SaaS Business Valuation: Check the Revenue, Cost and Retention

A subscription is a promise to keep serving a customer. It is not a promise that the customer will keep paying you.

That is the point I would keep in view when comparing a small profitable SaaS with a venture-funded company. Both sell software, but they can require very different amounts of money, work and patience from the buyer.

This guide is for owners and buyers of established SaaS businesses. It helps you prepare a defensible calculation or question an asking price. It does not produce an appraisal or a market multiple for a business we have not examined.

SaaS business valuation: choose the number before the multiple

Define the revenue and earnings basis before applying any valuation assumption. A small owner-operated SaaS may be discussed in terms of earnings, while a business reinvesting heavily in growth may need a broader revenue and cash-flow analysis; the label “SaaS” does not decide the answer.

FE International's SaaS valuation explanation discusses SDE, EBITDA and revenue approaches. Provider published terms, checked October 24, 2026. It is the provider's account of its process, not proof that its published ranges apply to your business. We have not adopted those ranges here.

Stripe explicitly distinguishes MRR from GAAP revenue and says its Billing calculation includes active and past-due subscriptions. It normalizes annual subscriptions into monthly amounts. Read Stripe's MRR and ARR documentation. Provider published terms, checked October 24, 2026. Other systems and custom configurations may use different definitions.

A $1,200 annual subscription can contribute $100 to MRR while producing $1,200 of cash in one month. A $5,000 implementation project can produce revenue without adding $5,000 to recurring monthly revenue. A past-due subscription can appear in a metric without becoming cash in your bank.

Those are reconciliation questions. They are not reasons to discard MRR; they are reasons to know what it means.

The earnings bridge: what survives the founder's departure?

Rebuild profit from the accounts and show owner labor separately. A profitable product can be much less profitable to a buyer who needs to hire a developer, cover support or restore deferred maintenance.

Fictional example: Slate Desk, a small subscription SaaS. USD, one trailing-12-month period. User assumption. None of these figures describes a real listing.

The $115,000 result is the example's normalized owner-operated earnings basis. The $55,000 result assumes hiring out the remaining necessary founder work. Neither is cash after acquisition debt, income tax or new investment.

An add-back is an expense restored to earnings because of a documented adjustment. The founder's pay may qualify when assessing a working-owner purchase. That does not mean the product can operate without anyone doing the founder's job.

The distinction is especially important in software. “Only a few hours a week” may mean the founder solved the hard problem last month and is waiting for the next one. A buyer still needs the skill, access or budget to handle it.

The Deal Decision Check

Say a fictional buyer needs $70,000 a year before financing and tax while hiring out all necessary work. Slate Desk's assumed $55,000 falls short: does not fit. It is irrelevant that $115,000 of owner-operated earnings looks attractive.

If the same buyer can and wants to do the founder's work, that is a different requirement. It needs an honest time and skill check rather than a quiet change to the spreadsheet.

A founder preparing to sell can use this bridge to separate the sale argument from the operating job. Our seller guide explains what records and decisions come next.

Prepare your numbers before choosing a sale route

Revenue can grow while the old customers leave

Read growth and retention together. New sales can hide a weakening existing customer base, while expansion from a few large accounts can hide customer losses elsewhere.

Here is a one-month revenue bridge for Slate Desk. All inputs are fictional User assumption. Each recurring-revenue amount is monthly-normalized USD; it is not cash collected.

The total grew by $2,000, or 10%. Yet the starting group of customers ended at a lower recurring run rate.

Net revenue retention, or NRR, includes expansion, contraction and churn from the same starting customer group. In this example:

($20,000 − $2,000 − $1,000 + $1,500) ÷ $20,000 = 92.5% monthly NRR.

Gross revenue retention, or GRR, excludes expansion:

($20,000 − $2,000 − $1,000) ÷ $20,000 = 85% monthly GRR.

New customers do not belong in either numerator. Stripe describes the same-cohort approach in its NRR explanation. Provider published terms, checked October 24, 2026. We use the formula and our own fictional inputs, rather than adopting a benchmark.

Do not compare a monthly result with an annual benchmark as if they share a time unit. Do not forecast a full year by mechanically repeating this one fictional month. Ask for the actual monthly history, annual cohorts where available, and consistent definitions.

At $22,000 of ending MRR, a simple MRR × 12 calculation gives $264,000 of annualized recurring revenue. It does not prove that the last year's revenue was $264,000 or that the next year's will be. Ask how the provider handles usage fees, discounts, trials, failed payments, pauses, refunds and one-time services.

For a small SaaS, look at customer counts and revenue together. Losing 20 small customers while one large customer expands can leave revenue flat and concentration much higher. You need to understand the business underneath that flat line.

A sensitivity table is not a market valuation

Show the effect of changing assumptions without presenting those assumptions as sale evidence. A useful worksheet makes uncertainty visible; a misleading one turns arbitrary inputs into a precise-looking estimate.

For Slate Desk, these are deliberately selected arithmetic inputs applied to its fictional $115,000 annual owner-operated earnings:

User assumption. These are not observed sale multiples, a market range, an appraisal or a prediction. The repeated $55,000 is deliberate: paying more does not create more operating earnings. Financing a higher price may leave less owner cash.

A credible comparable must define the business size, growth, revenue mix, profitability, retention, owner involvement, date and actual consideration. Public software-company multiples or a disclosed strategic acquisition can be poor comparisons for a small product supported by one founder. If you use them, explain the differences rather than presenting them as interchangeable deals.

