Draw Before You Decide

The Core Insight

Xerox priced the 914 photocopier too high to sell when it arrived in 1958. The company leased it instead, at 95 dollars a month with 2,000 free copies, plus five cents for each copy after that. Osterwalder and Pigneur file that move under finance-driven innovation, one of four places a redesign can start.

Business Model Generation, published in 2010, defines a business model as the rationale of how an organization creates, delivers, and captures value. The authors co-created it with 470 members of a business model community, all named as co-authors. Nine blocks on one sheet carry the definition.

Most managers treat alternatives as cheap and the choice between them as the hard work. Boland and Collopy name that a decision attitude. A design attitude runs the other way: producing one outstanding alternative is the hard part, and the choice becomes trivial once it exists.

Without a shared vocabulary a team cannot challenge its own assumptions, because each person argues about a different object. IBM, Ericsson, Deloitte, and Public Works and Government Services of Canada were using the canvas before the book was printed.

The Framework

Nine blocks cover four areas: customers, offer, infrastructure, and financial viability. Each block is a question with a written answer. The right side of the sheet holds value, and the left side holds cost.

Customer Segments asks who you serve and which customers matter most. Groups split when they need a different offer, channel, or relationship, or when their profitability differs. Credit Suisse separates customers with assets up to 100,000 dollars from clients worth more than 500,000 dollars.

Value Propositions asks which customer problem the offer solves. The book lists eleven kinds of value, including newness, performance, customization, design, brand, price, and risk reduction. Rolls-Royce sells the getting-the-job-done kind, charging airlines for each hour a jet engine runs.

Channels asks how each segment wants to be reached, and a channel does five jobs: awareness, evaluation, purchase, delivery, and support after the sale. Customer Relationships asks what a segment expects, across six categories including self-service, communities, and co-creation.

Revenue Streams asks what customers pay for today and how they prefer to pay. Seven mechanisms appear: asset sale, usage fee, subscription, lending, licensing, brokerage fees, and advertising.

Key Resources asks which assets the model needs, sorted into physical, intellectual, human, and financial. Key Activities asks what the company must do. Microsoft develops software, Dell manages a supply chain, and McKinsey solves problems. Key Partnerships asks who supplies what, for scale, for risk reduction, or for access to something you do not own. Cost Structure asks which resources and activities cost the most, and whether the model is cost-driven like Ryanair or value-driven like a luxury hotel.

Three layers sit on top: five patterns, six design techniques, and a five-phase process.

• Mobilize frames the objective and assembles the team.
• Understand means immersion: scan the environment, interview customers and experts, and study what failed before.
• Design means inquiry: brainstorm, prototype, test, and select.
• Implement executes the chosen model and communicates it across the organization.
• Manage keeps scanning and rejuvenates the model before the market forces it.

Key Ideas

Unbundle the Three Businesses Fighting Inside One

Three kinds of business hide inside most companies, each with different economics and culture. Relationship businesses find and keep customers. Product innovation businesses build new products. Infrastructure businesses run platforms for high volume repetitive work. Hagel and Singer named the split in 1999.

Maerki Baumann of Zurich spun its transaction platform into Incore Bank, which now serves other banks and securities dealers, and kept advice for itself. Pictet, the largest Swiss private bank, stayed integrated. The book lists six trade-offs. Selling products to competing banks creates a conflict of interest, and the platform needs scale one bank cannot reach.

Mobile operators ran the same play. Bharti Airtel was among the first, sending its network to Ericsson and Nokia Siemens Networks and its information technology to IBM. Their key asset became the brand and the customer relationships.

The Long Tail Needs Cheap Inventory

Selling less of more works when niche sales aggregate into real revenue, which needs low inventory cost and a strong platform. Three changes made it possible: cheaper production tools, cheaper distribution, and falling search costs. Lulu.com turned the bestseller model over by letting anyone publish, so authors became the customers and books print on order. A title that fails costs the platform nothing.

LEGO Factory opened in 2005 with a digital design tool, so customers built their own kits and ordered them boxed. Volumes stayed low, so LEGO adjusted existing resources rather than its supply chain, and the line stayed a small portion of revenue.

A Platform Lives or Dies on the Subsidy

A multi-sided platform serves two or more interdependent groups, and it is worth nothing to one group when the others stay away. Value grows with users through the network effect, so the fix for the first move is a subsidy on one side. Metro launched in Stockholm in 1995, gave the paper to commuters, drew advertisers, and turned profitable fast. Sony subsidized every PlayStation 3 and never recovered it, because fewer games sold than the company estimated. The operator settles two things first: which side is price sensitive, and whether the other side covers the subsidy.

Google runs three interdependent segments. Advertisers bid for keywords by auction, content owners take a share through AdSense, and free tools grow the surfer base. Osterwalder describes Nintendo before the Wii as losing share and near bankruptcy. The Wii aimed at casual gamers with cheap motion control, cost less to build, and needed no subsidy. The console earned from both sides.

Diamond Multimedia sold the Rio first, and Apple outpaced it with a better business model. The iTunes Music Store arrived in 2003 and made Apple the largest online music retailer. The App Store came in 2008, with every sale routed through Apple at a 30 percent royalty.

Free Is a Financing Decision

A free offer means one substantial segment pays nothing while another part of the model pays for them. Three versions exist: advertising, freemium, and bait and hook. Metro cut editorial spending to what a short commute needs. Controlling news racks at train and bus stations forced rivals into hand distribution at higher cost.

