Methodology
How ChurnCost.com sources and verifies benchmarks, derives the three-layer cost formulas, and refreshes the dataset. The audit goal is reproducibility: every number on the site should be re-derivable from a named primary source.
Sources
Every benchmark on the site traces back to one of the publishers below. Where two sources publish different numbers for the same metric, the page flags the divergence inline and shows both ends of the range.
| Publisher | Cadence | What we use it for |
|---|---|---|
| Bessemer Venture Partners - State of the Cloud | Annual | Median and top-quartile NRR / GRR by ACV band, annual cohort retention curves, public-SaaS reference data through the BVP Cloud Index. Anchors the NRR-benchmarks page and the segment-band rows on /benchmarks. |
| OpenView Partners - SaaS Benchmarks Report | Annual | Cross-segment retention, growth, and efficiency benchmarks pulled from OpenView's annual operator survey. Used as the secondary triangulation source against Bessemer State of the Cloud where the two reports overlap. |
| ChartMogul - SaaS Subscription Index | Quarterly (public summaries) | Aggregate churn, expansion, and contraction rates across ChartMogul's active customer base segmented by ACV. The public summary publishes median and top-quartile bands; the underlying dataset is proprietary. Used on /benchmarks and /nrr-benchmarks. |
| Recurly - Subscription Economy reports | Annual / event-driven | Failed-payment behaviour, smart-retry recovery rates, account-updater performance, and dunning recovery economics. Anchors the involuntary-churn page (9 percent of MRR figure, 20 to 40 percent of total churn share, 60 to 70 percent smart-retry recovery). |
| ProfitWell / Paddle - SaaS retention research | Rolling | Median B2B SaaS monthly churn rate (the 3.5 percent figure used on the home page), pricing-page churn-elasticity research, and expansion-revenue contribution bands. Cross-checks ChartMogul on overlapping metrics. |
| KeyBanc Capital Markets - Private SaaS Company Survey | Annual | Formerly the Pacific Crest SaaS Survey. Operator-level NRR, GRR, churn, and efficiency benchmarks across a few hundred private SaaS companies. Triangulates Bessemer / OpenView on the NRR median. |
| SaaS Capital - Churn-and-Valuation research | Annual / event-driven | Public quantification of how churn rate and NRR map to exit multiple in private-market SaaS transactions. Anchors /valuation-impact and the 0.5 to 1.5 multiple-compression-per-1-percent-excess-churn band. |
| Custify / Vena - SaaS vertical churn research | Annual | SaaS-vertical churn rate publications used for /saas vertical bands (fintech, healthcare, martech, devtools). Used as a tertiary source against Recurly + ProfitWell where vertical breakdown is needed. |
| Public 10-K and 10-Q filings (named SaaS comparables) | Quarterly | NRR disclosures from public SaaS filings (Snowflake, Datadog, Cloudflare, HubSpot, etc.) used as ground-truth anchors when the published-benchmark and the SEC-filing figure diverge. /valuation-impact references public NRR disclosures inline where relevant. |
| Battery Ventures - Cloud Software Spending Survey | Annual | Buyer-side cloud-software spend and renewal-rate research. Used as a check on the expansion-revenue contribution band and on the buy-side renewal-rate context for /valuation-impact and /scenarios. |
| FRED - Federal Reserve Economic Data (macro context) | Daily / monthly | Used only as a macro context check when a benchmark drift correlates with a public macro indicator (real rates, IT spending indices). Not used as a churn-benchmark source directly. |
In scope
- ✓Three-layer churn cost math (direct MRR loss, wasted CAC, destroyed LTV) with formula shown on every calculator
- ✓Monthly and annual churn rate conversion using the standard 1 minus (1 minus monthly)^12 derivation
- ✓Segment-aware benchmarks (B2C, SMB, mid-market, enterprise) with ACV-band breakouts where the source data permits
- ✓Involuntary churn modelling: failed-payment MRR loss, smart-retry recovery, account-updater contribution, dunning ROI
- ✓NRR and GRR formulas with worked examples and 2026 benchmark anchoring
- ✓Valuation-multiple sensitivity to retention with the SaaS Capital and public-comparables empirical band
- ✓Industry-subvertical bands (SaaS, e-commerce subscription, telecom) drawn from source-specific publications
Out of scope
- ×Per-company contract-level pricing for ChurnZero, Gainsight, Totango, or Vitally - list-price tiers and analyst-quoted bands only
- ×Single-customer probability-weighted churn modelling (cohort-level analytics is the right tool, not this site)
- ×Enterprise contract-renewal negotiation tactics and price-and-discount mechanics
- ×Cohort-decay modelling beyond aggregate segment bands (retentioncost.com covers this surface)
- ×Per-vendor CS-platform feature comparison beyond the comparison grid on /tools
- ×Single-jurisdiction tax-driven churn (US sales-tax nexus or VAT-threshold transitions) not modelled
Calculation framework
True churn cost = Direct MRR loss + Wasted CAC + Destroyed LTV. Direct MRR loss = MRR x monthly churn rate x 12. Wasted CAC = (annual churned customers) x (fully-loaded CAC). Destroyed LTV = (annual churned customers) x ARPU x (expected remaining lifetime months). The home page and /true-cost both implement the full three-layer sum.
