LTV:CAC ratio 2026: where 3:1 came from, and why yours is inflated

A founder I worked with last year walked into a board meeting with one slide that said "LTV:CAC 4.8x." The board loved it. Two quarters later the company was burning cash faster than the model predicted, growth had stalled, and nobody could explain why the "efficient" business was so hard up.

I have seen this exact movie enough times to know the ending before the lights go down. The 4.8x was fiction. Not a lie anyone told on purpose, just a number built on revenue that should have been margin, one blended cohort that described no real customer, and a CAC that quietly left out half the sales team.

One thing has changed since I first wrote this page, and it reframes the whole argument. I went looking for where the 3:1 rule everyone quotes actually comes from. It comes from one sentence in one blog post, and that sentence defined LTV on gross margin and handled expansion revenue separately. So the two biggest ways teams inflate the ratio are not loose readings of the rule. They are departures from the only definition the number 3 was ever set against.

Where the 3:1 rule actually comes from

The 3:1 benchmark traces to David Skok's SaaS Metrics 2.0, published on the For Entrepreneurs blog while he was a partner at Matrix Partners (For Entrepreneurs). The operative line is a guideline, not a finding: LTV should be about 3x CAC for a viable recurring-revenue business. That single sentence is the root of nearly every 3:1 citation since.

Two things about that are worth knowing before you hold yourself to the number.

The first is that nobody has ever produced the dataset. Blake Bartlett at OpenView said so in print in June 2016: "Why is a 3x LTV:CAC ratio the appropriate benchmark? No one knows." His answer was "It just is" (OpenView). Ten years on, the number still appears in board decks and investor memos as though it were derived from a broad sample. It is a rule of thumb that stuck. The write-ups that trace its history reach the same conclusion (The SaaS CFO).

The second is that the guideline came with conditions attached, and the conditions are the part that got dropped. It was written about mature companies at steady state: a stable churn rate, a realistic multi-year lifetime, and a payback period comfortably inside twelve months (Average CAC). It was never a seed-stage test. It gets applied as one constantly.

Here is the part that matters most for your own number. Skok's LTV was never revenue. The definition uses gross margin: ARPA times gross margin percentage, divided by churn (ChartMogul). And his advanced version handles expansion revenue as its own explicit term rather than folding it into the base (For Entrepreneurs definitions).

Sources disagree on the year the series first appeared, with write-ups placing it at 2010 and at 2013 and the page itself carrying later updates, so I am not going to pin a date on it. The wording and the definition are consistent across every account of it I could find.

What LTV:CAC is supposed to tell you

Lifetime value is what a customer is worth to you over the whole relationship, counted in gross profit. Customer acquisition cost is the fully loaded cost of landing them. Divide the first by the second and you get a rough read on whether your growth engine makes money or sets it on fire.

That framing is fine as a starting point. The problem is that four separate inputs feed the ratio, and each one has a default calculation that makes your business look better than it is. Stack all four and a real 1.8x can read as a 5x on a board slide. I have audited the models. It happens more than anyone admits.

Where the ratio gets inflated

There are four honest mistakes. I call them honest because smart people make them with clean intentions. They are still the reason your number lies.

Revenue instead of gross margin

This is the big one. Most teams calculate LTV off revenue. A customer paying $100K a year for three years becomes $300K of "lifetime value." But you do not keep $300K. You keep whatever is left after the cost of serving them: hosting, support, the customer success manager, third-party API fees, payment processing.

Run that same customer through the arithmetic at a $60K fully loaded CAC and watch the ratio move without anything about the business changing:

Same customer, same spend, three defensible-looking numbers, and the top and bottom rows point at opposite decisions.

Which margin you use is not a rounding question in 2026, and there is a genuine contradiction in the published data that is worth resolving rather than picking a side of. Median software gross margin is still 80% and has been stable for four years, so at the median AI infrastructure cost has not compressed it at all (Aleph, 2026). Meanwhile gross margin on AI products specifically is running around 52% in 2026, up from 41% in 2024, and pure application-layer AI products sit near 45%, up from 33% (ICONIQ State of AI, January 2026).

Both are true, and the gap between them is the trap. Company-level blended margin moves slowly, because the AI line is still small for most companies. Product-level margin has already moved. So a company at 80% blended margin shipping an AI feature at 52% will see nothing wrong in its blended number for a year or two, and will be pricing acquisition against an LTV that is wrong for the part of the business it is growing fastest. If you are running any real inference cost, the margin input belongs in the cohort split, not above it.

CAC that leaves out the expensive people

The second inflation lever is on the denominator. Ask most teams for their CAC and they will hand you ad spend divided by new logos. That number is almost always too low.

Real CAC includes the fully loaded cost of acquisition. Ad spend, yes, but also the salaries of your SDRs and AEs, the marketing team's comp, sales tooling, event budget, agency retainers, and the ramp cost of reps who have not closed anything yet. In a sales-led B2B motion, headcount is the biggest line by far.

