SaaS business model: how it makes money, how to price it, and what it costs to run

The standard definition comes from the US National Institute of Standards and Technology (NIST). Its definition of cloud computing (SP 800-145) describes Software as a Service as: "The capability provided to the consumer is to use the provider's applications running on a cloud infrastructure." The applications are reached through a web browser or a program interface, and the customer "does not manage or control the underlying cloud infrastructure including network, servers, operating systems, storage, or even individual application capabilities".

That is the delivery side. Stripe's Atlas guide to the SaaS business model, written by Patrick McKenzie, calls SaaS "a billing and delivery model for software" that "restructures businesses around itself". For a founder that means:

• Revenue repeats. The guide says SaaS revenue "is generally recurring and predictable", so cash flow is easier to plan than with one-off license sales.
• You run the software. Customers install nothing, and hosting, uptime and support are your job.
• Customers can leave every billing period, which is why churn sits at the center of every SaaS metric.

The Atlas guide excludes mobile apps, which "have a separate billing model, imposed by the platforms' app stores". This post covers software sold on the web.

The Stripe Atlas guide calls its simplified revenue equation "the Rosetta Stone to understanding all material facts about a SaaS business". In words: long-term revenue is the number of customers you win (acquisition times conversion) multiplied by the average revenue per customer per period (ARPU), divided by churn.

The terms, as the guide defines them:

• Acquisition: how well you attract prospects.
• Conversion: the percent of prospects who become paying customers.
• ARPU (average revenue per user): the average revenue for an account over a period, such as a month.
• Churn: "the percent of customers over a given period who do not continue paying for services." The guide's example: 200 customers pay in January and 190 of them pay in February, so churn is 5%.
• Lifetime value (LTV): with simplifying assumptions, expected customer lifetime is the inverse of churn. At 5% monthly churn a customer stays 20 months on average; at $30 a month that is $600 of lifetime revenue per new customer.

Three consequences from the same guide are worth knowing before you set a price:

1. Improvements multiply. A 10% gain in acquisition and a 10% gain in conversion give a 21% improvement (1.1 times 1.1), not 20%.
2. Price is the easiest lever. Acquisition, conversion and churn usually need cross-functional work; "updating your pricing model typically requires replacing a small number with a bigger one."
3. Growth stops at a plateau you can predict. With fixed acquisition, conversion and churn, the customer count levels off at acquisition times conversion divided by churn. To show the arithmetic with made-up inputs: a product that adds 40 paying customers a month and loses 5% of them each month levels off at 800 customers (40 divided by 0.05). If 800 customers do not cover your costs, one of the three inputs has to change.

The Atlas guide says there are, broadly, two ways to sell SaaS, and that the sales model "dictates almost everything else about the SaaS company and the product". It treats this choice as more important than whether you sell to consumers or businesses, or whether you are bootstrapped or funded.

Low-touch SaaS is designed so most customers buy "without sustained one-on-one interaction with a human being". The channels are the website, email marketing and very often a free trial. The guide puts typical prices at around $10 a month for consumer (B2C) apps and $20 to $500 a month for business (B2B) apps, an average contract value of roughly $100 to $5K a year. The headline metric is MRR (monthly recurring revenue). Its example of the model is Basecamp.

High-touch SaaS relies on a sales team to find, convince and keep business customers. The guide puts contract values for small and midsize businesses sold this way at $6K to $15K a year, and says true enterprise deals "start in the six figures". The headline metric is ARR (annual recurring revenue). Its example is Salesforce.

The guide warns that running both models at once usually fails: "only one of the models receives any traction". For an indie founder with no sales team, low-touch is the realistic start, and it shapes the rest: a self-serve pricing page, a trial or free plan that shows value quickly, and support that scales through docs and onboarding.

Low-touch SaaS usually presents its packages as columns on a pricing page. The Atlas pricing guide describes each column as a plan "offered at a different price, with differential access to features or maximum allowable usage along some axis interesting to the business", where the axis might be users (seats), servers, or "something with a closer connection to the problem domain". The examples below are from official pricing pages read on October 19, 2026; prices are in US dollars as shown on those pages.

