{"id":2049,"slug":"google-soasta-2017-bounce-probability","title":"Google/SOASTA 2017 — P(bounce) rises 32% at 1s→3s, 123% at 1s→10s","kind":"reference","scope":"business","status":"current","audiences":["kevin","candid-team"],"topics":["website-credibility-and-speed"],"reference_body":"**Claim:** Google/SOASTA 2017 (deep neural net, 90% prediction accuracy): probability of bounce rises **32%** as load goes 1s→3s, **90%** at 1s→5s, **123%** at 1s→10s.\n\nGoogle-published, A/B-controlled partner cases (web.dev): **Rakuten 24** — +53.37% revenue/visitor and +33.13% conversion from LCP optimization in an A/B test; **Vodafone Italy** — +8% sales from a 31% LCP improvement.\n\n**Source:** Google/SOASTA 2017; web.dev partner cases. Credible but self-selected (winners).\n\n**Confidence:** Verified.\n\n**Why this matters for Candid:** One of only three causal datapoints in the whole tier-mapping research. Page-speed → conversion is real. But note: speed → RANKING is overstated; see [[google-page-experience-not-ranking-lever]].","rationale_body":null,"metadata":null,"links":{"outgoing":[],"incoming":[{"slug":"research-brief-smb-widget-difficulty-to-work-june-2026","title":"Research brief: SMB widget difficulty-to-work mapping — three tiers of work for three sizes of gap (June 2026)","kind":"reference","scope":"business","link_type":"relates-to"},{"slug":"tier-1-foundational-web-presence","title":"Tier 1 — Foundational web presence (SMALL GAP)","kind":"reference","scope":"business","link_type":"depends-on"}]},"created_at":"2026-06-23T19:16:02.673Z","updated_at":"2026-06-23T19:16:02.673Z"}