The Retention Threshold
92% gross retention looks good. Above the 90% threshold most boards target.
But customers acquired eighteen months ago churn at 12% annually. The business loses its entire customer base every eight years. Growth depends entirely on new acquisition.
Retention isn’t a strength—it’s a constraint.
The compounding math
The difference between 90% and 95% retention feels small. The math compounds in ways that aren’t obvious until you run it.
90% annual retention: half your customer base turns over every seven years.
95%: half turns over every fourteen years.
97%: half turns over every twenty-three years.
The marginal improvement from 90% to 95% doubles revenue durability.
Customer acquisition is expensive. Retention is cheap by comparison. A business retaining 95% needs to replace 5% annually to stay flat. A business retaining 90% needs to replace 10%.
Same growth rate. Double the acquisition requirement. Double the cash burn.
The threshold
The retention threshold is the point where churn stops constraining growth. For most B2B SaaS, that sits around 95% gross retention.
Below it, you’re running to stay still. Above it, growth compounds.
Gross retention measures revenue you kept before any expansion. Net retention includes expansion revenue.
A business showing 110% net retention with 85% gross retention masks a churn problem with expansion. When expansion slows—and it always does eventually—the churn becomes visible.
Focus on gross retention first. Net retention comes from having a customer base healthy enough to expand into.
Leading indicators
Customer health score trending down. Support ticket volume trending up. Product usage declining in first ninety days.
All three surface months before churn.
Customer health scores only work if they’re honest. Most define health based on what’s easy to measure—login frequency, feature usage, support tickets—rather than what predicts churn.
Work backwards from churn. Pull customers who churned in last twelve months. Look at behaviour ninety days before they left.
What changed? Usage dropped below a threshold? Certain features went unused? Support tickets increased?
That behaviour becomes the health score.
Product usage determines retention
Product usage in first ninety days determines retention for the next two years.
A customer hitting activation milestones in month one retains at 96%. A customer taking four months to activate retains at 82%.
The difference sets before the first invoice.
Support ticket volume is a leading indicator most ignore. A customer who filed zero tickets last quarter and three this quarter is at risk.
Not because they have problems—all software has problems. Because their tolerance is declining.
Diagnose by cohort
Customers acquired through channel X churn at 15% whilst channel Y churns at 6%. Same product, same price, different retention.
The issue isn’t the product. Channel X attracts wrong-fit customers.
Segment retention by acquisition source, deal size, industry, implementation complexity, time to first value.
Cohorts that retain well reveal what good customers look like. Cohorts that churn reveal what to stop doing.
Pricing structure affects retention more than most realise. Annual contracts retain better than monthly. Customers who prepay retain better than those who pay in arrears. Customers on value-based pricing retain better than seat-based.
None of this shows on a P&L, but it compounds over years.
The fix
Understand why customers churn. Change what causes it.
They churn because onboarding takes too long? Fix onboarding.
They churn because product doesn’t deliver value fast enough? Change activation.
Wrong-fit customers buying? Change sales qualification.
The fix isn’t a retention initiative.
What to check
Pull retention data by cohort for last twenty-four months. Segment by acquisition channel, deal size, time to activation.
Retention varies by more than five percentage points across cohorts? Something in acquisition or onboarding is broken.
Check whether customer health score predicts churn. Pull churned customers from last year. Look at health scores ninety days before they left. Most showed green? Your health score is fiction.
Test whether usage in first ninety days predicts retention at eighteen months. Customers who activate fast retain materially better? Activation speed is the leverage point.
The retention threshold separates businesses that compound from businesses that churn.
Five percentage points of retention improvement often matters more than twenty percent revenue growth.
Most boards spend ten times more attention on growth than retention.
Related: Payback Over Ratios · The Unit Economics Test · Reading Guide