R2 — Treat every vendor retention number as survivor-only and quarantine it

Rule: When citing GoDaddy ~85%, Wix 105% NRR, or any builder vendor retention figure, always pair it with (a) the survivor-only caveat, (b) the independent SMB-SaaS churn range (SMB-focused SaaS churn benchmark — 3–7% monthly logo churn (≈31–58% annually); DIY builders fit this reference class), and (c) the .com/.net domain-renewal proxy (.com/.net domain renewal — preliminary 75.0% Q4 2025 (final 75.4% Q3 2025); low-to-mid-70s since 2004).

Why: Every vendor retention statistic excludes never-launched and abandoned accounts (Survivorship bias is the central, recurring threat in DIY-builder evidence — vendor stats and active-site surveys both exclude failures). Citing a vendor figure without the comparator misrepresents the category.

How to apply: No client-facing or article text should quote a vendor retention figure alone. Quarantine triple = (vendor figure) + (survivor-only flag) + (independent comparator). Cross-link GoDaddy FY2025 10-K — ~85% customer retention (84% in 2024 due to divestitures); 90% for >3-year cohort, Wix FY2025 net revenue retention 105% (Q4 earnings call) — NRR is not logo retention, NRR vs logo retention — distinct metrics; NRR > 100% does NOT mean low customer churn.