Markets and business models
Hybrid business models
Almost every large company mixes revenue lines by year ten. Start with one primary model, then add a second that reinforces it.
Almost every large software company is a hybrid by year ten. Slack sells seats, then enterprise plans, then AI usage add-ons. AWS is usage with enterprise commits. Uber takes a marketplace cut, then sells subscriptions and ads. Shopify stacks a SaaS subscription, a payments take rate, capital and an apps ecosystem.
Evolution is normal. Early on, choose one primary model so the product, metering and go-to-market stay coherent. Add a second revenue line when it reinforces the first — payments on a SaaS surface, seats on an API platform — not when you are bored.
Common migrations
- Self-serve SaaS → enterprise packaging (security, SSO, contracts).
- Pure seats → seats plus usage (AI features, storage, compute).
- Pure usage → commits or prepaid credits (forecastability for both sides).
- Marketplace → managed marketplace (ops in the middle for quality).
- Open source → cloud control plane.
Each migration changes unit economics: CAC payback, gross margin, and what "a customer" means in the data model.
Walking through a stack
Shopify is a useful picture. Merchants pay a subscription (SaaS). Many also process payments, so Shopify keeps a take rate on volume (transactional). Capital and the app store are extra lines that sit on the same merchant relationship.
The payments line works because the merchant already lives in the product. A random new marketplace on the side would split focus. The test: does the second line use the same buyer, the same data model and the same reason the first line exists?
OpenAI is another stack: consumer subscriptions, API usage and enterprise contracts. Three animals, one capability. The risk is that each line needs different packaging, support and margin math. See usage economics for the API side.
What the second line forces you to build
| Added line | New obligation |
|---|---|
| Usage on top of seats | Metering, quotas, overage invoices |
| Seats on top of an API | Entitlements, admin, dunning |
| Payments on a SaaS surface | Ledgers, reconciliation, risk |
| Ads on a consumer app | Inventory, auctions, fraud |
Billing is not a later feature. Metering mistakes and entitlement bugs become revenue bugs. If you cannot explain how a dollar is earned and recognized, the hybrid is incomplete.
Common mistakes
- Kitchen-sink pricing. Three meters on day one confuse buyers and your own finance team.
- A second line that fights the first. Selling cheap one-time projects that train customers not to subscribe.
- Migrating before the first model works. Enterprise packaging will not save a product nobody keeps.
Check yourself
Why add usage to a seat product instead of raising the seat price?
Value may grow with consumption (tokens, storage, compute), not with headcount. A usage line captures that without punishing light teams.
Name a hybrid that reinforces the core versus one that is a distraction.
Shopify payments on top of merchant SaaS reinforces checkout. A side marketplace with a new buyer is a distraction.
Key takeaways
- Hybrids appear later, when a second line uses the same buyer and relationship.
- Start with one primary model so metering and GTM stay coherent.
- Each migration changes CAC, margin and the definition of a customer.
- Instrument billable events from day one. Retrofitted metering is expensive.
Further learning
Next, how to choose a model.