How We
Do Research
The Florida Local Search Index is not one dataset collected one way. It is a multi-method measurement program that we test, version, and refine over time. Here is how we measure, what we stand behind, and where we are still improving.
Three Tiers of Measurement
We do not use one tool for everything. We use the best available measurement for each signal, and we upgrade modeled inputs to directly measured ones as our capability grows.
Manual Deep Audits
Ground-truth audits of individual businesses (97 in Brevard County, 47 data points each): GBP completion, schema, page speed, reviews, and response patterns, reviewed by hand. This is our accuracy anchor.
Statewide Market Intelligence
Across 90 Florida cities and 12 industries, we measure real market structure from the Google Places API and by fetching the top-ranking businesses' sites: listing density, review depth, ratings, and structured-data adoption. Measured, not estimated.
AI Visibility Instrument
We record what AI assistants (ChatGPT, Gemini, Grok, Claude) actually recommend for local searches, run repeatedly and captured verbatim. Whether a business appears is a reproducible fact about that assistant at that time. Original data nobody else publishes locally.
We are transparent about how we measure and what each number means. The captured response corpus, the exact prompt sets we run against AI assistants, and our scoring configuration stay proprietary. We publish the method and the findings, not the turnkey recipe.
Inside a Deep Audit
How the Tier 1 ground-truth audit works, the process that anchors the accuracy of everything else.
Define the Sample
For a deep audit we identify the businesses in the target industry and market using Google Maps and local sources, and include every verifiable business, not a curated subset. Our first full deep audit covered Brevard County; the same process extends to new markets.
Manual Audit
Each business listing is individually reviewed. We document GBP completion, website presence, schema markup, mobile responsiveness, page speed, review count, rating, and response patterns.
Automated Testing
We run Google PageSpeed Insights and schema validation tools against every website. Results are captured with timestamps for reproducibility.
Data Normalization
Raw data is cleaned, deduplicated, and normalized. Businesses with multiple locations are counted individually. Permanently closed businesses are excluded.
Analysis & Peer Review
We analyze the data for patterns, compute benchmarks by industry and city, and have findings reviewed by at least two team members before publication.
Publication
Results are published with full methodology documentation, data sources, and known limitations. We update studies quarterly when new data is available.
Analysis Framework
GBP Completeness Score
We score each profile on 15 factors including business description, categories, hours, photos, posts, Q&A, and attributes. Each factor is weighted by its impact on local rankings.
Technical SEO Audit
Schema markup presence, mobile responsiveness, HTTPS adoption, page speed (LCP, FID, CLS), and indexation status are tested for every website.
Review Health Score
We measure total review count, average rating, review recency, response rate, and response time. These factors are combined into a single review health score.
Industry Benchmarking
All metrics are segmented by industry and city. This allows businesses to compare themselves against direct competitors, not the market average.
Known Limitations
All data is captured at a point in time and reflects what we measured on that date. AI answers, listings, reviews, and rankings change, so we re-measure and version each release.
Tier 2 statewide data samples the leading listings in each market, not every business. It reflects competitive structure, not a full census.
AI-visibility runs use the assistants' APIs with live search, the best measurable proxies for the consumer apps but not identical to them. Google AI Overviews have no API and are not yet included.
We cannot measure some ranking factors (domain authority, backlink profiles) without third-party tools that carry their own accuracy limitations.
Review counts and ratings are public data. We cannot verify review authenticity.
The program is evolving. We continuously test methods, benchmark results, and upgrade modeled inputs to directly measured ones as capability allows.
We publish these limitations because we believe transparency is more valuable than appearing infallible. If you spot an error in our research, please .
Market Opportunity Score
The Market Opportunity Score (0 to 100) is the single most important number in the Florida Local Search Index. It tells a business owner, in one number, how much untapped local search potential exists in their city and industry.
Opportunity Score = a versioned, weighted blend of measured market inputs: competition depth, rating floor, structured-data gap, and AI-visibility gap.We publish the inputs and what each one means. The exact weights are versioned and refined as we add directly measured signals; the current configuration is proprietary.
Competition Depth
- Review-count distribution among the top listings
- How many established competitors hold the market (100+ reviews)
- How crowded the top of the map pack already is
Rating Floor
- The average rating you need to be competitive
- The lower-quartile rating among the leaders
- How much headroom exists below the top tier
Structured-Data Gap
- Share of top-ranking sites with schema markup, measured by fetching them
- How many leaders are still missing it
- The opening that gap leaves for AI and rich results
AI-Visibility Gap
- How often local businesses surface in AI assistant answers (Tier 3 instrument)
- Whether the market has a clear AI-recommended leader yet
- How contested those answers still are
Tier 2 statewide market data is measured from the Google Places API and by fetching the top-ranking businesses' sites: real listing counts, review-count distributions, ratings, and structured-data adoption. These are measurements, not estimates, though they sample the leading listings in each market rather than every business and reflect a point in time. Tier 1 adds manual, 47-point Brevard deep audits. Tier 3 records AI-assistant recommendations directly. The opportunity score is a versioned formula over those measured inputs, refined as we add signals.