Skip to main content
Last updated April 19, 2026

How we rank every local business in America.

Operator scores 35,305,944 businesses across 774 cities every night. Here is exactly how. No black boxes. No paid placement.

Refresh cadence: Nightly 2am HSTSources: 6 primary + 3 enrichmentPrimary signal: Health score (0-100)
The rules

Six principles we refuse to break.

Ranking local businesses at scale is an exercise in restraint. These are the guardrails we commit to publicly.

01

No paid placement

Rank is determined by the model. No promoted listings, no sponsored slots, no ad inventory in any leaderboard.

02

Same rules for everyone

Claimed businesses on paid plans rank with the same formula as unclaimed businesses.

03

Full methodology is public

Weights, inputs, and source boundaries are documented publicly and versioned.

04

Nightly refresh, never stale

Every business gets rescored nightly, so rank can move within 24 hours of meaningful signal change.

05

Businesses can see their own score

Claimed profiles can inspect component-level score detail without a paid requirement.

06

Appeal is a real process

Scoring disputes are reviewable by a human and corrections are tracked through transparency reporting.

The core score

Health score.

A 0-100 composite telling you whether a local business is winning or losing in its market, refreshed nightly.

The six components

Reviews25%
Market Rank20%
Financial Health20%
Response Rate15%
Jobs Flow10%
Operator activity10%

Reviews (25%)

Composite of rating quality, review volume against category baseline, 90-day velocity trend, and recency decay.

reviews_score = 0.40·rating_norm + 0.25·volume_vs_median + 0.20·velocity_90d + 0.15·recency_decay

Market Rank (20%)

Normalized percentile rank inside each city + category market.

Financial Health (20%)

Estimated from category/geography revenue bands, tenure, digital footprint, and operating signals. This is an estimate, not disclosed revenue.

Response Rate (15%)

Share of trailing reviews that received an owner response, weighted by latency and response quality characteristics.

Jobs Flow (10%)

Completed work velocity and reliability where first-party data exists; otherwise a category baseline is applied.

Operator activity (10%)

Captures adoption depth for active Operator workflows. Non-users receive a neutral category baseline.

Why composite scores beat single metrics. A single score can be gamed. The composite balances volume, velocity, recency, and response quality.
Leaderboard order

How we compute market rank.

Each market is a city + primary category tuple. Rank sorts by market-share signal, with Health score and tenure as tie-breakers.

market_share_scorePrimary sort key

A normalized estimate of local demand share based on review depth, rating quality, velocity, and response performance.

share = (review_volume·rating·velocity_factor·response_factor) / sum(market)
  • market_rank - integer position from 1 to market_total.
  • market_total - total competitors in this city + category.
  • market_share_score - normalized 0.00 to 1.00 estimate.
New signal

AI Visibility score.

A 0-100 signal for how findable a business is across AI answer surfaces.

Inputs to AI Visibility

Schema.org markup20%
llms-full.txt quality15%
robots.txt AI policy15%
Citation frequency20%
Content structure15%
Site speed + crawlability15%

Citation frequency is measured from representative local intent query sets and tracked on rolling windows.

Where data comes from

Data sources.

Operator combines licensed/open substrate data, structured enrichment, and owner-verified corrections.

Overture Maps FoundationPrimary

Base layer for business existence, location, and category. Licensed under CDLA Permissive 2.0.

Google Places metadataPrimary

Ratings, review counts, and profile metadata via official APIs under provider terms.

USPS + state registriesPrimary

Address normalization, tenure signals, and licensing context where publicly available.

First-party web crawlEnrichment

Structured extraction of business-owned domains, respecting robots.txt posture.

Structured extractionEnrichment

Model-assisted extraction to normalize service, pricing, and content quality signals.

Owner-verified claim editsOwner

Claimed businesses can correct key profile fields; owner edits supersede inferred values.

Boundary condition: no paid consumer location SDK dumps, no leaked personal datasets, and no prohibited scrape sources.
Update cadence

Freshness.

Everything in Operator updates on one of four cadences:

Nightly2am HST

Health score, market rank, and AI Visibility scores recompute nightly.

WeeklySunday

Site enrichment and model-assisted extraction refresh on a weekly cycle.

Monthly1st of month

Base map and directory substrate synchronize with upstream source updates.

Real-timeOn claim edit

Owner profile changes apply immediately; composite score updates on the next nightly pass.

Epistemic honesty

What we do not score.

  • Actual revenue. Financial signals are modeled estimates.
  • True craft quality. Reviews are a proxy, not absolute truth.
  • Ethical fit. Compliance and trust indicators are surfaced separately.
  • Personal fit. Rank is a filter, not a final buying decision.
Answers

Methodology FAQ.

How often does Operator update business rankings?

Every night at 2am HST. Health scores, market ranks, and AI Visibility scores refresh on the same nightly cycle.

Where does Operator get its business data?

Primary source is Overture Maps Foundation, enriched with official metadata APIs and first-party crawl/extraction. Claimed businesses can override key fields with verified edits.

What is the Health score?

A 0-100 composite from six weighted components: Reviews (25%), Market Rank (20%), Financial Health (20%), Response Rate (15%), Jobs Flow (10%), and Operator activity (10%).

How is market rank calculated?

Rank is computed within each city + category market using market-share signals first, then Health score and tenure as tie-breakers.

Can a business pay to rank higher?

No. There is no sponsored placement and no paid rank boost.

What is the AI Visibility score?

A 0-100 measure of how findable a business is in AI answer surfaces, based on structured data, citations, crawlability, and content quality signals.