What is business design for a startup?

A first version that people love and that loses money on every user is not a product. It is a gift. Business design holds three questions at once, do people want it, can we build and deliver it, does it close as a business, and treats each as a set of hypotheses with a number and a date. Startups postpone the third question because the first two are more fun. This whitepaper is about asking all three before building.

What do desirability, feasibility and viability mean in product?

A first version that people love and that loses money on every user is not a product. It is a gift. Business design holds three questions at once, do people want it, can we build and deliver it, does it close as a business, and treats each as a set of hypotheses with a number and a date. Startups postpone the third question because the first two are more fun. This whitepaper is about asking all three before building.

Why do startups postpone thinking about the business model?

A first version that people love and that loses money on every user is not a product. It is a gift. Business design holds three questions at once, do people want it, can we build and deliver it, does it close as a business, and treats each as a set of hypotheses with a number and a date. Startups postpone the third question because the first two are more fun. This whitepaper is about asking all three before building.

Introduction · Business Design

§The fourth whitepaper ended with a and a first unit of work at the bottom of it. Something is about to be built. This whitepaper asks the question that the ladder deferred: when it exists, is it a business?

§The three lenses are old. Tim Brown's IDEO framed innovation as the overlap of what people desire, what is technically and organizationally , and what is financially viable, and Marty Cagan's four risks say the same thing in product's vocabulary: value and usability are desirability, feasibility is feasibility, viability is viability. What is new is how cheaply a small team can now satisfy the first two and how completely that cheapness tempts it to skip the third.

§ 0.1Why the third lens gets postponed#

§ is the fun part. It is conversations, prototypes, strangers lighting up. Feasibility is the part engineers love: can it be built, and how. Viability is a spreadsheet with a stranger's willingness to pay in one cell and your costs in another, and the cell you fear is the one at the bottom.

§Startups postpone it for three reasons, all of them understandable. The first is that early of desire feels like evidence of a business, and it is not: people desire many things they will not pay for. The second is that the founders' instinct says growth first, monetization later, an instinct inherited from an era of cheap capital and near-zero marginal cost that the AI era has quietly ended. The third is that viability questions are the ones most likely to kill the idea, and nobody wants to kill the idea in week six.

§The postponement comes back, and it comes back at the worst moment: after the first version exists, when the team discovers that the segment that loved it is the segment that cannot pay, or that every user costs more in inference than they return, or that the operation cannot deliver what the prototype promised at any price.

§ 0.2Hypotheses, numbers, dates#

§The discipline in this whitepaper is Rita McGrath's, stated thirty years ago and rarely followed: a new venture is a set of assumptions, and the work is to list them, attach a number to each, and test the one that would kill the plan first. Every chapter here ends by adding lines to one running list, , and chapter 12 collects the list with an owner and a date on each line.

§The numbers are not forecasts. They are thresholds: if fewer than this many pay, if it costs more than this to serve one, if the operation can absorb fewer than this per shift, the is in doubt. A number you can be wrong about is a hypothesis. A number you cannot be wrong about is a slide.

§ 0.3What this whitepaper is not#

§It is not go-to-market. Channels, launch, positioning, sales motion, and acquisition are deliberately absent. They are downstream of the question this whitepaper asks, and a team that plans a launch before it knows whether one customer is profitable has planned the wrong thing. It is not the conversations, which the third whitepaper owns, nor the direction, which the fourth owns, nor the experiments that will test these hypotheses in market once the product exists, which the sixth owns.

§ 0.4The running case#

§One case runs through: a multi-brand hospitality group in 2026, whose and strategy the previous two whitepapers followed. Its business design had four segments, frequent customers, occasional customers, staff and managers; a business model built around loyalty tiers; monetization hypotheses on several fronts; and one false assumption at the center, that customers would need to be paid for information they turned out to give gladly. The case is told without names, with the numbers.

§Part I is desirability: what each segment values and where segments pull apart. Part II is feasibility: can we build it, can we deliver it, and in what order. Part III is viability: the model, the unit economics, monetization as hypotheses, and who pays first. Part IV stress-tests all of it against data the company already has and collects the list. Part V is how it fails, and the case. The conclusion is what AI drafts and what still needs a spreadsheet and a person who has run a profit-and-loss statement.