Google Reviews Benchmark Study 2026

We analyzed 116,220 Google Maps SERP observations across 108 US metropolitan areas and 14 industries. This study measures what actually predicts who appears in the local map pack — and what doesn't.

The finding that matters: the strongest signal we tested is inverted, and every other signal the local SEO industry sells correlates with rank at or below ρ = ±0.1. Proximity to the searcher explains 96% of top-3 variance at a 5km offset. Everything else is fighting over the remaining 4%.

The Signal Strength Ranking

Across all observations, we tested six signals using Spearman rank correlation against Google Maps position. Spearman ρ measures monotonic association between a signal and rank position — non-parametric, no normality assumption. Values near zero mean the signal does not predict ranking in a monotonic way.

The strongest signal is inverted. Every other signal the local SEO industry sells — reviews, velocity, recency, GBP completeness, review content quality — registers at or below ρ = ±0.1. For reference, a correlation of 0.1 explains roughly 1% of ranking variance.

Methodology note: All correlations are observational, not causal. We avoid “ranking factor” language throughout. Sample enrichment: 3,375 backlinks profiles, 2,397 GBP profiles, 141,900 reviews across 2,286 top-3 businesses.

Proximity Is the Algorithm

We searched from a point 5 km away from each metro center. Across 10 major metros and 9 industries — 90 queries total — 96% of the top-3 results changed.

This is not a ranking signal in the traditional sense. It IS the algorithm. Everything else is fighting over the remaining 4%.

Three verticals — Dental, Auto Repair, Real Estate — showed literally zero top-3 overlap at a 5 km offset. Veterinary was the only vertical where any meaningful portion of the top-3 persisted across the offset, and even that was only 16.7%. For a searcher who walks 5 km in any direction, the dentist Google recommends is almost certainly a different dentist.

Proximity test: 90 queries (9 verticals × 10 MSAs). Offset: +0.045° latitude (~5 km north of canonical metro center). Overlap is the fraction of businesses that appear in BOTH the original and offset top-3 for the same keyword.

Organic and Local Are Separate Universes

Across 954 city-industry pairs, the Google Maps top-3 and the organic search top-3 share exactly 0 businesses.

Traditional web SEO — blog posts, backlinks, content marketing — operates in a completely different system from local pack ranking. The businesses dominating organic search for local queries are Yelp, Angi, BBB, and Zillow. The businesses winning the local pack are independents with reviews.

If you pay an agency to rank you on Google, you need to know which Google they're ranking you on. Ranking at position 8 in organic for “plumber near me” does almost nothing for your phone. Ranking at position 2 in the map pack does everything.

The Domain Authority Inversion

Higher domain authority correlates with worse local pack ranking across every industry tested.

Every vertical's correlation is positive. In Spearman terms, positive ρ between domain authority and rank number means higher authority associates with larger (worse) rank positions. The businesses with the strongest web presence rank at positions 8, 10, 15. The businesses winning the top 3 often have small websites or no website at all.

Review Benchmarks by Industry

The number of reviews needed to compete in the top 3 varies 83x across industries — from Restaurant at 830 median reviews down to Accounting at 10.

The “min to compete” column is the 25th percentile of top-3 review counts — the number a business in your industry needs to not look like an outlier in the local pack. Below this, you are structurally disadvantaged before proximity is even considered.

Density tiers are assigned by MSA population. DENSE = top third of the 108 markets by population, THIN = bottom third. In most verticals the competitive bar is not dramatically different between tiers — plumbers in Akron need roughly what plumbers in Boston need. Legal is the big exception: personal injury attorneys in dense markets need 1.8x more reviews than in thin ones.

Review Quality Doesn't Predict Rank

We scored review semantic quality — whether reviews mention specific services, locations, and outcomes — for 2,286 top-3 businesses using 141,900 pulled reviews.

The correlation between review quality score and rank position: ρ = −0.036. Indistinguishable from noise.

This was tested twice. First via a crosswalk between scouting data and SERP data (ρ = −0.005, n=2,070). Then via direct review pulls for all top-3 businesses (ρ = −0.036, n=2,286). Both confirm: what reviews SAY does not predict where the business ranks.

Methodology: Semantic quality was measured by heuristic keyword matching — service-specific terms, location mentions, and outcome words (fixed, solved, helped, etc.). This is a proxy for review depth, not sentiment. Reviews with zero matching terms score 0; reviews with all three dimensions score 3. The per-business score is the mean across all pulled reviews. No AI/LLM classification was used in the primary analysis.