An asking price is not a completed transaction. A broker's opinion is not independently verified sale evidence. A headline price may include an earnout, seller financing or stock rather than cash on the closing date.

An earnout is a payment that depends on later conditions. A seller note is debt the buyer owes the seller under agreed terms. A holdback is consideration kept back for a period or condition. For valuation and negotiation, record each component and the chance of collecting it separately from cash already payable.

If a provider offers a headline number, ask: “How much is fixed, how much is cash at closing, and what has to happen for the rest to be paid?”

Annual plans and lifetime deals create service obligations

Cash collected before a sale can pay for work the buyer must perform afterward. Map the remaining service obligations and agree their treatment in the transaction, rather than assuming every past dollar is free cash for the seller.

Say a fictional customer paid $1,200 for an annual plan one month before closing. Depending on the service dates and accounting policy, much of the service period remains. The buyer may inherit the need to host, support and serve that customer after the seller has received the cash.

That does not automatically require a dollar-for-dollar price deduction. It requires an explicit agreement about prepaid revenue, costs, liabilities and any closing adjustment. An accountant can help reconcile the accounting, and counsel can help define the transfer.

Lifetime offers deserve the same attention. The customer may owe no more subscription fees while continuing to consume hosting, support or API usage. A historical launch spike can therefore coexist with a recurring cost burden. Split those users from paying subscription cohorts and measure their costs separately.

This is also why profit alone needs context. A business can look cash-rich after collecting annual subscriptions and still need a reserve to perform the work.

Field Notes: the product is more than its repository

Operator experience. Building and running SaaS teaches you that the visible feature is only part of the job. Support, billing failures, migrations and keeping old integrations working can consume time that never appears on a product tour.

That experience comes from operating software businesses, not from brokering or completing an acquisition through the providers covered here. We have not done those transactions.

For a buyer, I would want a demonstration of four ordinary events:

1. A new customer signs up, pays and gets access.
2. An existing customer changes a plan or cancels.
3. A failed payment or support problem is resolved.
4. A code change is deployed, monitored and rolled back if needed.

Then ask what happens when the founder is absent. A product that is easy for its creator to maintain may be difficult for the next owner, especially if the creator holds the billing credentials, deployment knowledge and customer relationships.

Source-code access does not settle intellectual-property ownership. Request the relevant employee and contractor agreements, licenses and a list of third-party components. Do not assume that permission to use one service account includes the right to transfer it. Verify important services and customer contracts individually.

Scaling also has a different cost shape from a storefront screenshot. A new enterprise customer can bring implementation, security questionnaires and bespoke support. More usage can increase API or infrastructure costs. If the seller describes all growth as nearly free, ask for contribution by customer or usage segment.

My view on AI and a small SaaS valuation

AI can reduce development work and make a narrow product easier to copy at the same time. I would examine actual customer behavior and costs before giving either possibility much weight in a price.

For an AI-enabled product, request usage-based gross margins, rate limits, provider dependencies and the process for errors. A subscription price is fixed more often than the user's consumption is. If one heavy user can consume the margin from several quiet users, an average is hiding the part you need to investigate.

For a product AI might replace, ask why customers keep it: workflow, integrations, records, trust, distribution or a result the generic tool does poorly. “We have AI” is not a reason by itself. “Customers tried the alternative and renewed” is a useful lead that still needs evidence.

Those are editorial judgments and diligence questions, not a forecast that AI will raise or destroy a particular business's value.

Build a valuation packet a buyer can question

A good packet lets another person trace the claim to a source and see what remains unresolved. It should work even if that person disagrees with your proposed price.

Do not send secrets, private customer data or unrestricted production access just to get an initial price opinion. Share proportionate evidence through an appropriate confidential process. A serious buyer can inspect more as the transaction progresses, with the right access controls and professional support.

If you are selling, the strongest next step is a price discussion based on that packet. If you are buying, compare the asking price with your all-in capital, operating skills and cash requirement before moving into full diligence.

Compare SaaS sale and acquisition routes

The Deal Decision Check remains unresolved when core revenue, rights or costs cannot be established. A high growth rate does not cancel that uncertainty. If the records support the claim, the result is supported by the stated evidence for that claim alone; it is not an approval to buy.

How we researched this page

Deal Field Guide helps you buy or sell an online business with clear eyes — comparing the real routes, showing the money after fees, and telling you what owning each kind of business is actually like, with the source behind every number.

Research note. On October 24, 2026, we read Stripe's recurring-revenue and retention documentation and FE International's SaaS valuation explanation. Provider statements are identified as such. Our earnings bridges, cohort figures and sensitivities are fictional calculations, not private transaction data. Field Notes draw on the publisher's experience running SaaS; we have not bought or sold through the providers covered. We use software, including AI, to help organize research and drafting and remain responsible for the page. See our methodology.

This is educational information, not legal, tax or investment advice. A real valuation and acquisition need professionals who can inspect the business and agreement.

Sources checked October 24, 2026

• FE International: SaaS metrics and valuation — provider's description of valuation approaches; no published multiple range adopted.
• Stripe: Understanding MRR and ARR — monthly normalization, past-due subscriptions and distinction from accounting revenue.
• Stripe: Net revenue retention — same-cohort revenue retention and calculation.
• Stripe: Essential SaaS metrics — definitions and relationships among acquisition, retention, growth and economic measures.