Methodology

How Law Leaderboard Ranks Lawyers

Objective, data-driven lawyer comparisons using NLP review analysis, market intelligence, and public records. Every metric is computed from publicly available data using reproducible methods.

500
Law firms tracked
25
US cities covered
250,863
Reviews retrieved
212,755
Text reviews analyzed

Unlike traditional directories that rely on self-reported ratings or pay-to-play placements, every metric on Law Leaderboard is computed from publicly available data. Our database spans six categories: profile data, review intelligence, market metrics, responsiveness, case type coverage, and computed scores.

1

Firm Identification

For each city and practice area combination, we identify the top 20 law firms from Google Maps search rankings. We use the same query a consumer would type — for example, "personal injury lawyer Houston" — and capture the top 20 results from Google's local search algorithm. This approach reflects real consumer discovery behavior.

2

Profile Data Collection

For each identified firm, we collect comprehensive profile data from their Google Business listing.

  • • Name, address, phone, website, coordinates
  • • Hours of operation, wheelchair accessibility, online appointments
  • • Primary and additional practice area categories
  • • Photo count, logo, claimed/verified status
3

Review Collection

The current public build stores the Google review records we can retrieve for each firm from the source/API. Depending on the profile, that can range from a handful of reviews to several thousand. Public Google review counts, retrieved review records, and text reviews analyzed are related but different numbers. Each stored review includes:

  • • Full review text when available
  • • Star rating (1-5) with exact timestamp
  • • Reviewer name and Local Guide status
  • • Owner response text and timestamp (if present)
4

NLP Review Analysis

This is the core of our methodology. We run Natural Language Processing (NLP) keyword extraction across all collected review text to derive quantitative metrics that no other directory publishes.

5

Computed Metrics

We derive several unique metrics from the raw data:

6

Market Intelligence

For each city and practice area, we collect market-level data including search volume, cost-per-click, trend data, and AI Overview presence.

7

City & State Aggregation

We aggregate individual firm metrics to produce market-level insights: average rating, median review count, average owner response rate, and the most commonly praised attribute across all firms in a market.

NLP Analysis: What We Extract

Praise Attribute Scoring

We scan each review for 50+ keyword patterns across 10 praise categories: communication, professionalism, results, compassion, recommendation, expertise, accessibility, dedication, transparency, and speed.

Complaint Analysis

For 1-3 star reviews, we extract 6 complaint categories: poor communication, slow process, low settlement, unprofessional behavior, being passed around, and hidden fees.

Case Type Detection

We identify case types mentioned in reviews: car accident, truck accident, motorcycle, slip and fall, medical malpractice, workplace injury, wrongful death, dog bite, product liability.

Computed Metrics

Owner Response Rate — What percentage of reviews the firm responds to. Computed from the presence/absence of owner responses.
Average Response Time — Mean days between a review being posted and the owner responding, calculated from timestamps.
Review Velocity — How many reviews the firm receives per month, indicating current client volume.
Polarization Index — Difference between 5-star % and 1-star %. A consistently good firm scores above 90.
Recommendation Rate — Percentage of reviewers who explicitly use the word "recommend" in their review text.

Data Freshness

All data is refreshed monthly. Review analysis is based on the most recent reviews available at collection time. Market data (search volume, CPC) is sourced from Google Ads API. Last full refresh: April 2026.

Known Limitations

We want to be transparent about what our methodology does not cover:

⚠ Public Google review counts, retrieved review records, and text reviews analyzed are different counts and should not be treated as interchangeable.
⚠ Some very large Google profiles still exceed source/API collection ceilings, so the retrieved dataset can remain below the public review count.
⚠ NLP keyword extraction is pattern-based, not AI-inference-based. It may miss nuanced sentiment.
⚠ Google Maps rankings fluctuate. A firm in our top 20 today may not appear tomorrow.
⚠ We do not verify attorney credentials, bar status, or disciplinary history (planned for future updates).
⚠ Ratings and reviews can be manipulated. We do not currently perform fake review detection.

Questions about our methodology? We welcome scrutiny.

Explore Rankings →