Outrigger AI Visibility Index: Off-Page Decomposition

“The audit covers 2,729 businesses across 14 verticals (personal injury law; real estate; dental; accounting; med spa; home services / plumbing; financial advisors; SaaS — CRM; SaaS — project management; digital marketing agencies; e-commerce — baby/kids/family DTC brands; boutique hospitality; insurance brokers and comparison platforms; personal-finance and money-management apps) and four metropolitan markets (Los Angeles, New York, Chicago, Sydney)”

“Sampling proceeded in two waves. Phase 1 (n = 1,004) sampled businesses URL-first, beginning with an industry-city directory query and verifying first-party domain ownership. Phase 2 (n = 1,725) sampled name-first, beginning with a business name and resolving the URL post-hoc via Apify Google SERP.”

“The single largest off-page partial r anywhere in the table is has_wikipedia → Claude (+0.219), and the second largest is quora_mention_count → Claude (+0.203).”

“Domain Authority has a raw Pearson correlation of 0.337 with the composite visibility score on the Phase 1 sub-sample (n = 1,004) and 0.217 on the full sample (n = 2,200 businesses with non-null DA). After BH-FDR correction it remains significant as a raw correlation. In the multivariate L2-regularised logistic regression on n = 2,729 — which simultaneously estimates 22 features — its standardised coefficient is +0.139 with a bootstrap 95% confidence interval of [−0.031, +0.316], which crosses zero.”

“We interpret this as confounding rather than absence of effect: high-DA businesses are systematically more likely to have Wikipedia entries, LinkedIn company pages, BBB and Yelp listings, and active YouTube channels, and the off-page indicators absorb most of the explanatory power that DA appeared to carry alone.”

“We split the sample into Domain Authority quartiles and compared mean visibility scores between businesses with above-median directory_count and below-median directory_count within each quartile.”

“One reading is that off-page presence has the most leverage for already-authoritative businesses; another is that directory coverage itself becomes more complete as DA rises, and the uplift in Q4 reflects the joint product of authority and coverage rather than a pure off-page effect. The dataset is consistent with either reading; the controlled trial in §7 is designed to discriminate.”

“All of these findings are correlational. The dataset cannot adjudicate causality. A business that is mentioned on Reddit and on Wikipedia and that ranks in BBB and Yelp directories is likely to be a more established business with more brand equity, and brand equity itself plausibly drives both off-page presence and AI mention probability.”

“Observational, not causal. The dataset cannot distinguish causation from confounding. Partial-correlation and OLS-residualisation analyses isolate linear contributions of a predictor controlling for a measured set of confounders; unobserved confounders (brand age, marketing spend, executive PR effort, demand-side prominence) are not controlled for.”

“Stochastic AI outputs. Each (business, model, prompt) cell is a single sample from a stochastic process. We did not measure within-cell run-to-run variance.”

“Single point in time. All probing was conducted in April–May 2026. AI model versions, training cutoffs, and grounding behaviour change frequently. Replication on the same sample at a later date would likely produce a different mention pattern at the margin.”

“The author is the founder and operator of MentionLayer, a generative engine optimization platform whose product thesis overlaps with the research questions of this paper.”

“These results are correlational. To test causal magnitude, a Phase 3 controlled trial is planned for May–July 2026, with the protocol pre-registered on the Open Science Framework before enrolment.”