Flat or capped pricing. One price per tier, no matter how many people use it. Basecamp's pricing page says "No per-user fees, everyone's included" and sets its tiers by active projects and storage: Studio is $59 a month for up to 10 active projects with unlimited users, and the Unlimited edition is $299 a month, billed annually. Easy to buy and easy to explain, but revenue does not grow when a customer adds people.

Per seat (per user). The price scales with the number of users. Linear's Basic plan is $10 per user per month and Business is $16 per user per month, both billed yearly, with a free plan for small use. Revenue grows as a customer's team grows; small teams may pick a cheaper tool.

Tiered by a usage measure. Plans step up along a measure tied to the product's value. Plausible Analytics prices its plans by monthly pageviews, starting from up to 10k monthly pageviews, with a 30-day free trial and no credit card required.

Usage-based (metered). A base fee includes a quota, and overage is billed per unit. Supabase's Pro plan starts from $25 a month and includes 100,000 monthly active users, then $0.00325 per additional monthly active user, and 250 GB of egress, then $0.09 per GB. The price follows the customer's actual use, which makes bills less predictable for them and revenue less predictable for you.

Freemium. A permanent free plan with limits, and paid plans for more. Calendly's free plan includes one event type and one calendar connection; its Standard plan is $10 per seat per month billed yearly. Free users cost you hosting and support, so the free plan has to lead naturally to the paid one.

Many products mix these: Linear and Calendly pair a free plan with per-seat paid plans. The table just above the FAQ puts the five models side by side.

Two pieces of pricing advice from the Atlas pricing guide that fit a first launch:

• Show few choices. Reviewing one founder's pricing page with eight plans, the guide recommends condensing to "three plus 'call us for enterprise pricing'".
• Revisit prices often. The guide says most SaaS companies set prices "without much consideration and leave them alone for years at a time" and encourages revisiting them quarterly.

Card up front or not. The Atlas guide says the decision to require a credit card for a trial "dominates the character of the free trial". Its rule of thumb for low-touch trials that do not require a card: substantially below 1% conversion to paid is "generally evidence of poor product-market fit", about 1% is a baseline for competent execution and 2% or more is "extremely good". For trials that require a card, the same guide gives substantially below 40% as a warning sign, 40% as a baseline and 60% as doing well. It adds that requiring a card usually brings more paying customers on net early on, until you get good at onboarding and lifecycle email. These are one author's rules of thumb, not survey data.

Annual billing. The guide calls selling annual accounts "a good idea", both for the cash collected up front and because annual accounts have lower churn rates. Linear's and Calendly's paid prices above are their billed yearly rates.

Every subscription payment carries a fee. Here are the published US list prices from each provider's own pricing page as of October 2026:

• Stripe Payments: 2.9% + 30¢ per successful transaction for domestic cards, plus 1.5% for international cards and 1% if currency conversion is required.
• Stripe Billing (subscription management on top of payments), pay as you go: 0.7% of Billing volume.
• Stripe Managed Payments: adds 3.5% per successful transaction to standard Stripe processing fees. Stripe says this covers global indirect tax (sales tax, VAT, GST), fraud prevention, dispute responses and transaction-level support.
• Paddle: 5% + 50¢ per Checkout transaction, as a merchant of record.
• Lemon Squeezy: 5% + 50¢, as a merchant of record.

To show the arithmetic on a single $20 domestic card payment, using those list prices: Stripe Payments plus Billing comes to $1.02 (58¢ + 30¢ + 14¢). Paddle and Lemon Squeezy each come to $1.50 (5% of $20 is $1.00, plus 50¢). Managed Payments on top of standard processing comes to $1.58 (6.4% of $20 is $1.28, plus 30¢), before any optional products such as Billing. Fixed per-transaction fees weigh more on cheap plans, which is one more reason low-touch SaaS rarely charges a few dollars a month.

What a merchant of record does. Paddle defines a merchant of record as "a legal entity responsible for selling goods or services to an end customer" that manages payments and takes on "the associated liabilities, such as collecting sales tax, ensuring Payment Card Industry (PCI) compliance, and honoring refunds and chargebacks." Lemon Squeezy says it takes on "the liability of tax collection and calculation and pay[s] taxes on your behalf." The trade-off is a higher fee in exchange for not handling sales tax and VAT yourself. Which side of that trade is right depends on where your customers are; this post does not give tax advice.

