It is the quarterly board meeting. The deck has a single tile in 48pt: LTV:CAC = 3.5. The slide title reads “healthy unit economics.” Halfway through the call, the CFO asks why the line of credit is being drawn down again if the unit economics are healthy. Nobody on the team has a clean answer, because the answer is not in the tile.
The LTV-to-CAC ratio is a fine snapshot. It is a poor operating metric. Underneath the ratio sit three numbers that decide whether the business actually has working capital next quarter, whether the next acquisition dollar pays, and whether the lifetime number you are dividing by is one you can borrow against. This post is about reading those three numbers, and about why the headline ratio (and especially the 3:1 heuristic) earns less weight than it gets.
The formula, briefly
The ratio puts lifetime value in the numerator and customer acquisition cost in the denominator. LTV is the cumulative gross contribution a customer generates over a chosen time window: revenue minus COGS, shipping, payment fees, and returns, summed over months 1 through N. CAC is total acquisition spend in a period divided by new customers acquired in that period. The two glossary entries cover the mechanics; this post is about what the ratio of those two numbers does and does not tell an operator.
The shorthand version of the ratio’s verdict is the one most people have heard: 3:1 is healthy, below 1:1 is on fire, above 5:1 means you are underspending on growth. The shorthand has the shape of a constant of nature. It is not one.
Where the 3:1 number came from
The 3:1 LTV:CAC heuristic crystallized in mid-2010s SaaS-metrics writing, the David Skok / Bessemer-orbit body of work that gave the industry most of its unit-economics vocabulary. It travels with a specific set of implicit assumptions. SaaS gross margins typically sit at 70–90%. Revenue is contractual and recurring, with a cancellation event that makes churn observable. Customer lifespan is measured in years, and monthly churn in low single digits. Under those conditions, a 3:1 ratio describes a business that earns back its CAC quickly in cash terms and then compounds for years on near-pure-margin recurring revenue.
DTC does not look like that. Gross margins typically sit in the 20–50% range, varying widely by category. Apparel, supplements, and consumer electronics are not the same business. There is no cancellation event, so customer lifespan is inferred from observed repeat behavior rather than contracted. Repeat purchase is a probability, not a subscription. The mechanical ratio number that worked for SaaS does not transfer to a brand where every dollar of “lifetime value” is itself an estimate, where the margin on that dollar is half what SaaS earns, and where the cash takes longer to come back. Reusing the SaaS target as if it described DTC unit economics quietly imports assumptions that do not hold.
Payback period is the cash constraint
The number the ratio hides, and the one that actually constrains a DTC business week to week, is payback period: the number of months until cumulative gross profit per customer equals CAC. Two brands can sit at identical LTV:CAC ratios and live in completely different operational realities depending on how quickly the cash comes back.
A brand running a 4:1 ratio with a 24-month payback period is profitable on paper. It is also, mechanically, financing every new customer it acquires for two years before the cash returns. If the acquisition volume is growing, which is the whole point of healthy unit economics, the working capital required to fund that growth grows faster than the cash returning from older cohorts. The brand can show a beautiful ratio in the deck and run out of money paying for ads before the LTV arrives. A brand running a 3:1 ratio with an 8-month payback is less impressive on the slide and more durable on the bank statement: the cash recycles fast enough to fund the next cohort.
6–9 months is a common practitioner target for DTC payback when the brand has no external working-capital line. Longer paybacks are workable (many brands run at 12 or 18 months) but require committed capital sized to the gap: debt, equity, or supplier terms. The ratio is silent about which of these worlds you are in. Payback is the question to ask after the ratio.
Marginal CAC vs average CAC
The CAC in LTV:CAC is an average: total acquisition spend in a period over total new customers in that period. Spend decisions, though, happen at the margin. The number that matters for the next ad dollar is what that dollar gets you, not what the blended history says.
A worked example. A brand acquires 1,000 customers in a quarter at a $40 blended CAC. The first 800 came from saturated, well-optimized prospecting at roughly $30 each. The last 200 came from expansion audiences (broader lookalikes, new placements, channels with thinner conversion data) at $90 each. The blended number averages to a healthy $40, and the LTV:CAC ratio computed against it looks fine. But the marginal customer cost $90, and the next 100 customers will cost more than that. The ratio describes a business that has already passed the point where the next acquisition push pays.
Most dashboards default to period-aggregated CAC, which makes marginal CAC hard to see without separate cohort or channel-level views. The discipline is to ask, at the moment of the spend decision, what did the last cohort actually cost, and what is the trajectory. The blended ratio will not warn you when the answer is breaking.
”Lifetime” is a window choice
“Lifetime” in LTV is a measurement window, not a property of the customer. A 12-month LTV, a 24-month LTV, and a five-year extrapolated LTV are three different numbers from the same data, and each is correct for a different decision.
A 12-month LTV is observed cash. Twelve months of repeat behavior have actually happened; you can borrow against the number, and a working-capital lender will. It is the right input for cash planning. A 24-month LTV is partly observed and partly modeled, depending on cohort age. It is the right input for channel-level scaling decisions where you are pricing a marginal cohort against its expected contribution. A five-year extrapolated LTV is mostly modeled, projected from early-cohort repeat curves under retention assumptions that may or may not hold. It is fine for valuation conversations, where the tolerance for assumption error is higher. It is dangerous as a spend-gating input, because using it bets cohorts on retention behavior nobody has yet observed.
The cargo-cult version of this is picking the longest horizon the model will produce, dividing by CAC, and reporting the resulting ratio. The ratio will always look better at five years than at twelve months. It will also always be the least defensible of the three. The grown-up version is to name which LTV horizon answers which question, and use the matching one.
Track all three
None of this makes the LTV:CAC ratio useless. It is a fine snapshot for the deck, a quick sanity check against catastrophic unit economics, and a number that summarizes a lot of work into a single tile. The mistake is to operate from it.
The operating triangle is the ratio, payback, and marginal CAC, read together. The ratio is the snapshot: is the business roughly viable on a unit basis. Payback is the cash constraint: can the business afford to grow at its current pace without choking on working capital. Marginal CAC is the next-spend gate: does the next dollar pay. A brand whose ratio looks fine but whose payback is stretching and whose marginal CAC is rising is a brand whose deck is lying to it, even if the lie is technically true.
The instinct here is the same one we wrote about in reading ROAS, AOV, and CAC together when planning ad spend: ROAS in isolation hides margin, just as LTV:CAC in isolation hides payback. The single-number reflex is comforting because it produces a verdict. The joint reading produces the question, which is what an operator actually needs. Treat any single number as a starting point for a question, not a conclusion. The LTV:CAC ratio, for all its board-deck gravity, is no exception.