Survivorship bias is the central, recurring threat in DIY-builder evidence — vendor stats and active-site surveys both exclude failures

Claim: Every vendor retention statistic and every active-site survey in the DIY-builder category systematically excludes never-launched and abandoned sites by construction. Wix's F-1 (Wix Form F-1 (SEC, March 20 2014) — verbatim admission that many registered users never publish a website) explicitly admits the platform does not measure publishing or engagement.

Source: Methodological synthesis across the brief.

Confidence: This is a structural observation about the available evidence, not a metric.

Why this matters for Candid: Any retention number ≥80% from a builder vendor measures retention among survivors, not retention of started accounts. The denominator gap is the most important thing to remember when reading any vendor or aggregator statistic in this category. Cross-link R2 — Treat every vendor retention number as survivor-only and quarantine it.