Founders pick a revenue model by copying whoever they admire, and then spend two years discovering that the model decides their burn rate, their sales cycle, and what an investor will pay for the company. Choosing it on purpose costs a weekend. Discovering it costs a round.
Key takeaways
A revenue model is how you charge. A business model is the whole system, of which charging is one part.
Subscription, transactional, marketplace, usage-based, licensing, advertising, and services each suit a different kind of product and a different kind of customer.
The model decides your cash flow more than your revenue does. Two companies at the same ARR can have very different bank balances.
At pre-seed, the most common pricing mistake is charging too little, and the second is not charging at all.
Investors value revenue models differently, so the model you choose changes your multiple before it changes your revenue.
What is a revenue model, and how is it different from a business model?
A revenue model is the mechanism by which money moves from a customer to you. Monthly fee, one-off purchase, a cut of a transaction, a charge per unit consumed.
A business model is larger. It covers who the customer is, what you make, how it reaches them, what it costs you to deliver, and how you charge.
The revenue model is one component inside it, which is why the two terms get used interchangeably and why the confusion matters. You can keep the same business model and change how you charge. Changing the revenue model alone is a pricing decision. Changing the business model is a pivot.
For a founder at the earliest stage, the practical version of the question is narrower: what triggers a payment, and how often does it repeat.
What are the main revenue models, and who does each one suit?
Subscription. A fixed fee for ongoing access, charged monthly or annually. Suits products the customer needs repeatedly rather than once, which is why software defaulted to it. Works badly when the underlying need is occasional, because customers notice they are paying for something they used twice.
Transactional. They buy, they pay, the transaction ends. Consumer brands, e-commerce, and most physical products. Simple to explain and simple to start, and every month begins at zero.
Marketplace. You connect two sides and take a percentage of what flows between them. The percentage is your take rate, and it is the number that decides the business. Hard to start because you need both sides at once, and durable once you have them.
Usage-based. The customer pays for what they consume, metered. Common in infrastructure, APIs, and anything where consumption varies a lot between customers. Revenue grows as your customer succeeds, which founders like, and it is difficult to forecast, which finance teams do not.
Licensing. Someone pays to use your intellectual property in their own product. High margin, small number of large contracts, long sales cycles.
Advertising. You give the product away and sell the attention. Needs scale before it produces anything, and the audience you need is larger than early-stage founders assume.
Services. You sell time and expertise. Cash comes in quickly, and the business is capped by how many people you can hire, which is the trade a lot of founders make without realising they have made it.
Freemium sits slightly outside this list. It is a way of acquiring customers rather than a way of charging them, and what sits behind the free tier is one of the models above.
How do you choose a revenue model at pre-seed?
Three questions settle most of it.
How often does the customer have the problem? Daily or weekly problems support subscriptions. Annual or occasional problems do not, and trying to force a subscription onto an occasional need produces high churn that looks like a product failure.
Who decides, and how long do they take? A model that requires an annual contract needs a buyer with budget authority, which lengthens your sales cycle. A transactional model can be sold to the person with the problem.
Can you deliver it repeatedly without proportional cost? If serving the tenth customer costs the same as the first, you are closer to a services business than a product one, and the model should reflect that until the cost curve bends.
A useful exercise before you commit: write down what the customer is buying in their own words. If the answer is “an outcome, once”, you are transactional. If it is “the ability to keep doing something”, you are closer to subscription.
What does each model do to your cash flow?
Revenue and cash are different things, and the model is where they come apart.
Subscription builds slowly and predictably, and it is hardest on cash in the first year because you are paying your acquisition cost upfront and collecting it back a month at a time. Annual prepayment solves this and is worth a discount to get.
Transactional gives you cash on the day of sale, which feels healthier than it is. There is no base to build on, so every month you start again, and growth means spending more on acquisition every month.
Marketplace has money flowing through your account that is not yours. The number that matters is net revenue after payouts, not the gross value of transactions. Confusing the two is the most common way a marketplace founder overstates the business, including to themselves.
Usage-based revenue is lumpy. A good month from one large customer can hide the fact that nobody else expanded.
Services are cash positive fastest. It is also the model where revenue and headcount move together, so growth does not improve your margin.
How do you price what you are selling?
Three ways to arrive at a number. Cost-plus, which works out what it costs you and adds a margin. Competitor-based, which looks at what others charge. Value-based, which starts from what the outcome is worth to the customer.
Most early-stage companies use competitor-based pricing while telling investors they use value-based pricing.
Two errors show up repeatedly at pre-seed. The first is pricing too low, often from a fear of rejection, which is expensive twice over because a low price attracts customers who were never going to be good ones and because raising a price later is harder than setting it correctly now.
The second is not charging at all, running a long free pilot in the hope that usage converts into willingness to pay. It rarely does. Interest and payment are different signals, and only one of them tells you anything.
The cheapest test available to you is to ask for money earlier than feels comfortable. A price is the only validation that cannot be given politely.
What do investors read into your revenue model?
The model affects your valuation before it affects your revenue, because different models carry different multiples. Recurring revenue is valued higher than one-off revenue, for the simple reason that next year’s number is more predictable.
A services business and a software business at identical revenue will be valued very differently, and the founder of the services business is often surprised by this at exactly the wrong moment.
What gets examined underneath the headline number:
Gross margin, not revenue. Revenue with thin margin buys less than a smaller number with a fat one.
Net revenue for marketplaces. An investor will want your take rate, and quoting gross transaction value instead reads as avoidance.
Revenue concentration. Three customers making up most of your revenue is a different risk from thirty.
Expansion. If existing customers spend more over time, that is the strongest signal a revenue model is working.
At pre-seed, none of these numbers are large yet, and nobody expects them to be. What an investor is reading is the shape of the business underneath, which is set by the model you picked rather than by how much revenue you have so far.
When should you change it?
Changing how you charge is cheaper than changing what you build, and founders leave it later than they should.
Reasonable triggers: churn concentrated in a particular customer type, a sales cycle that keeps stalling at the same objection, or customers consistently asking to buy in a way you do not offer. That last one is the clearest signal available and the easiest to ignore.
The cost of changing is real. Existing customers on old terms, contracts that have to be honoured, and a period where your numbers are hard to compare to last quarter. Doing it once, deliberately, is manageable. Doing it three times in a year tells an investor you have not worked out who the customer is.
Frequently asked questions
Can a startup have more than one revenue model?
Yes, and many do as they grow. At pre-seed it is a bad idea, because two models mean two sales motions, two pricing conversations, and a team too small to run either properly.
Which revenue model is best for a SaaS startup?
Subscription or usage-based, and the choice turns on how much consumption varies between customers. Where every customer uses roughly the same amount, a flat subscription is simpler to sell. Where usage varies by an order of magnitude, metering captures value you would otherwise leave behind.
How do I price when I have no customers yet?
Price against the cost of the problem rather than the cost of your product. Ask what the customer currently spends in money or hours on the thing you are replacing, and price as a fraction of that.
Is freemium a revenue model?
No. It is an acquisition strategy with a revenue model behind it. The question to answer is what the paid tier charges for and why a free user would ever cross over.
Does the revenue model affect my valuation?
Yes, through the multiple applied to your revenue. Predictable, high-margin, expanding revenue is valued more highly than lumpy, low-margin revenue at the same headline number.
