every founder deck has the same slide. "ltv/cac of 4.0x." the partner nods. the founder breathes. the room moves on.
that number is six to twelve months old. it was already old when it was computed. and unlike most stale numbers, it gets more flattering the longer you wait to recompute it, because the customers who churned aren't in the denominator anymore. you're not looking at the engine. you're looking at the survivors.
the founders raising series B in 2026 stopped showing the ratio. they show monthly cac payback by cohort and current-month blended cac instead. the partners stopped asking for ltv/cac somewhere around 2024. founders kept volunteering it.
what ltv/cac actually measures
the formula is honest. ltv = arpu × gross margin ÷ monthly churn rate. cac = s&m spend ÷ new customers acquired. divide the first by the second and you have a ratio that says "every dollar of acquisition spend returns four dollars of margin over the lifetime of the customer."
the problem is in the inputs. monthly churn rate is computed from cohorts that are old enough to have a churn rate — usually 9-12 months minimum. s&m spend is whatever you booked last quarter. so the numerator is built from customers acquired before the product worked, and the denominator is built from spend before the channels saturated. the ratio is a story about a company that doesn't exist anymore.
worse, ltv has a survivorship problem baked in. as a cohort ages, the customers who were going to churn already churned. the ones still paying you in month 18 are, by definition, the stickiest ones. their implied lifetime is enormous. they pull the average lifetime up. the ratio climbs. the company looks healthier on paper while the new cohorts behind them quietly churn at twice the rate.
four views of the same company. the first is what's in the board deck. the second is the cohort that signed up six months ago — they haven't paid back yet, and the curve is bending the wrong way. the third is the cac you'd report if your books were closed yesterday. the fourth is the magic number, which is the leading version of the same question.
the survivorship problem, in numbers
a saas company adds 100 customers in january at a blended cac of $2,000. arpu is $200/month, gross margin is 80%. monthly churn is 4%.
at month 6, 22 customers have churned. 78 remain. the implied lifetime of the survivors, computed from their retained arpu, looks much longer than 25 months — because the ones who were going to churn already did. the ratio you compute against the survivors is structurally higher than the ratio you would have computed against the original cohort at month 0.
founders feel rewarded for waiting to compute the number. "let's wait until the cohort matures." by the time it matures, the only customers in it are the ones who like the product. their ltv looks great. it tells you nothing about whether the next 100 customers you acquire will behave the same way. they almost never do.
what to measure instead
three numbers, in the order you should look at them on the first of every month.
monthly cac payback by cohort. not blended. by month of acquisition. january '26's cac payback at month 4, february '26 at month 3, march '26 at month 2. you want to see the curve flattening. if march's payback at month 2 is worse than january's payback at month 2, your acquisition cost is rising faster than your funnel is converting. that's the leading signal. ltv/cac would have hidden it for another six months.
current-month blended cac. total s&m spend in the month divided by net new customers in the month. it's a noisy number. that's the point — it moves with reality. when your linkedin campaign stops working, this number jumps the next week. ltv/cac wouldn't have noticed until the q3 board meeting.
net new arr ÷ s&m, last quarter. this is magic number, and it's the closest thing to ltv/cac that's actually a leading metric. it asks a simple question — for every dollar you spent on s&m last quarter, how much recurring revenue did you book? a 1.2x answer means hire more. a 0.5x answer means stop hiring AEs. ltv/cac of 4.0x with magic number of 0.4x is a company that was healthy. magic number tells you what it is today.
ltv/cac is the most-cited metric in venture decks and the least useful one in venture math — it confirms what you already lived through and tells you nothing about what you're about to.
why the metric persists
investors stopped asking for it years ago. founders keep volunteering it because it's the only ratio they can compute that's reliably above 1.0. that's not a metric. that's a comfort blanket.
the second reason it persists — it's the only ratio that flatters companies whose recent cohorts are bad. a series A founder with deteriorating cohort retention can still report ltv/cac of 4.0x by averaging in the survivors from the 2024 cohort. the math is technically correct. the story is misleading. partners know this, which is why the ones who matter stopped asking.
how zift handles this
zift computes monthly cac payback by cohort, current-month blended cac, and magic number from your stripe + ad platform + payroll data, every fifteen minutes. on the first of the month you get all three, the delta from last month, and the cohort that moved the curve. no waiting for a cohort to mature before you know whether your funnel is breaking.
if you're a finance lead at a series A team running this across multiple segments or geographies, zift handles that too.
the founders who don't get caught by a deteriorating funnel aren't smarter. they just stopped looking at the ratio that gets more flattering the longer they wait.