The standard LTV formula divides average revenue per account by the churn rate, which means the answer approaches infinity as churn approaches zero. Any error in a churn estimate is amplified into the LTV, and every LTV claim is therefore a churn claim wearing a dollar sign.
TL;DR: LTV computed as ARPA divided by churn is arithmetically valid and practically unusable at low churn rates. Here is how to test an LTV claim and rebuild it on a bounded horizon.
Sellers quote LTV because it is the number that justifies acquisition spend, and the formula is genuinely standard. The trouble is structural rather than motivational: the formula assumes a constant churn rate applied forever, and no real business has either. At 2% monthly churn it implies a fifty-month average life; at 5% it implies twenty. The claim is not usually inflated on purpose, it is inflated by a denominator that was already the weakest number in the file.
Four mechanisms account for most of the gap between this claim and what the raw rows show. They are not mutually exclusive and they compound.
Dividing by churn implicitly extrapolates forever. Recompute on twenty-four or thirty-six months of contribution, which is a horizon you can actually observe and underwrite, and the figure usually falls by half or more.
LTV should use contribution after cost of revenue: hosting, payment fees, support and third-party API costs. Payment processing alone is typically a few percent, and support costs are frequently the largest single omission.
Early-tenure churn is almost always much higher than late-tenure churn. Using a blended rate overstates the life of new customers, which is exactly the group the acquisition spend is buying.
Revenue arriving in month forty is worth materially less than revenue in month two. An undiscounted LTV overstates a long-horizon figure precisely where it is least reliable.
Every step below runs on a subscription-level export in a spreadsheet. None of it needs access to the seller's live billing account, which matters, because as a buyer you will not get one.
These are the thresholds we use in our own reports. They are working thresholds rather than industry standards, and the right line for a given deal depends on contract length, tenure and how transferable the customer relationships are.
| What you find | Verdict | What to do about it |
|---|---|---|
| Bounded 24-month margin LTV within 25% of the claim | Green | The claim is conservative and usable. Rare, and worth noting when you see it. |
| Bounded figure is 40–60% of the claim | Investigate | Normal for an unbounded, gross-revenue formula. Recalibrate the acquisition maths on your number rather than theirs. |
| Bounded figure is under 40% of the claim | Price it in | The acquisition spend that the business has been running may not have been profitable. Check whether historical spend was justified by the real figure. |
| Payback period above 18 months | Investigate | Cash-hungry regardless of what LTV says. This is a working-capital question for the first year of ownership. |
| Cost of revenue not available at all | Investigate | LTV cannot be computed on margin without it. Treat any quoted figure as revenue-based and discount accordingly. |
Ask for these before the LOI. After the LOI you are renegotiating rather than negotiating, and a seller who will not produce subscription-level rows has told you something useful either way.
Illustrative. ARPA of $60 and a quoted 2% monthly churn gives $3,000, which is the seller's figure and the formula is applied correctly. Recomputing churn on a paying-only, revenue-weighted base gives 4.1%, which alone takes the naive figure to $1,463. Applying a 78% gross margin gives $1,141. Bounding the horizon at thirty-six months with an observed retention curve, where first-year churn is higher than the blend, gives roughly $780. The claim and the rebuilt figure differ by nearly 4×, and every step between them is a definitional choice rather than a disagreement about the data. At a $1,000 acquisition cost, the first number says spend and the last says stop.
LTV is the most leveraged number in a SaaS deal because it sits downstream of churn and multiplies its error. That makes it a poor primary metric for a buyer and an excellent secondary check: if the seller's LTV cannot be reproduced from your own churn figure, you have found a disagreement about retention, which is the thing you actually care about. Underwrite on payback period and bounded contribution instead, both of which are observable within a horizon you will own.
The relevant tool on this site is the free LTV calculator, which runs the arithmetic above on a file you paste in. The full method is documented in the 5-risk buyer-side method and the due-diligence checklist.
Every check on this page can be run by hand in a spreadsheet, and if you have the time you should. If you would rather not: send us the target's subscription export and we run the full human-reviewed analysis — logo churn, revenue churn, customer concentration, annual-plan decay, zombie MRR and an A–F revenue-quality grade. The free Starter tier covers one CSV per month, which is enough to check a single deal.
See a sample report → · Get the free 23-point checklist →
Because dividing ARPA by the churn rate extrapolates a constant churn rate to infinity, so the answer approaches infinity as churn approaches zero. At 2% monthly churn it implies a fifty-month customer life; at 4% it implies twenty-five. Small errors in the churn estimate become large errors in LTV, and churn is usually the least reliable number in the file.
Gross margin, always, for any decision about acquisition spend. Hosting, payment processing, support and third-party API costs all come out before the customer contributes anything. Using gross revenue overstates LTV by whatever the cost of revenue is, and support cost is the component most often left out.
Payback period and bounded contribution. Acquisition cost divided by monthly gross-margin contribution gives a payback in months, and cumulative contribution over twenty-four or thirty-six months gives a value on a horizon you can actually observe. Both avoid extrapolating a churn rate forever.