Annual churn = 1 minus (1 minus monthly churn)^12. The naive monthly-times-12 approximation overstates annual churn by a small amount at low rates and a large amount at high rates (5 percent monthly is 46 percent annual, not 60 percent). The /calculator page documents this trap explicitly.
CAC waste per churned customer = fully-loaded CAC at the time of acquisition. Fully-loaded CAC includes sales salaries, commissions, marketing spend, content and product-led acquisition cost, and onboarding cost. Enterprise outbound CAC bands run $15K to $80K; inbound SMB bands run $500 to $3,000. /true-cost and /saas reference these bands inline.
Destroyed LTV per churned customer = ARPU x expected remaining lifetime months. Remaining lifetime is the inverse of monthly churn rate adjusted for expansion contribution. The conservative bound used on /true-cost assumes flat ARPU; the optimistic bound used on /scenarios layers 20 to 30 percent expansion contribution per Battery Ventures and ChartMogul.
Recovered MRR per cycle = (involuntary churn share x failed-payment MRR) x recovery rate. The recovery rate combines smart-retry (typical 50 to 60 percent recovery on the first three retry windows per Recurly) and account-updater (an incremental 10 to 15 percent recovery on the cards-on-file population). The /involuntary-churn page implements the combined 60 to 70 percent recovery band.
NRR = (start-period MRR minus churn minus contraction plus expansion) divided by start-period MRR, expressed as a percentage. GRR is the same numerator without the expansion term (caps at 100 percent by construction). The /nrr-benchmarks page documents both formulas with worked examples.
Refresh cadence
The benchmark dataset is re-checked on the first business week of each month against the publishers listed above. The last verified label currently reads May 2026.
The verification date is held in one constant (LAST_VERIFIED_DATE) imported by every page. Footer text, schema dateModified, and visible "Updated" stamps all read from that single source so the date cannot drift across the site.
Out-of-cycle refreshes happen on any of the following triggers:
- ●Bessemer State of the Cloud update (typically Q4 each year)
- ●OpenView SaaS Benchmarks Report annual release
- ●Recurly / ProfitWell research publication touching involuntary churn or expansion economics
- ●Material vendor pricing change at ChurnZero, Gainsight, Totango, or Vitally
- ●SaaS Capital churn-and-valuation publication update
- ●Public 10-K NRR disclosure from a named SaaS comparable that materially diverges from the published-benchmark median
Limitations
- ●Benchmark ranges are mid-market central tendencies, not predictions. A specific company can sit anywhere within the range based on cohort acquisition mix, contract structure, and payment-network behaviour.
- ●Vendor pricing pages drift between monthly verifications. Where ChurnZero or Gainsight publish only contact-sales pricing, the page flags this and uses analyst-quoted bands.
- ●Segment definitions (enterprise vs mid-market vs SMB) vary slightly between Bessemer, OpenView, KeyBanc, and Recurly. Cross-source bands on this site use the median-of-medians where definitions diverge, with the divergence flagged on /benchmarks.
- ●The expansion-revenue contribution band used in the three-layer destroyed-LTV math is conservative. Operators with strong expansion motion (best-in-class 20 to 30 percent of revenue from expansion per Bessemer and Battery Ventures) should adjust the LTV multiplier upward.
Corrections process
Spotted a stale benchmark, a missing segment, or a source we have not caught? Email [email protected] with the page URL, the specific number in question, and the source you would like cited. Substantive corrections (benchmark band shifts, vendor pricing changes, methodology errors) are typically actioned within five business days. The LAST_VERIFIED_DATE is rolled forward when the change ships.
Updated May 2026