40+ Customer Retention Statistics for 2026

How this list is different

Customer retention is the home of the most-quoted statistic in business: "a 5% increase in retention boosts profits 25-95%." We traced it to its origin. It comes from "Zero Defections: Quality Comes to Services" by Frederick Reichheld and W. Earl Sasser in Harvard Business Review, September-October 1990, and the original range was 25-85%, not 25-95%. It is thirty-five years old and behind a paywall, which is why almost nobody quoting it has read it.

The rest of this page is current, verified benchmark data from SaaS Capital, Benchmarkit, and ChartMogul, with every figure traced to its dataset.

Key retention statistics at a glance

• Median B2B SaaS net revenue retention is 102% for mid-market ACVs (SaaS Capital, 2025)
• Acquiring new revenue costs twice as much as expanding existing customers ($2.00 vs $1.00 CAC ratio) (Benchmarkit, 2025)
• Companies with NRR above 100% grow 43.6% per year versus 13.1% for those below 60% (ChartMogul)
• AI-native companies show the weakest retention of any segment: 40% median GRR (ChartMogul, 2025)
• The famous "5% retention" statistic dates to 1990 and says 25-85%, not 25-95%

SaaS retention benchmarks (NRR and GRR)

• Median net revenue retention is 102% for private B2B SaaS companies with ACVs of $25,000-$50,000, with the top quartile at 111% and the bottom quartile at 97% (SaaS Capital, survey of companies with $1M+ ARR, 2025)
• Median SaaS net revenue retention is 101% across the broader benchmark set, which the publisher reads as a sign that retaining and expanding existing customers is getting harder (Benchmarkit, 2024 data, published 2025)
• In a much wider dataset of roughly 2,700 B2B SaaS companies with at least $250k ARR, median NRR was 82% with an upper quartile of 97% (ChartMogul, 2025)
• B2C SaaS retention is far weaker: median NRR of just 49% in the same dataset (ChartMogul, 2025)

The spread between these medians (82% to 102%) is a sampling artifact worth understanding: SaaS Capital and Benchmarkit survey larger, funded companies, while ChartMogul measures every business on its platform including very small ones. Both are real; quote the one matching your segment.

Retention and growth

• SaaS businesses with NRR over 100% grow 43.6% per year on average, versus just 13.1% for businesses with NRR below 60%, roughly 1.5-3x faster growth (ChartMogul SaaS Retention Report)
• The correlation holds in independent data: companies with NRR of at least 110% grew faster than the median, while those below 100% grew slower (SaaS Capital, 2025)

Acquisition vs expansion cost

The modern, verifiable replacement for the 1990 folklore.

• The median New CAC Ratio rose 14% to $2.00 of sales and marketing spend per $1.00 of new-customer ARR (Benchmarkit, 2024 data)
• The median Expansion CAC Ratio is $1.00, meaning acquiring new revenue costs roughly twice as much as expanding existing customers (Benchmarkit)
• The sales-side equivalent: expansion deals win at 45% and close in 52 days, versus 19% and 91 days for new business (Ebsta x Pavilion, in our sales statistics)

Note what this does to the famous claim: real data supports "expansion is roughly 2x cheaper than acquisition," not "retention is 5-25x cheaper." The directional advice survives; the multiplier does not.

The AI churn wave

The most interesting new finding in retention data, and a caution for anyone selling AI products.

• AI-native software companies show the weakest retention of any segment: median GRR of 40%, worse than B2C, and median NRR of 48% (ChartMogul, 2025)
• Retention scales sharply with price tier: AI-native products above $250/month see 70% GRR and 85% NRR, comparable to healthy B2B SaaS
• Products at $50-$249/month see 45% GRR and 61% NRR
• Products under $50/month see just 23% GRR and 32% NRR (all ChartMogul, 2025)

The pattern suggests cheap AI tools are being tried and abandoned at extraordinary rates, while higher-priced AI products embedded in real workflows retain like normal software.

Famous retention statistics, fact-checked

What the verified numbers say together

The retention data tells a coherent story: net revenue retention has compressed to roughly break-even (101-102% at the median for funded B2B SaaS), expansion remains about twice as capital-efficient as acquisition, retention and growth rates are tightly correlated, and a new category (cheap AI tools) is churning at rates the industry has not seen before.

The operational implication is the same one the 1990 article was groping toward with better data behind it now: the cheapest revenue is the revenue you already have, and the systems that watch account health, usage, billing signals, and engagement, are the ones that catch churn while it is still reversible. That monitoring is precisely what a unified customer view and automated health tracking exist to do.

What these numbers do not tell you

Retention benchmarks are unusually easy to misread, for three reasons worth knowing before you quote them.

Medians hide segments. A 102% median NRR covers companies with wildly different contract sizes. Retention correlates strongly with ACV: enterprise contracts churn less because switching costs more. Comparing a self-serve product to a blended median is comparing two different businesses.

Gross and net retention answer different questions. NRR includes expansion, so a company losing customers steadily can still report NRR above 100% if the survivors expand fast enough. GRR is the number that tells you whether customers stay. Quoting NRR alone can hide a leaking bucket for years.

Cohort age changes everything. Retention measured across all customers looks better than retention measured on cohorts, because long-tenured customers are, by definition, the ones who did not churn. A company growing quickly has a young customer base and will look worse on the same underlying performance.

The practical test: before acting on any retention figure, ask what population it covers, whether it includes expansion, and how old the cohorts are.

The takeaway

Verified retention data: NRR has compressed to break-even for most B2B SaaS, expansion is twice as efficient as acquisition, retention predicts growth more reliably than almost any other metric, and cheap AI products are churning spectacularly. The famous statistics that made retention a strategic priority are real but thirty-five years old and routinely misquoted, the modern numbers make the same case with evidence you can actually check.

Where retention is actually won

The CAC ratio is the number to act on: expansion revenue costs half what new revenue costs.

Most teams accept that and then staff the opposite way, because new business is visible and renewals are quiet. The customer who goes silent three months before renewal looks identical to the happy one, and nobody is watching. It is the failure described in re-engaging cold leads, moved later in the lifecycle.

Outsales treats a dormant customer the way it treats a dormant prospect: an open relationship with nobody minding it. It reads the thread, decides when contact is due and writes it. The autonomous CRM model exists for this, where a system of record stores the fact and does nothing with it.

Related reading: sales statistics for the acquisition side of the same equation, CRM reporting for building the retention view, and the Harvard Business Review work on response speed, which applies to renewals as much as to new leads.

Two more worth reading alongside this: lead generation statistics for what new revenue costs to acquire, and CRM ROI for measuring what the system gives back.