Subscriber Acquisition Cost for SaaS: What Actually Counts (Free Trials, Dunning, and Expansion Revenue)

Subscriber acquisition cost (SAC) is the fully loaded cost of getting one new paying subscriber: total acquisition spend divided by new paying subscribers in a period. It uses the same math as CAC, but three subscription-specific decisions, free trials, dunning, and expansion revenue, determine whether your number is honest or inflated.

Get one of those three wrong and you're not computing SAC, you're computing a nicer-looking number that happens to share its name.

Updated August 2026.

Subscriber Acquisition Cost (SAC) vs. CAC: What's Actually Different

Nothing in the core formula changes between SAC and CAC. What changes are three decisions a plain CAC glossary never forces you to make: whether a free trial start counts as an acquired subscriber, whether a dunning recovery counts as a new one, and whether an expansion upgrade counts at all.

The standard SAC vs. CAC framing defines SAC generically as the cost of acquiring a new subscriber, without saying which of those three decisions belong in the count. That gap is exactly where founders get the number wrong, usually not out of carelessness, but because nobody ever told them the formula had a decision to make in the first place.

Worth noting: the three pages ranking #1-3 for this query, Amplitude, Simon-Kucher, and Paddle, never use the term “subscriber acquisition cost” at all. They rank on generic customer acquisition cost content anyway. That's evidence the two terms get used interchangeably in practice, which is exactly why a definitive answer beats another glossary entry. If even the top three results won't commit to the term, don't expect your invoicing software to sort this out for you either.

Want one subscription-metrics breakdown like this in your inbox every week? Subscribe, that's what we send.

The SAC Formula, and Why the Denominator Is Where Founders Get It Wrong

SAC equals total acquisition spend divided by new paying subscribers acquired in the period. That's the whole formula. Genuinely, that's it, no asterisk. Mailchimp's SAC formula states a version of this same ratio. Baremetrics' CAC breakdown states another, and the numerator side is genuinely simple.

The numerator holds three standard components: paid ad spend, sales or founder time on acquisition calls (amortized across the period), and acquisition-tied tooling cost. For the levers that actually lower each of these, see the guide on reducing SaaS CAC. This piece is about getting the number right, not lowering it, so the numerator gets one paragraph and no more.

The denominator is where the real content lives. Counting a trial signup, a dunning recovery, or an expansion upgrade as a “new subscriber” distorts SAC, and each mistake pulls in a different direction: trials make it look artificially cheap, dunning double-counts spend against a customer you didn't newly acquire, and expansion blends near-zero-cost growth into an acquisition number. Sections below handle each one on its own, and the decision grid gives you the full rule set as a single table.

One honest concession before you build a dashboard on this number: even a perfectly loaded SAC is a lagging indicator. It tells you what last period's acquisition cost you, not what this period's will cost, because spend and conversions both take time to settle. Build your dashboard around that limitation, not against it.

Do Free Trials Belong in Your SAC Denominator?

No. Count paid conversions, not trial starts, unless you're deliberately tracking trial-acquisition cost as its own separate, earlier-funnel metric.

The mechanism is simple: trial-to-paid conversion is always well under 100%, so counting trial starts as “acquired subscribers” divides the same spend across a bigger denominator. SAC looks artificially cheap right when the business needs the real number most, during early-funnel tuning. DealHub's own glossary entry flags this ambiguity directly, noting that some companies count only paying subscribers while others include trial users who convert within a window; it doesn't resolve which is right, but the admission itself confirms this is a real, unsettled gap in most SAC guidance.

The fix is two metrics, not one. Track Trial Acquisition Cost (spend divided by trial starts) as a leading indicator of top-of-funnel efficiency, separately from SAC (spend divided by paid conversions) as the lagging outcome metric. Don't blend them into one number; a business can have cheap trial acquisition and expensive real acquisition at the same time, and collapsing the two hides that.

One contrast worth naming: stay.ai, writing for eCommerce subscription boxes, recommends removing free trials from the acquisition program entirely. That logic doesn't transfer to a SaaS free-trial funnel, where the trial itself is usually the primary activation mechanism, not a discount tactic. Copy that advice from an eCommerce playbook and you'd be cutting the exact thing that gets people to pay. For trial-to-paid conversion benchmarks and the freemium-vs-trial decision itself, see the freemium vs. free trial breakdown; this section is only about which number belongs in your SAC denominator.

Dunning and Failed-Payment Churn: The Subscription-Specific SAC Trap

SAC math breaks the moment a subscriber's card fails, because most founders never decide whether a subscriber recovered through dunning counts as a new acquisition.

It doesn't. A dunning-recovered subscriber was already counted once, in the denominator, the month they first converted. Folding the recovery back into SAC double-counts that subscriber's original acquisition spend against a customer you didn't newly acquire; the spend didn't repeat, so the count shouldn't either. It sounds obvious written out like that, and it still trips people up the first time a recovered subscriber shows up in their dashboard.

