two companies at $1m arr can be worth 5x apart. same revenue, same logo count, same headline number. the difference isn't in the spreadsheet you show investors.
it's in net revenue retention — what happens to revenue after a customer signs. churn means you're refilling a leaky bucket every month. nrr above 110% means revenue compounds on itself. one number, sitting quietly under the arr line, deciding whether the company is worth four times revenue or twenty.
most founders don't know theirs to the dollar. the ones who do don't lose 5x by accident.
what nrr is, and where the bands sit
net revenue retention measures what happens to a cohort of customers over a year. take everyone who paid you anything 12 months ago. now look at what they're paying you today.
nrr = current revenue from that cohort / their revenue 12 months ago
if that group was paying you $100k a year ago and is now paying you $115k (after some upgrades, some downgrades, some churn), nrr = 115%. the thing this captures that gross retention misses: expansion. some customers shrink, some grow. nrr lets you net those out into one honest number.
industry standard reads in 2026:
- under 90%: leaky bucket. you're losing more revenue than you're keeping, and growth comes entirely from new logos. sustainable only if you're absurdly good at acquisition.
- 90–100%: stable. you're keeping what you have. fine if your acquisition is efficient.
- 100–110%: healthy. expansion offsets churn. most public saas companies live here.
- 110–130%: excellent. revenue compounds on the existing base. this is where 5x multiples live.
- over 130%: extraordinary. snowflake-class. usually only seen in usage-based pricing or companies where existing customers expand rapidly.
the median for public saas sits around 105-110% in 2026, down from 115%+ in 2021. the chart below tells you why the difference between 90% and 120% isn't a rounding error.
why the gap between two $1m arr companies is 5x
let's run the math. two companies, same starting point: $1m arr, both growing top-line at 100% year-over-year by adding the same volume of new logos.
company a has nrr at 90%. company b has nrr at 120%.
after three years, both have added the same gross volume of new logos. but company a lost 10% of existing revenue each year. it's burning new logos to replace churn before any of the new growth shows up. company b grew existing revenue 20% per year, in addition to new logos.
projected arr in year 3 lands at $5.8M for company a and $11.2M for company b. at typical 2026 revenue multiples (4-8x for healthy saas), that's a difference of $22M to $44M in enterprise value from the same starting line. nothing else changed. no different sales motion. no different product. one of these companies just stopped leaking and the other didn't.
this is why investors lead diligence with the nrr question and why founders who don't know their nrr by heart get marked down before the conversation even gets to product.
what nrr actually catches
three failure modes that arr alone can't see.
failure mode 1: the wrong customers. you're closing logos that don't expand or stay. the headline is fine, the retention isn't. this usually shows up as nrr in the 80s with great new-logo growth. founders mistake this for traction. it's not.
failure mode 2: pricing left on the table. if your nrr is at 100% it doesn't mean you're keeping everyone. it might mean expansion exists but you're not charging for usage growth. nrr below 110% in a usage-based saas is a pricing problem, not a retention problem.
failure mode 3: cohort decay. older cohorts are leaving but new cohorts are bigger, masking it. the rolling 12-month nrr stays flat but the underlying behavior is getting worse. the only way to catch this: look at nrr by cohort, not just blended.
nrr is the single metric that turns a saas business from a thing that makes money into an actual asset.
how to actually move it
most teams attack churn. that's half the picture. the levers, ranked by what actually shows up in nrr.
one: expansion via usage. tier-based or seat-based pricing where customers grow into more revenue without a renegotiation. typically the highest-leverage move.
two: second-product attachment. cross-sell to existing customers. requires sales motion, but the close rate on existing customers is 4-5x the rate on cold.
three: reduce involuntary churn. failed payments, expired cards. dunning, smart retries, card updater services. usually a 1-3% lift in nrr from operational work alone.
four: annual contracts. shifts the churn event from monthly to yearly, lowers the gross churn rate, lets expansion compound between renewals.
five: cut churn from the wrong customers. the boring, hard one. requires saying no to the wrong customers at the top of funnel. founders almost never do this until the board meeting where someone makes them.
the operational truth
most early-stage founders don't actually know their nrr to the dollar. they know logo retention. "we kept 9 out of 10 customers this quarter" is the version they can recite. but logo retention is a vanity number next to net dollar retention. losing one $50k customer and keeping nine $5k customers is great logo retention and terrible nrr.
the calculation is doable from any decent billing system. the data is in stripe. the work is in pulling it cleanly, attributing it to cohorts, and looking at it monthly instead of when an investor asks for it.
how zift handles this
zift computes nrr, mrr, churn, and expansion from your stripe data, every fifteen minutes. on the first of the month you get the number, the delta from last month, and the customer that moved it. no spreadsheet. no manual reconciliation.
if you're a finance lead at a series A team and you need this with multi-entity reconciliation and cohort analysis, zift handles that too.
nrr is the metric that quietly decides whether your $1m arr is worth $4m or $20m. most founders find out which one they had at term sheet, which is the wrong week to find out.