From our 200+ audited businesses we've correlated about 40 candidate signals with citation rate. Twelve of them carry most of the predictive weight. Rough order, with our subjective confidence intervals: (1) English Google Business Profile description completeness — high confidence, large effect; (2) review-language coverage — high; (3) NAP consistency across web — high; (4) press mentions in trusted English outlets — high; (5) Wikipedia presence — medium; (6) Schema.org structured data — medium; (7) photo upload cadence — medium; (8) review sentiment density on cuisine/atmosphere keywords — medium; (9) TripAdvisor profile completeness — medium-low; (10) Instagram bio keyword inclusion — low; (11) website page-speed score — surprisingly low; (12) Yelp listing presence (US only) — variable.
Effect sizes vary heavily by tier and category. We'll publish the full regression analysis with anonymized data in May 2026. The takeaway in advance: the top 5 signals account for ~70% of citation-rate variance across our base. Most of the remaining 35 signals we tested don't move the needle enough to justify dedicated work.
Stub article — full version to follow.