Audience alone does not produce advertising revenue. Facebook had more than 200 million active users in 2009, and its users responded to Facebook advertising less than to traditional web ads.

Freemium runs on two numbers: the average cost of serving a free user, and the rate at which free users convert. Fewer than 10 percent subscribe. Red Hat sells subscriptions instead of per-release licenses, supports each release for seven years, and takes community improvements at no cost.

Skype ran calls peer to peer over user hardware, so it carried little infrastructure of its own. It had more than 400 million registered users and more than 100 billion free calls since 2004, on 2008 revenue of 550 million dollars. More than 90 percent of users stayed free.

King C. Gillette discounted or gave away razor handles in 1904 to create demand for blades, and blocking patents kept cheaper blades from fitting. Insurance runs freemium backward, and REGA funds Swiss air rescue from more than two million patrons who are then exempt from its costs.

Buy the Innovation You Cannot Staff

Open models create value with outside partners in two directions. Outside-in brings external ideas in. Inside-out sends idle ideas and assets out.

A.G. Lafley became chief executive in June 2000 during a share-price slide and rebuilt the culture rather than the research budget. He set a goal of creating 50 percent of innovations with outside partners, at a time when the figure sat closer to 15 percent. Procter and Gamble passed the goal in 2007, and research productivity rose 85 percent on modestly higher spending. Three bridges carried the work: technology entrepreneurs, internet platforms, and retirees.

InnoCentive sells the same trade. Seekers post challenges, solvers attempt them, and prizes range from 5,000 to 1,000,000 dollars. GlaxoSmithKline pools patents for drugs against diseases of the poorest countries, so no single rights holder blocks an advance.

Prototype at Four Scales, Then Run the Scenarios

In the summer of 2000 Matt Fineout of Gehry and Associates tore up two days of work. The team removed about 5,500 square feet from a floor plan. They had shown it can be done, and now needed to decide how.

A prototype is a thinking tool, and interaction with prototypes produces ideas faster than discussion does. The design attitude means exploring crude ideas, discarding them fast, and accepting uncertainty until a direction matures. Jim Glymph of Gehry Partners prices the alternative: freeze an idea too early and you fall in love with it. Four scales run from sketch to evidence.

• A napkin sketch holds the idea, the value proposition, and the main revenue streams.
• An elaborated canvas adds business logic, a market estimate, and fact checking.
• A business case turns the canvas into a spreadsheet of costs and revenues.
• A field test puts the value proposition, channels, and pricing in front of customers.

Scenarios do the same job for futures the company does not control. The pharmaceutical example runs on two drivers: personalized medicine arriving, and the shift from treatment toward prevention. Crossing them produces four futures: business as usual, my.medicine, the healthy patient, and reinventing pharma. Prepare two to four before the workshop, tell each as a story, and build a model for each.

Practical Applications

Draw the current model before proposing a change to it. The authors run it large and on paper, one element per sticky note, a few words each. A note that moves in a second keeps a group willing to change its mind.

Profile one customer before designing the offer. The empathy map, built by a visual thinking company, works on a named person with real demographics. The six prompts cover what she sees, hears, thinks, feels, says, and does, plus her pain and her gain.

Pick the epicenter of the redesign, because a model can be rebuilt from any of four. Amazon Web Services started from the resource side, built on retail infrastructure. Cemex started from the offer, promising poured cement on site within four hours against an industry standard of 48.

Run the Four Actions on every block. Kim and Mauborgne ask which taken-for-granted factors must go, which drop below industry standard, which rise above it, and which get created. Cirque du Soleil cut animals and star performers, added theme and refined music, and combined circus with theater to reach adults instead of families.

Decide integrate or separate before the pilot ends. Nicolas Hayek launched the Swatch inside SMH, and each brand kept autonomy over product and marketing while manufacturing and research stayed central. Nestle put Nespresso in a wholly owned company in 1986, and the separation kept the project alive when 1987 sales sagged. Markides sorts the choice on conflict severity and strategic similarity. Risk to brand and earnings is the third variable, and ING Direct nearly toppled ING in 2008.

Who This Is For

Founders designing a first model get the most, along with operators inside a company whose model is aging. The canvas is a group instrument, and its value appears when four people who disagree point at the same block. Workshop facilitators get a full method, from the empathy map to the scenario process.

Skip it if you want positioning language, a market entry sequence, or a way to test demand. The book describes what a business model contains and how to redesign it, and it stops at that edge.

The canvas is a description tool. A team can fill nine blocks completely and still hold a business nobody pays for, because the sheet has no field for evidence. Winning firms supply nearly every case: Apple, Google, Skype, Nespresso, the Swatch, and the Wii. Those canvases were drawn after the outcome was known, and none of them explains why the same drawing failed elsewhere. The figures date the book to 2009 and 2010, when eBay owned Skype. The closing sections on information technology alignment read as promotion for the authors' own tool.

The Decision

Draw three canvases this week. Start with the model as it runs today, then change one block, then change the revenue mechanism. An afternoon covers all three at napkin scale.

Then run the book's kill and thrill session on each one. Spend twenty minutes on why the model does not work, then twenty minutes on why it flies. Keep whichever canvas survives both halves.

The canvas gives a room one object to argue about, and the arguing is most of the value. Whether a customer pays gets settled elsewhere. That answer sits in the field test, the fourth prototype scale and the only one with a customer in it.