You do not need a dashboard on day one, but you do need these numbers, all defined in the Stripe Atlas guide:

• MRR: monthly recurring revenue, the main number for low-touch SaaS.
• ARR: annual recurring revenue. The guide describes it as "essentially all of the non-churned revenue of the company minus certain nonrecurring items, such as one-time setup fees".
• Churn: the share of customers who stop paying in a period, quoted monthly in low-touch SaaS.
• CAC: customer acquisition cost, "the marginal spending on marketing and sales per customer added".
• Net revenue churn: revenue lost from cancellations and downgrades minus revenue gained from upgrades in the same customers. The guide calls negative net revenue churn the gold standard for high-touch SaaS and says virtually no low-touch business achieves it.

For low-touch monthly churn, the Atlas guide's rules of thumb are: 2% is "a very sticky SaaS product", 5% is "roughly where you expect to start", 7% points to fixable problems or a hard market, and 10% or more is "evidence of very poor product-market fit". It also notes that selling to freelancers and other informal businesses raises churn, because those customers exit their own businesses more often, and that raising prices can lower churn by changing which customers buy.

It can be, and the shape of the costs explains why it often takes time. The Atlas guide says SaaS companies "frequently spend less than 5~10% of their marginal revenue per customer on delivering the underlying service", so serving one more customer is cheap. The expensive part comes first: marketing and sales costs land early in a customer's life, and the revenue that pays for them arrives over the following months. A company chasing fast growth "will almost always spend more money in a given period than they collect from customers."

On timing, the guide says bootstrapped SaaS businesses "often take 18 months before they're profitable enough to be competitive with reasonable wages for the founding team", and that for a bootstrapped company 10 to 20% yearly revenue growth "can produce very, very happy outcomes". Funded companies are held to a different bar: the guide's rule of thumb for aggressive growth from over $1 million ARR is to triple revenue two years in a row, then double it three years in a row. It describes the slow start many companies go through as "the Long, Slow SaaS Ramp of Death".

These are the author's observations, not measured averages.

The order below is our suggestion, built from the sources above. It is a planning aid, not a formula.

1. Pick the sales model first. If a buyer can sign up, see value and pay without talking to you, plan for low-touch. If every deal needs a call or a custom contract, you are building high-touch.
2. Pick the value metric. Charge along the axis that grows with the value a customer gets: seats, projects, pageviews, active users or API calls. If none fits, use a flat price per tier.
3. Publish three plans, named so a buyer can place themselves.
4. Choose a trial or a free plan. A free plan suits products that spread through teams or shared links; a short trial suits products whose value shows within days.
5. Offer annual billing next to monthly.
6. Choose how you collect payments, using the fee arithmetic above for your actual price.
7. Run the equation before launch. Estimate monthly new customers, price and churn, and work out the plateau. If it does not cover your costs, change the price, the channel or the market before writing more code.
8. Launch small, then measure. A soft launch to a limited group lets you test the price and the trial before a public launch, and a Show HN or directory listings can bring the first visitors. Long term, low-touch SaaS depends on being found, which is what our SaaS SEO guide covers.

What we could not verify. The churn, conversion and growth figures in this post are rules of thumb from the Stripe Atlas guides, which do not show a publication date on the page and do not cite a dataset. We found no official or survey source for typical indie SaaS revenue, profit margins or time to profitability, so this post gives none. Prices and fees change often; the figures above are list prices shown on the official pages on October 19, 2026, and some pages show yearly billing by default. We did not check sales tax rules in any state or country.

How this post was put together. We did not build or run a SaaS for this post, and it contains no statistics of our own. On October 19, 2026 we read NIST SP 800-145, two Stripe Atlas guides, the official pricing pages of Stripe, Stripe Billing, Stripe Managed Payments, Paddle and Lemon Squeezy, Paddle's merchant of record explainer, and the pricing pages of Basecamp, Linear, Plausible Analytics, Supabase and Calendly. The fee examples are our own arithmetic from those list prices. VibeLaunch, which publishes this post, helps founders submit their product to directories at launch. The source for every figure is in the list below.