Reviews · 2026-08-21

Google Review Statistics for 2026: The Numbers That Actually Have Sources

Why an audited list matters

Review statistics circulate fast and degrade faster. The "90% of consumers read reviews before visiting a business" figure that appears on hundreds of marketing blogs is widely repeated without a traceable source. The current BrightLocal figure below is 97%, with its survey vintage and sample size stated.

Every number below is traced to a named study, with its methodology, sample size, and data year stated. Where a figure's evidence grade is weaker than (b), it is flagged.

Consumer behaviour thresholds (BrightLocal 2026)

All figures from the BrightLocal Local Consumer Review Survey 2026 — a panel survey of 1,080 consumers evaluating local business selection criteria (grade b, published January 2026).

• 97% of consumers read online reviews for local businesses.
• 41% always read reviews before choosing — up from 29% the prior year.
• 68% reject businesses rated below 4.0 stars.
• 31% only use businesses rated 4.5 stars or higher — up from 17% in 2025.
• 47% refuse to use a business with fewer than 20 total reviews.
• 74% care only about reviews written in the last three months.
• 32% only trust reviews under two weeks old.

The year-on-year comparisons point towards tighter thresholds. The minimum credible profile is getting harder to maintain passively.

Ranking weight (Whitespark 2026)

From the Whitespark 2026 Local Search Ranking Factors survey — a Delphi panel scoring 187 individual factors across 47 expert SEO practitioners (grade b, operational for 2026, published November 2025).

• Google Business Profile signals: 32% of Local Pack weight.
• Review signals (velocity, recency, rating, keyword density, owner response rate): 16–20% of Local Pack weight, increasing year-on-year.
• On-page / website signals: 15–19%.
• Behavioural engagement: 8–9%.
• Link and authority: 8–15%.
• Citation / NAP consistency: 6–7%.

Together, GBP and reviews account for roughly half the ranking engine. Proximity remains the largest single factor (approximately 55% of algorithmic variation in location queries) but is uncontrollable.

Revenue impact

• 108% revenue increase for businesses with 25+ reviews versus thin profiles — Womply 2026 local merchant database study, compiled by Sixth City Marketing (grade b, correlation of transactional data with active GBP and Yelp profiles).
• 5–9% revenue increase per one-star rating improvement — Harvard Business School (Luca, M.), regression discontinuity analysis of Yelp ratings and independent restaurant revenues (grade a, peer-reviewed academic study, re-evaluated 2026).

The first figure is a correlation, not a controlled experiment — businesses with more reviews likely also invest more in marketing. The Harvard figure uses a rigorous regression discontinuity design that isolates the rating effect, which is why it earns grade (a).

Review freshness decay

The BrightLocal 2026 data on freshness is worth isolating because it is the most commonly ignored threshold. Seventy-four per cent of consumers care only about reviews from the last three months. Thirty-two per cent set the bar at two weeks.

A profile with 50 reviews, all from 2024, performs worse on this dimension than a profile with 15 reviews, all from the last quarter. Volume alone does not resolve the freshness problem.

Channel performance for review solicitation

From the Wylto 2026 performance benchmark — analysis of 2.3 million anonymised messaging campaigns (grade b, published April 2026).

The WhatsApp figures include native quick-reply CTA buttons; the email figures reflect HTML links. The gap in click-through rate — roughly 10x — is large enough to be structurally significant for review collection.

AI discovery thresholds

These figures are newer and less well-established than the Google-specific data.

• ChatGPT recommends only 1.2% of analysed business locations, versus a 35.9% visibility rate in Google's Local Pack for the same set — SOCi 2026 Local Visibility Index (grade b). AI discovery is roughly 30 times more selective than traditional local search.
• 150+ detailed reviews is the emerging threshold for consistent AI recommendation — BizIQ 2026, citing multi-profile crawl data (grade b).
• 74% of local businesses with NAP inconsistencies across three or more directories are excluded from Google AI local answers — Semrush 2026 (grade b).

These thresholds will move. Treat them as directional, not fixed.

What the numbers do not tell you

Attribution in local search is messy. A parent may find your profile on Google Maps, visit your website, and then phone — blending organic SEO and conversion metrics in ways that no single study cleanly separates. The Womply revenue figure is correlational. The Whitespark weights are expert-scored, not experimentally derived. And every consumer survey carries self-report bias.

The numbers are useful as a framework, not as engineering tolerances. The thresholds — 4.0 stars, 20 reviews, three-month freshness — are the clearest actionable signals in the data.