is dead regardless of how much customers love it. The test is therefore procurement, not marketing: over three to four weeks, walk into 12 labs across the two clusters with a written commitment of 300 samples a month and collect signed rate cards, while running 60 real collections with one phlebotomist and a manual WhatsApp number to measure actual route density and repeat intent. Total cost is roughly Rs 1.5 lakh, versus a full launch.
Takeaway: The binding constraint is supplier pricing, not customer demand: the whole model rests on a roughly 45% discount off the lab's list price, and at 25% the contribution per order falls to about Rs 60 and the business cannot exist. So the recommendation is not to fund a launch but to fund a four-week, roughly Rs 1.5 lakh test that gets 12 signed lab rate cards and 60 live collections, and to proceed only if the discount clears about 40% and 90-day repeat exceeds 20%.
Common pitfalls
- •Filling the canvas as a description instead of a set of hypotheses. If you never say 'this box is a guess and here is how I would test it', you have written a poster, not a business model. Always finish by ranking the boxes by risk.
- •Blending multiple customer segments onto one canvas. A marketplace with buyers and sellers, or a healthtech with hospitals and patients, needs one canvas per side; otherwise the problem, channel and pricing boxes become vague averages that describe nobody.
- •Confusing Existing Alternatives with competitors. Listing three funded startups misses the point. The real alternative is usually an Excel sheet, a phone call to a broker, a WhatsApp group, or simply putting up with the problem, and each of those implies very different willingness to pay.
- •Over-stuffing the Solution box. Ten features signals you have not decided what matters. Three, one per problem, forces prioritisation and makes the MVP concrete.
- •Claiming an unfair advantage that is not one. First-mover advantage, a passionate team, a great UI and 'our technology' are all copyable within a quarter. If you genuinely have none yet, say so and describe how one would accrue over time.
- •Skipping the unit economics. Revenue Streams and Cost Structure are where most canvases go quiet, and it is exactly where a case interviewer will push. One contribution-per-unit number and a payback period turn the whole exercise into an argument.
Interview tips
- •Do not narrate all nine boxes in order like a checklist. Open with the segment, the problem and the UVP in about ninety seconds, then say 'the two boxes that decide this business are pricing and unfair advantage, let me go deep there'. Prioritising boxes out loud is itself a signal of judgement.
- •Always name the early adopter, not just the segment. 'Riders at the three highest order-density hubs in Bengaluru who already rent petrol bikes' beats 'gig workers in India' and instantly changes the quality of the rest of your answer.
- •Bring one number to the Revenue and Cost boxes even if you have to assume it. Contribution per unit and payback period, stated with your assumptions visible, is the single highest-leverage thing you can add to a startup case.
- •End on the riskiest assumption and a specific, cheap, time-boxed test. 'The thing that kills this is X; I would test it with 40 vehicles at two price points over four weeks' is a much stronger close than a summary of what you just said.
- •Use the canvas as scaffolding for B-plan competitions but present the story, not the grid. Judges want problem, why now, why you, and the unit economics; the canvas is how you make sure none of those is missing.
- •Know how it differs from the Business Model Canvas and be ready to say which you would use. Lean Canvas for high-uncertainty new ventures, Business Model Canvas for an established firm where partners, activities and resources are the real levers.
Test yourself
Best video explainers
Capture Your Business Model in 20 Minutes - Lean Canvas
Ash Maurya
The creator himself walking through the whole canvas in one sitting. Start here; it is the closest thing to a primary source on video.
Lean Canvas Intro - Uber example 🚘
Railsware Product Studio
Fills a real canvas box by box using Uber as the running example, which makes the abstract blocks concrete fast.
Understanding the Lean Canvas by Ash Maurya - WF Masterclass
Wadhwani Foundation Africa
A longer teaching masterclass format aimed at first-time founders, useful if the 20-minute version moved too quickly.
Ash Maurya: Business Model Canvas vs. Lean Canvas - Running lean @ Lean Startup Night Frankfurt
AlexBoerger.com
Maurya explains exactly which four blocks he swapped out and why, which is the comparison interviewers most often probe.
Lean Canvas Creator Ash Maurya on Running Lean Third Edition at Lean Product Meetup
Dan Olsen
The most current version of his thinking, including how the canvas connects to traction, business modelling and choosing which risk to attack first.
Go deeper
Why Lean Canvas vs Business Model Canvas?
Ash Maurya (Lean Stack)
The original essay by the creator explaining which four Business Model Canvas blocks he replaced and the reasoning behind each swap. Free to read and the single most authoritative source.
Lean Canvas
Business Model Toolbox
A clean, neutral reference page defining all nine blocks with short prompts for each, plus the guidance to draw a separate canvas per customer segment.
Lean Business Model Canvas template (PDF)
UC San Diego Office of Innovation and Commercialization
A free printable blank canvas from a university startup toolkit. Print it and fill it by hand before a B-plan competition.
Lean Canvas
Tilburg University Entrepreneurship Toolbox
A university entrepreneurship-centre explainer that maps each Lean Canvas block against its Business Model Canvas counterpart, useful for revision.
Lean Canvas vs. Business Model Canvas: A Quick Guide
Miro
A practical side-by-side on when to choose each tool, which is exactly the judgement call an interviewer tests when they ask why you picked this framework.
Now use it on a real case
Reading a framework isn't the same as applying it under pressure. Practise with an AI interviewer that pushes back.
Practise a case free