Review Velocity: Weak but Real

Review velocity — the rate at which new reviews arrive — shows a weak but consistent signal. Overall: ρ = −0.063. The strongest per-vertical effects appear in Real Estate (−0.166) and Insurance (−0.146). Restaurant showed the weakest effect (−0.021).

Velocity matters more in low-review-count verticals where each new review represents a larger percentage change. In high-review verticals like Restaurant and Veterinary, velocity is negligible — any individual new review is a drop in a 300-plus bucket. Even in the strongest verticals, a correlation of −0.17 explains less than 3% of ranking variance.

GBP Completeness: Near Zero

We profiled 2,397 Google Business Profiles. Completeness — whether a business has a description, hours, phone, website, photos, and categories filled — correlates at ρ = −0.038 with rank.

Near zero. Filling out your GBP is good practice for conversions — a complete profile converts a visitor into a call more reliably than a half-built one. The data does not support treating completeness as a ranking lever.

This finding contradicts the most common piece of local SEO advice: “fill out every field on your Google Business Profile to rank higher.” Our measurement says fields beyond the minimum threshold (name, category, address, phone) do not move ranking position in any direction that beats noise. Fill out your profile because customers read it, not because Google counts fields.

Stability and Methodology

Before the signal analysis, we ran stability checks to confirm the local pack is stable enough to measure. Results:

The local pack is stable across time, device, and query variation — but not across location. 4.1% overlap at 5 km offset means proximity overrides every other factor we measured combined.

Methodology

• Source: Google Maps SERP via DataForSEO API.
• Period: April 2026.
• Scope: 116,220 observations across 108 US MSAs and 14 industries, 3–4 keywords per vertical.
• Enrichment: backlinks for 3,375 domains, review histories for 2,397 businesses (141,900 reviews), GBP profiles for 2,397 businesses, organic SERP for 972 queries, search volume for 14 keywords.
• Statistical method: Spearman rank correlation (ρ). Non-parametric, no normality assumption.
• Google AI Overviews for “near me” queries: 0% (n=50). Excluded from signal findings.

What we did NOT measure

• Click-through rate
• Conversion rate
• Phone call volume
• Whether the business actually got customers from ranking
• Causal relationships — all findings are correlational

Caveats

• Per-city benchmarks for core verticals are based on 3 top-3 observations per city (the actual SERP, not a sample from a larger pool).
• Temporal stability of 83–93% means the top-3 persists across snapshots — results are stable but not frozen. Expect ~1 business in the top-3 to rotate on any given day.
• The proximity finding means our center-point choice affects which businesses we observe, but the relative dynamics between businesses (ρ values, benchmark distributions) hold.
• Signal correlations are reported for 9 core verticals with full data enrichment (plumbing, electrical, roofing, auto repair, dental, legal, insurance, real estate, veterinary). The other 5 (restaurant, moving, locksmith, chiropractic, accounting) are reported only on review-count and rating distributions.

Data license: CC BY 4.0. Cite as: DuBois, IMPIOUS LLC, 2026. Aggregated benchmark data is downloadable as JSON at impious.io/data/benchmark_data.json.

What This Means

For business owners: proximity decides who Google shows. You cannot change your location. What you CAN control is having a claimed, complete Google Business Profile and a steady accumulation of authentic reviews. That combination — existing and being findable — is the whole game. Everything else the industry sells you is fighting over single-digit percentages of variance.

For agencies: the data suggests that traditional local SEO packages heavy on backlinks, content optimization, and citation building are not justified by measurable ranking correlation. The defensible service is the boring one — GBP optimization, review generation coaching, and honest diagnostics. If you sell anything else as a ranking lever, this data is going to keep getting harder to explain away.

For the industry: we publish this data under CC BY 4.0 because local SEO needs more original research and less recycled conventional wisdom. If you replicate this study and get different results, we want to hear about it.

Download the Dataset

The full benchmark dataset — 1,512 rows covering 14 industries across 108 US metros with per-market medians, national statistics, and Spearman signal correlations for the 9 core verticals — is available as a free download.

Download: benchmark-2026.csv (CC BY 4.0)

If you use this data in your own research or content, we ask that you link back to impious.io/research/google-reviews-benchmark-2026.

About This Study

This study was conducted by IMPIOUS LLC (impious.io), a local visibility research agency. Data was collected via DataForSEO APIs in April 2026. For methodology questions or custom analysis for your city or vertical: impy@impious.io.

A markdown-only version of this study is available at /research/google-reviews-benchmark-2026.md. Full site index for LLMs: /llms.txt.