This section won't re-teach dunning setup, that playbook lives in the dunning and payment-recovery guide. What matters here is the mechanism, not the tactics: as documented by Stripe, Smart Retries defaults to 8 payment-retry attempts within a 2-week window, configurable to windows between 1 week and 2 months. Stripe skips retries entirely on hard-decline codes like lost or stolen card, and on cases where no payment method is available. No recovery-rate figure is published on that documentation, so none gets attached here; the mechanism is the citable fact, not a percentage.

For the decision grid: dunning-recovery cost belongs in retention-ops cost, never in the SAC numerator or denominator. It's a retention win, not an acquisition event, and mixing the two categories makes both numbers less useful.

Expansion Revenue and Referral Signups: What NOT to Fold Into SAC

Two of the most common SAC-distorting mistakes point in opposite directions: counting expansion as new acquisition, and blending free referral signups into paid-channel SAC.

Expansion first. An existing subscriber upgrading their plan is not a new acquisition. Counting that upgrade as a “new subscriber” in the denominator makes SAC look cheaper than it is, because expansion revenue carries near-zero incremental acquisition cost, you didn't spend anything new to get it. Paddle's general rule that only new customers count toward CAC backs the same logic here. For how to build and track expansion revenue on its own terms, see the SaaS expansion revenue playbook; it doesn't belong in your SAC math at all.

Referral works in the opposite direction. If organic or referral signups, which carry near-zero acquisition cost, get blended into the same denominator as paid-channel signups, blended SAC understates true paid-channel cost. That can lead a founder to under-invest in a paid channel that's actually working, because the blended number makes it look worse than it is. The fix: compute SAC per channel, at minimum splitting paid from organic and referral, rather than reporting one blended figure. Two minutes in a spreadsheet, and the paid channel gets judged on what it actually did instead of what averaging did to it.

Once SAC is computed cleanly, the next question is what to do with it, pairing it against LTV, which is covered in the LTV to CAC ratio piece rather than re-derived here.

The SAC Line-Item Decision Grid

Ten line items, one call each. This is the single reference worth screenshotting.

Screenshot it, then watch what it does to an actual SAC calculation next.

A Worked Example: SAC for an $8K MRR Solo SaaS

Say a solo-founder SaaS around $8K MRR spends $4,000 in a month across paid search, paid social, one acquisition tool, and founder time on sales calls. In that same period: 100 people start a free trial, 25 convert to paid, and one of those 25 is a subscriber who churned involuntarily and got recovered through dunning mid-period, meaning they were already counted once, at their original conversion.

Same $4,000 of spend, three different denominators, three different answers:

Corrected SAC ($160) is 4x the naive number ($40), a 300% increase, driven entirely by the denominator shrinking from 100 to 25. Spend didn't move. The naive number understates true SAC by 75%, since $40 is only a quarter of $160. The dunning adjustment then moves SAC from $160 to $166.67, a further roughly 4% increase, purely from excluding one subscriber who was already counted once.

That gap exists entirely because of which subscribers got counted, never because spend changed. That's the whole argument for reading Sections 3 through 5 before you trust any SAC number you compute. Trust the shortcut version instead, and you'll end up building a dashboard on a number that was never real to begin with.

Frequently Asked Questions About Subscriber Acquisition Cost

What is the formula for subscriber acquisition cost?

SAC equals total acquisition spend divided by new paying subscribers acquired in the period. The formula is identical to CAC; what changes is which subscribers you're allowed to count in the denominator, covered above for trials, dunning, and expansion.

What's an example of calculating subscriber acquisition cost?

See the worked example above: $4,000 in monthly spend divided by 25 paid conversions gives a corrected SAC of $160, versus a naive $40 if trial starts get counted instead. The gap comes entirely from the denominator, not the spend. Those figures are illustrative arithmetic, not a benchmark.

Is subscriber acquisition cost the same as customer acquisition cost?

Functionally, yes, same formula, applied to a recurring-billing relationship. The difference that matters is practical: a subscription business has to make explicit calls on trials, dunning recoveries, and expansion revenue that a one-time-sale CAC calculation never has to make.

What is a good SAC for a SaaS company?

This piece deliberately doesn't set a benchmark number; getting the calculation right comes before comparing it to anyone else's. For segmented SaaS CAC benchmarks by ACV tier, see the CAC benchmarks breakdown.

Can SAC be negative?

Not in the literal formula, since spend and subscriber count are both non-negative, but “negative CAC” is a real term for growth loops where referral or viral acquisition brings in paying subscribers faster than paid spend accounts for them. In that case, computing SAC per channel (as recommended above) shows the paid channel's real cost instead of a misleadingly low blended number.

Get the Next SaaS Metrics Breakdown

We break down one subscription metric like this every week, with the mechanism, a decision grid, and a worked example, not just a formula restated. Next up: what actually belongs in your LTV calculation before you pair it against SAC. Subscribe to get it when it publishes.