SaaS Valuation Multiples: Use Evidence, Not a Magic Number

SaaS valuation multiples are not a universal price list: choose the denominator, verify recurring revenue, normalize owner labor and costs, then test whether the business can transfer to a buyer. We would value a solo SaaS from defensible ARR or owner earnings and refuse to quote an exact acquisition price when no buyer offer or closed-sale price is disclosed. Cal AI reports $25M/yr net [V], but that verified result alone cannot produce a transaction multiple.

Contents

• ·What SaaS valuation multiples actually measure
• ·Evidence comes before the multiple
• ·Real cases and the inputs a buyer can inspect
• ·Where the popular valuation claim breaks
• ·How to build a defensible valuation
• ·FAQ

What SaaS valuation multiples actually measure

A SaaS valuation multiple is a ratio between company value and a maintainable financial metric, usually recurring revenue or earnings. It is not the valuation itself. The useful calculation is maintainable metric × negotiated multiple; the hard work sits in proving the metric and deciding which risks a buyer will inherit.

Wikipedia's software-as-a-service entry defines the delivery model, while Stripe's SaaS metrics reference separates recurring-revenue metrics from churn and acquisition economics. Those distinctions matter because identical top lines can hide very different retention, support, infrastructure, and growth profiles.

For example, Data Fetcher reports $23K/mo [F], 600 paying customers [F], and an 85% margin [F]. That is materially more useful than a revenue screenshot alone, yet it still lacks a closed purchase price. Treat an asking multiple, an investor valuation, and a completed acquisition as three different facts.

Evidence comes before the multiple

Start by grading the revenue claim, then inspect its quality. ProvenStartups uses a four-class evidence method: third-party verified [V], founder-reported [F], creator-relayed [C], and unverified [U]. A precise number with weak provenance should not outrank a less dramatic number that a buyer can reproduce from billing records.

The full ProvenStartups index contains 406 graded startup ideas, including 266 software or SaaS products and 38 documented cautionary tales. The full matching valuation cohort covers 229 projects; 138 are solo-run, and 86 publish a clean monthly figure. Those are cohort counts, not a market multiple.

Before discussing price, request:

• ·billing exports matched to bank deposits;
• ·churn and cohort retention;
• ·customer concentration and contract terms;
• ·normalized infrastructure, support, and owner-labor costs;
• ·code, domain, data, and account transferability.

Letterly at $250K/mo [C] is an impressive relayed claim, but its evidence class tells a buyer what still needs verification. ProvenStartups would not quietly promote [C] to [V] because the number is large.

Real cases and the inputs a buyer can inspect

These are real operating cases with disclosed revenue, profit, customers, or margins, not hypothetical valuation models. None of the named records below discloses both a closed acquisition price and the normalized metric used by that buyer. We therefore use them to compare valuation inputs and refuse to manufacture SaaS acquisition multiples from missing deal terms.

This table is useful precisely because it does not pretend the figures are equivalent. A single-client retainer, a one-month profit result, a multi-app portfolio, and a mature subscription product need different normalization. The evidence grade says who supplied the claim; it does not certify retention, transferability, or a buyer's willingness to pay.

Where the popular valuation claim breaks

The popular claim is that a founder can take ARR, apply the current SaaS multiple, and call the result a valuation. ProvenStartups data contradicts that shortcut. The records mix SaaS, consumer apps, tools, services, and portfolios, while their disclosures range from verified net results to relayed top-line snapshots.

HabitKit reports $15K MRR [F] with only $200–300/mo in costs [F]. That cost profile may support strong owner earnings, but a buyer still needs retention and dependency data. Conversely, MeetOscar reached $45,000 MRR in 60 days [F]; speed is attractive, but a short history gives less evidence that revenue will persist.

Revenue scale does not erase provenance either. Minea / DropMagic reports Minea peaking at $750K MRR [F] and DropMagic reaching $45K MRR in four months [F]. Those are valuable operating signals, not verified sale comps. Any table that turns them directly into exit prices is adding an undisclosed assumption.

How to build a defensible valuation

Build the valuation as an auditable range, not a single authoritative number. Start with revenue and earnings that survive diligence, document adjustments, then let buyer fit determine the negotiated multiple. This approach will not predict every offer, but it exposes exactly which assumption changes the result and which evidence is missing.

1. 1.Choose the metric. Use maintainable recurring revenue for a genuinely recurring product; use normalized owner earnings when labor and margins dominate.
2. 2.Reconcile it. Match invoices, processor exports, refunds, taxes, and bank deposits.
3. 3.Normalize costs. Add market-rate replacement for founder work and remove expenses that will not transfer.
4. 4.Map risk. Record churn, concentration, platform dependency, security obligations, and code ownership.
5. 5.Separate evidence from negotiation. Build a low and high case, then document why a strategic buyer might depart from either.

Investopedia on exit strategy frames an exit as a planned way to transfer or reduce ownership. The SBA's guide to selling a business covers valuation and sale preparation. We would take reconciled records into that process; we would refuse a headline multiple unsupported by a completed deal.

FAQ

The short answers below keep operating evidence separate from acquisition evidence. That distinction matters more than decimal precision: revenue can be real while the proposed multiple is still speculative. When a required input was not disclosed in the supplied case record, the correct answer is “not disclosed,” not an estimate dressed as a benchmark.

What is a typical SaaS valuation multiple?

No typical multiple is supported by the supplied case data, so ProvenStartups will not invent one. The appropriate range depends on the chosen metric, growth durability, churn, margins, concentration, owner dependence, and buyer fit. A market benchmark can start a negotiation, but only closed deal terms can establish a transaction comp.

Should I value a SaaS on revenue or profit?

Use the metric that best represents transferable economics. Revenue may fit a stable recurring product; normalized profit may fit a lean, owner-operated asset. nano-banana.ai disclosed ≈$115K/mo net profit for a single month [C], but that short period still needs normalization before it can anchor a valuation.

Does verified revenue mean the valuation is verified?

No. Cal AI's $25M/yr net [V] is third-party verified operating evidence, not a verified sale value. A valuation also requires the denominator period, transferable costs, risk adjustments, and either an actual offer or a disclosed transaction price. Evidence grades tell you how much to trust the claim presented, not what a buyer must pay.

Can a very early SaaS be valued?

Yes, but uncertainty should be explicit rather than hidden inside a confident multiple. The AEO Service record has a $2,000/mo retainer from one client [F], so concentration is central. With little history, document contracts, pipeline, delivery labor, and transferability, then present scenarios instead of treating early revenue as mature ARR.

What should a solo founder prepare before selling?

Prepare clean billing and bank records, customer cohorts, churn, contracts, source-code ownership, infrastructure access, support workload, and a replacement-cost estimate for your labor. The matching cohort contains 138 solo-run projects, so owner dependence is not an edge case. The goal is to show which cash flow survives after the founder leaves.