Choosing a business model

If value is "hours of human work replaced," usage or outcome pricing fits better than seats. If value is "a place the team lives every day," seats fit. If value is "matching scarce supply to demand," a take rate fits. The wrong model fights the product forever.

Ask four questions.

1. Where does value accrue as the customer succeeds? Seats, usage, transactions, or attention?
2. What is the natural buyer and budget? Card on file, IT budget, performance marketing, COGS line?
3. What is the marginal cost of serving more? Near-zero (classic SaaS) versus real (inference, shipping, drivers)?
4. What distribution can you actually win? Community, sales team, two-sided liquidity, brand?

What the model forces you to build

Billing is not something you add later. Metering mistakes become revenue bugs. If you cannot explain how a dollar is earned and recognized, the model is incomplete.

Venture fit versus lifestyle fit

Not every good business is venture-scale. A profitable $8M ARR niche SaaS with 90% retention can be an excellent outcome for founders who want dividends — and a poor fit for a fund that needs power-law returns. Venture models usually need a path to a very large outcome: huge TAM, network effects, or expansion that does not die at the first customer profile. Be honest about which game you are playing; the funding path depends on it.

Pair the model with market sizing. A beautiful model in a tiny TAM is still a lifestyle business.

Pricing is packaging of the model

Price is how the model shows up on a checkout page. Three packaging mistakes:

1. Wrong value metric — charging seats when value is API calls. Customers feel nickel-and-dimed or undercharged.
2. No expansion path — one flat SKU so land equals ceiling.
3. Hidden COGS — "unlimited" plans when inference or SMS fees scale with power users.

Good packaging mirrors how buyers budget: monthly cards for consumers and SMB, annual invoices for mid-market, commits for infra. Revisit packaging when gross margin or sales-cycle data say the old meter is lying.

A one-page model memo

Before a pitch or a job decision, write six lines: (1) primary revenue line, (2) who the buyer is, (3) ACV or ARPU band, (4) gross margin band, (5) how a customer expands, (6) the kill condition. If you cannot fill those without hand-waving, you do not understand the company yet.

When you join, reverse-engineer the model from the billing page and the latest S-1 or blog metrics. Are they selling seats, usage, take rate, or ads? That tells you whether a reliability incident hits revenue immediately (usage), whether expansion depends on seat growth (SaaS), or whether the hard problem is liquidity (marketplace). In design reviews, ask: "Does this feature expand ACV, reduce churn, or only burn eng time?"

Common mistakes

• Picking seats because SaaS multiples look nice. If value is consumption, seats will feel unfair and cap expansion.
• Unlimited plans on variable COGS. Power users become your worst customers.
• A model you cannot instrument. If you cannot meter the event, you cannot charge for it cleanly.

Check yourself

Value is "hours of support work replaced" and your COGS are tokens. Seats or usage?

Usage or outcome pricing. Value and cost both scale with work done, not with headcount.

What are the six lines of the model memo?

Primary revenue line, buyer, ACV/ARPU, gross margin, expansion path, kill condition.

Key takeaways

• Four questions pick the model: value metric, buyer, marginal cost, distribution.
• Each model has product obligations. Billing and metering are the model, not extras.
• Venture fit and lifestyle fit are different games. Say which one you are in.
• Price packages the model. Wrong meter, no expansion path, or hidden COGS will fight you.

Further learning

Depth pages after the business models hub: SaaS, marketplaces, consumer, usage/API.