You will see a claim that leaving people out understates CAC by 40% to 60%, or by 40% to 80%. I could not find a primary study behind either range, so I am not going to quote one as a fact. You do not need it. The benchmark you are measuring yourself against settles the question on its own: the $2.00 new CAC ratio is total sales and marketing expense divided by new ARR (Benchmarkit, 2026). Every credible external benchmark is built that way. So an ad-only CAC is not a stricter read of the same metric. It is a different metric, and comparing it to a fully loaded benchmark tells you nothing except that you are ahead, which you are not.

One blended cohort that describes nobody

Here is the mistake that fools even careful teams. The standard LTV formula assumes every customer behaves the same. They do not.

Take two cohorts. Early adopters pay $300 a month and churn at 2%. Newer self-serve customers pay $150 a month and churn at 8%. Assume 80% gross margin on both and an equal number of customers in each, because otherwise the blend depends on the mix, which is itself the point. Those are assumptions, not data, and I am writing them out so you can swap in your own.

• Cohort A: $300 times 0.8, divided by 0.02, is $12,000. Average lifetime 50 months.
• Cohort B: $150 times 0.8, divided by 0.08, is $1,500. Average lifetime 12.5 months.
• Blend the inputs first: $225 ARPA at 5% churn gives $3,600.
• Average the two LTVs instead: $6,750.

Cohort A is worth eight times cohort B. The two ways of producing "the company LTV" are nearly 1.9x apart from each other, and neither one lands near either real cohort. Now divide all four by a $3,000 blended CAC and you get 4.0:1, 0.5:1, 1.2:1 and 2.25:1. One business, four ratios, and only one of them clears the 3:1 line. In reality CAC differs by cohort too, which widens the spread rather than narrowing it.

Acquisition CAC measured against expansion LTV

The last one is subtle and it is how good teams accidentally cheat. Your LTV number usually includes expansion: upsells, seat growth, tier upgrades over the customer's life. Your CAC number only covers what you spent to land the logo. So you are crediting acquisition spend with revenue that expansion earned.

This one used to be an argument. It is now measurable, and the numbers are blunt about it. Expansion ARR is around 40% of all new ARR, and above $50M ARR it is more than half. Winning it costs an expansion CAC ratio of $1.00 against $2.00 for new ARR (Benchmarkit, 2026).

Read that twice, because it cuts in two directions. If expansion is 40% of your new ARR and you have folded its revenue into LTV while leaving its cost out of CAC, you have not made a small accounting slip. You have credited the more expensive motion with the output of the cheaper one. And when you separate them properly, you usually find the cheaper motion is the one worth funding, which is a decision you could not see before.

There is a reporting artefact in the same data worth knowing about. Blended CAC ratios improved about 12% year over year, but the new CAC ratio held at $2.00. The blend got better because expansion grew as a share of the mix, not because acquiring a new logo got cheaper. If your own blended efficiency improved last year, check whether you earned it or whether your mix shifted underneath you.

What good actually looks like in 2026

Here is what is actually measured, with the sample behind each number.

Note what the median is: 3.2:1, which sits almost exactly on the rule of thumb. That is the strongest thing anyone can say for the 3, and it is worth saying. A guideline written about mature companies in the early 2010s turns out to describe the middle of the market in 2026. It is still not a derivation, and it is still a median across segments whose LTVs differ by more than an order of magnitude.

On the by-stage ladder you will find on every roundup, under $2M ARR aim for 2:1 to 3:1, $2M to $10M aim for 3:1 to 4:1, above $10M aim for 4:1 or 5:1: I could not find a dataset behind it. Every page that publishes those bands publishes them without a citation. So I will give you my own version and label it as what it is, which is judgment from running these audits rather than a measurement. Early on I do not care about the ratio at all, because CAC is noisy and the lifetime is a guess. From the point where churn is stable enough to trust, I want the ratio above 3:1 per segment and I want to know which segment is carrying it. Above $10M ARR with a mature motion, a ratio still stuck at 3:1 usually means churn is eating LTV rather than that acquisition is efficient.

And there is a trap in chasing a high ratio, which I want to be blunt about because it is also judgment rather than data. A 6:1 or 7:1 usually means underinvestment. You have found something that works and you are being too timid to fund it. Sitting on a beautiful ratio while a competitor takes the segment is a slow way to lose.

A note on provenance. Every source above was read through search result extracts rather than a direct fetch of the publisher's page, because the egress proxy on this machine blocks those domains. Treat the figures as reported rather than verified first hand. The Skok attribution and the exact wording of the guideline were corroborated across three independent search passes, the Bartlett quote across two, and the $2.00 and $1.00 CAC ratios across two. Confirm anything headed for a board deck at the source.

The number I actually watch instead

Here is my real opinion after ten-plus years doing this. LTV:CAC on its own is a vanity metric. It measures theoretical lifetime money against cash you have already spent, and the lifetime part can take three or four years to show up. A cash-constrained company can die waiting for a gorgeous LTV to materialize.

So I never look at the ratio alone. I pair it with CAC payback: how many months of gross-margin revenue it takes to earn back what you spent to acquire the customer. The B2B SaaS median is 16 months on full-year 2025 actuals, down from an 18-month peak in 2024 and the first improvement in four years. The healthy line most investors still hold is 12 months or under. I took the whole reversal apart in CAC payback period, including why the benchmark improved while most companies' own numbers did not.

