Choose a business model

Why this step matters

A business model answers three questions: who pays, what they pay for and how often. The same product can succeed or fail depending on those answers. A tool sold as a monthly subscription to small businesses needs very different marketing, support and cash flow from the same tool sold as an annual contract to large companies, or given away and paid for by advertisers.

Choosing early does not lock you in, but it shapes everything that follows: what you build first, which channels make sense, how much you can spend to win a customer and whether venture capital is a good fit at all.

The best model is usually the one your customers already understand, priced against the value of the outcome rather than your costs, with margins that survive realistic acquisition spending. Simple beats clever at this stage: you want a model you can explain in one sentence to a customer, a co-founder and an investor, and test within a few weeks.

How to do it

1. List the models that fit your product. Common ones are subscription (SaaS), usage-based pricing, one-off purchases, transaction fees or commissions (marketplaces and payments), services or productised services, licensing, advertising and sponsorship, and affiliate revenue. Many startups combine two later, but start with one.
2. Check what the market already accepts. Your competitor sheet shows how customers are used to paying. Going against that habit (for example charging per seat where everyone charges per project) needs a strong reason.
3. Identify the payer and the user. They are not always the same. In B2B the buyer may be a manager or a finance team; in marketplaces both sides may pay; in ad-funded products the payer is the advertiser.
4. Build simple unit economics. In one sheet estimate price per customer, gross margin after hosting, AI model, payment and support costs, expected customer lifetime, and the cost to acquire a customer through your likely channels. Use pessimistic numbers. If lifetime gross profit per customer is not comfortably above acquisition cost, the model needs work.
5. Test the price with real prospects. Go back to the people you interviewed and show a simple pricing page or proposal. Ask them to choose a plan or sign a pilot. Their reaction to a real number is more useful than any survey.
6. Decide how you will fund the company. Some models need upfront capital (hardware, marketplaces that must seed both sides, enterprise sales cycles). Others can be bootstrapped from early revenue. Be honest about which path you are on, because it changes how fast you must grow.
7. Write it down in one paragraph. Payer, price, billing frequency, main channel, target margin and funding path. Share it with your co-founders and advisers and revisit it after your first 20 paying customers.

What it costs and how long it takes

This step is thinking and spreadsheet work, so it costs nothing. Two to five days is enough for a first version. Expect to revise it after launch when real conversion and churn data arrive.

Free credits, perks and services that help

Your model decides which programs fit. Bootstrapped founders should focus on free tiers and perks that cut running costs; see /for/bootstrapped. If you plan to raise, accelerators and investor programs give funding and pricing advice; see /accelerators. AI-heavy products should model inference costs early and look at /best/ai-credits, while infrastructure-heavy ones should check /best/cloud-credits. When you get to billing, compare options at /best/payments-billing.

Common mistakes

• "Free now, monetise later." Without a clear plan this usually means never. Even free products need a known path to revenue.
• Cost-plus pricing. Customers pay for the value of the outcome, not your costs. Start from what the problem costs them today.
• Ignoring variable costs. AI model usage, payment fees and support time grow with every customer. Leave them out and your margins look better than they are.
• Too many models at once. Subscriptions plus ads plus a marketplace splits your attention. Prove one first.
• Choosing venture funding by default. VC suits models that can grow very large very fast. Many good businesses do better bootstrapped.

Go deeper

Monetizes compares monetization methods (subscriptions, memberships, digital products, ads, affiliate and sponsorship) with the platforms and networks that pay for each, plus an earnings calculator.