Revenue Model Selection

Choosing how a company captures value (subscription, usage, transaction take-rate, marketplace commission, licensing, services, or advertising) and reading how that choice shapes the fundraising story, margin profile, and defensibility for years.

Two companies can sell to the same customer, solve the same problem, and grow at the same rate while investors read them very differently. One charges a fixed subscription. The other takes a percentage of every transaction it processes. At a Series A, the subscription company gets credit for recurring revenue; the take-rate company has to prove how much of its volume will repeat. The product may be similar. The revenue model is doing most of the talking.

Context

This pattern sits at the founding-formation stage, though the choice is often revisited as customers reveal what they will actually pay for. It applies to a pre-revenue startup choosing its first model, a company with early traction deciding whether to add a second, or a founder writing the deck and realizing the model is the part investors will price.

The decision is distinct from the value proposition, which settles what value the company creates. The revenue model is the mechanism that captures a share of that value and routes it back to the company. The same value can be captured many ways — a tool that saves a customer money can charge a flat subscription, a percentage of the savings, a per-use fee, or a license — and the choice among them is not a billing detail. It sets the shape of the company’s economics, the story it tells investors, and the moat it can or cannot build.

The common model families a software company chooses among:

The company isn’t confined to one row forever. NFX has argued repeatedly that durable companies stack models over time: a marketplace adds subscription tools for sellers, a SaaS product adds a usage-based tier, or a transaction business adds SaaS to smooth revenue. But stacking is a sequence of deliberate additions. It isn’t a reason to dodge the first choice.

Problem

A founder has to pick a primary model before the company has enough data to know which one the market will reward. The choice then compounds through every round, margin calculation, and defensibility claim. The wrong model isn’t usually fatal at the start; early customers will pay almost any reasonable way to get a product they need. It becomes a drag later. Services revenue gets discounted because it doesn’t recur. A thin take-rate can’t fund an enterprise sales motion. Advertising requires an audience scale most startups never reach. By the time the problem is visible, the model is woven through contracts, pricing pages, and customer expectations.

Forces

• Predictability versus upside. A fixed subscription is easy to forecast and easy to price at a premium multiple. Usage or take-rate models can capture more as customers grow, but they’re lumpier and harder to underwrite.
• Value alignment versus billing friction. A fee tied to savings or outcomes can feel fair because revenue rises with customer value. It can also be hard to measure, invoice, and budget, which makes it harder to buy.
• Margin profile versus market access. Pure software subscription carries high gross margin and the cleanest fundraising story. Services or transaction-heavy models may reach customers software alone can’t, but investors discount the lower-margin revenue.
• Defensibility versus simplicity. Models with recurring contact and switching costs, such as subscription, marketplace, and usage tied to integrated data, can harden into a moat. A one-time charge is simpler, but it leaves the company re-winning the customer at every sale.

Solution

Choose the primary model by matching three things: the shape of the value, the buyer’s budgeting reality, and the fundraising story the founder needs to tell. Then design the early contracts so a second model can be stacked later without re-pricing the base.

First, match the model to how value accrues. Continuous value fits subscription. Consumption-scaled value fits usage. A product whose value is matching two parties usually fits a commission on the match. The first question is not “what’s easiest to bill?” It is “where does value show up, and over what period?”

Second, respect how the buyer buys. Enterprise buyers like predictable line items and struggle with variable bills they can’t forecast, which is why enterprise software often uses committed subscriptions even when usage would capture more. Self-serve buyers tolerate usage or freemium because the first commitment is small. The model has to fit the customer’s buying process, not only the company’s preferred spreadsheet.

Third, read the fundraising consequence before committing. Recurring subscription revenue with strong net retention is the shape venture investors price most generously, because it is predictable and compounds. Usage-based revenue is well understood now, but it invites questions about committed versus discretionary spend. Transaction and marketplace revenue is read on take-rate and on whether the gross figure or net figure is the real business. Services revenue gets the deepest discount because it scales with headcount rather than software.

Fourth, preserve room to stack. A marketplace that may later sell subscription tools to sellers shouldn’t bury those tools inside the commission. A usage business that may add committed subscription tiers shouldn’t price usage so low that the floor looks like a price hike. Stacking is how the company expands its share of the value it creates; the option is cheap to keep early and expensive to retrofit.

How It Plays Out

Shopify and Stripe both monetize commerce, and both layer models, but investors read the layers differently. Shopify charges merchants for software and takes a cut of payments processed through it. The subscription gives investors a recurring base to underwrite; the transaction revenue scales with merchant success. The stack is the point. The subscription stabilizes the revenue investors price, and the take-rate captures upside when merchants grow.

The cautionary shape is the startup that chooses advertising before it has audience scale. Advertising monetizes attention, and revenue per user is small, so the model needs enormous volume before it produces meaningful revenue. Founders are drawn to “free” because it removes the friction of charging, but it also delays monetization until a future that has to be very large. For most startups, advertising means the company needs a hit-scale audience before the business model works.

Consequences

Benefits. A model matched to the shape of the value lets the company capture a fair share of what it creates instead of leaving it on the table. A recurring model compounds lifetime value across renewals rather than resetting it at every sale. The right model also tells a cleaner fundraising story: the investor spends the meeting on the business, not on what the revenue really is. And a model with recurring contact and switching costs can turn billing into part of the company’s defensibility.

Liabilities. The choice is sticky. Once customers are priced on a model, moving them to another means re-pricing the base, risking churn, and stalling the company. Optimizing for the fundraising-friendly model can pull the company away from the model customers would most readily buy. Stacking models adds operational complexity: multiple billing systems, sales motions, and sets of unit economics. And no model rescues a weak value proposition. A precise capture mechanism on a product customers don’t need captures a percentage of nothing.

Related Articles

Sources

• Alexander Osterwalder and Yves Pigneur, Business Model Generation (2010) — the canonical framework for the components of a business model, including the revenue-streams building block that distinguishes recurring from transaction-based capture and asset-sale from usage and subscription pricing.
• NFX’s writing on business models and the case for stacking multiple revenue models over time — the practitioner source for the view that durable companies layer models (marketplace plus SaaS, usage plus subscription) rather than staying on a single one.
• a16z, “16 Startup Metrics” and the companion “16 More” — the standard practitioner reference for how investors read recurring revenue, net retention, and the metrics that differ by model type, and why subscription revenue commands the cleanest underwriting.
• Clayton Christensen, Karen Dillon, and others’ work on jobs-to-be-done informs the value-alignment principle here: a model that charges for the job the customer is hiring the product to do aligns capture with value in a way that survives competition better than a model bolted on after the fact.