Subscription, transaction, marketplace, usage, licence. The model shapes your margin, your sales motion and how much money you need.
The main models:
Subscription — recurring fee for access. Predictable, fundable. Usage-based — pay per unit consumed. Scales with value, less predictable. Transaction / commission — a cut of each transaction. Needs volume. Marketplace — connect two sides, take a fee. Hardest to start, strongest once running. Licence — one-off or annual fee to use software. Advertising — free to users, monetise attention. Needs enormous scale. Services — paid for people's time. Low margin, no leverage.
Because the model determines nearly everything downstream: your gross margin, whether you can afford salespeople, how much cash you need, how predictable revenue is, and whether the business is fundable at all.
And because changing it later is close to starting a different company. It is worth thinking about properly once, early.
The same scheduling product, four ways:
Subscription — $99/user/month. Predictable MRR, easy to forecast, ~85% margin. Sales motion: direct. Most fundable.
Usage-based — $2 per rota published. Revenue scales with value delivered, so small firms pay little and grow into it. Harder to forecast and harder to sell to procurement, who want a fixed number.
Marketplace — connect firms with relief drivers, take 8%. Potentially the largest outcome and by far the hardest start: you need both sides at once, and neither shows up for an empty market.
Services — set up rotas for firms at $2,000/month. Fastest to revenue, ~30% margin, and it does not scale — every new customer needs a person.
Many companies start with services to fund themselves and learn the workflow, then productise. That is a legitimate path, provided you are deliberate about it rather than drifting into an agency by accident.
First-time founders often pick subscription by default because it is what startups do, without checking whether their customer's value is actually recurring. If someone uses your product twice a year, a subscription creates churn you did not need to have.
The second mistake: drifting into services. Customers ask for help, you say yes, and eighteen months later 70% of revenue is consulting and the product has not moved. It pays the bills and it quietly becomes the company.
The third: choosing advertising for a product with thousands rather than millions of users. Ad revenue at small scale is close to nothing.
Continuously → subscription. Per event → usage or transaction. Once → licence. Match the model to the shape of the value.
If delivery requires people per customer, you have a services margin whatever you call it.
A $500 ACV cannot support a salesperson. That forces self-serve, which forces product decisions.
Convention matters to buyers. Deviating can be an advantage or an obstacle — choose knowingly.
Set a percentage ceiling and a date. Otherwise it grows until it is the business.
Early, before pricing, and again whenever you enter a new segment.
Do not agonise pre-validation. Establish that people want it first; the model can be adjusted before you have many customers.
Apply this to your own startup in My Full Journey (free account).