Given the choice, I will take a 2:1 ratio with a 6-month payback over a 5:1 ratio with a 36-month payback every time. The first one recycles cash fast enough to fund the next customer. The second one looks brilliant on a slide and quietly bleeds your runway. Use both numbers as a filter: payback under 12 months AND LTV:CAC above 3:1, per segment. Either one alone will let a false positive through.

How to build this so it holds up

The reason most teams calculate LTV:CAC wrong is not that they are bad at math. It is that the inputs live in five different places and nobody owns stitching them together. Revenue sits in the billing system, churn is in the CRM, sales comp is in a spreadsheet, and marketing spend is in the ad platforms. To get an honest ratio you have to pull all of it into one model, by segment, and keep it current. That is a RevOps job, not a finance-once-a-year job.

Here is the sequence I run when a client asks me to fix their unit economics reporting.

The payoff is not a prettier board slide. It is that you finally know which segment to feed. When the enterprise cohort shows a 5:1 ratio with a 9-month payback and the self-serve cohort shows 1.4:1 with a 22-month payback, the budget decision writes itself. You stop pouring acquisition spend into the leaky segment and back the one that compounds. That decision is worth more than the metric itself.

Step 01 is the one people skip and the one that costs the most. If your team cannot agree whether ARR includes services, or whether a logo counts on signature or on go-live, the ratio will move every time somebody rebuilds the model. We wrote the definitional layer up separately in ARR vs bookings vs revenue, because it sits underneath every number on this page.

None of this works if the underlying data is dirty either. If your CRM has half your closed-won deals missing an amount or a segment tag, the model inherits that mess. Cleaning that up is the boring prerequisite nobody wants to fund, and it is exactly where I start. A model built on bad CRM data will lie to you with more confidence than a guess.

For the rest of the cluster: the SaaS magic number uses these same honest inputs at company level. If retention is your weak input, start with net revenue retention, because churn is the single biggest lever on the LTV side of this equation. And if expansion turns out to be carrying your ratio, the motion behind it is worth building deliberately rather than hoping for, which is what the land and expand playbook is for.

Frequently asked questions

Where does the 3:1 LTV:CAC rule come from?

From David Skok's SaaS Metrics 2.0 on the For Entrepreneurs blog, written while he was a partner at Matrix Partners. The line says LTV should be about 3x CAC for a viable recurring-revenue business, and it was offered as a guideline rather than a finding from a dataset. Blake Bartlett at OpenView made the point directly in 2016: nobody knows why 3 is the right number. It is worth knowing that the original definition used gross-margin LTV and treated expansion revenue separately, which is not how most teams calculate the ratio they compare to it.

What is a good LTV:CAC ratio for B2B SaaS?

The measured median across 939 B2B companies is 3.2:1, close enough to the rule of thumb to keep using 3:1 as a floor. Above that, the useful answer is per segment rather than per company: segment LTVs in that sample range from $15K for SMB to over $1M for enterprise, so a company-wide ratio can hide a broken segment completely. A ratio above 6:1 usually means you are underinvesting in growth rather than winning, though that reading is judgment rather than a measured finding.

How do you calculate LTV correctly?

Use gross-margin contribution, not revenue. The formula is average revenue per account times gross margin percentage, divided by your churn rate. So $300 monthly revenue at 80% margin and 2% monthly churn gives you $240 divided by 0.02, or $12,000 in LTV. Do it per cohort, not company-wide, and split gross margin by product line if any of your revenue carries AI inference cost, because AI product margins are running near 52% against 80% for software at the median.

Should CAC include salaries?

Yes. Fully loaded CAC includes SDR and AE salaries, marketing comp, tooling, events, and agency fees, on top of ad spend. In a sales-led motion the people are the largest cost. The reason it matters is comparability: every external benchmark, including the $2.00 new CAC ratio, is built on total sales and marketing expense, so an ad-only CAC cannot be measured against any of them.

Why is LTV:CAC alone not enough?

Because it measures theoretical lifetime money against cash you have already spent, and the lifetime part can take years to arrive. A high ratio with a long payback can still starve you of cash. Pair the ratio with CAC payback period, currently at a 16-month B2B SaaS median, and treat both as a filter: payback under 12 months and ratio above 3:1.

How often should we recalculate it?

Monthly, by segment, from live data. The once-a-year version that lives in a finance spreadsheet is already stale by the time the board sees it. When it updates itself from your billing system and CRM, you can actually steer acquisition spend with it instead of just reporting it.

The takeaway

LTV:CAC is worth calculating. It is just worth calculating the way the rule you are measuring against defined it, which almost nobody does. Use gross margin, split it by product line, load every acquisition cost, split your cohorts, give expansion its own denominator, and never look at the ratio without payback next to it. Do that and the number stops being a slide you defend and starts being a decision you can trust.

If you want a second set of eyes on yours, that is the kind of work we do at Ziel Lab. We would rather hand you an ugly true number than a pretty false one.