Free local business review data for journalists
ReviewTactic Data Desk gives journalists free review data on local businesses in the US, Canada, Europe, Australia, New Zealand and Asia: ratings, review volume, response rates, complaint themes and suspicious-review rates by city, category or brand, in 1 to 7 business days. Source mention requested, not required.
Last updated 14 October 2026 Slots this week: 3 of 3 Briefs accepted: 0 of 100 3 a week, 100 in total, through 30 June 2027
Negative Google reviews checked against Google's content policy
Scan dataset, measured 6 September 2026
24.3%
of 369,688 negative reviews matched at least one Google policy.
- Business profiles
- 15,954
- Largest category
- Spam, 41.2% of 89,694 flagged
Limits. Flagged is not removed; Google alone decides. Sample is self-selected: businesses that ran a scan.
Source: ReviewTactic analysis of 369,688 Google reviews, 15,954 businesses, September 2026.
-
1,000,000+
local business profiles in the benchmark database
-
420,000+
restaurants among them
-
100+
review platforms monitored, Google the deepest
-
3
briefs a week, 100 in total, through 30 June 2027
What is the ReviewTactic Data Desk?
The ReviewTactic Data Desk is a free research service for working journalists, run by ReviewTactic, a review intelligence platform that monitors reviews across Google and 100+ review sites and keeps a benchmark database of more than 1,000,000 local business profiles, including more than 420,000 restaurants, across the United States, Canada, Europe, Australia, New Zealand and Asia.
You tell us the story and the question. We pull the numbers, write a short method note, and send you a summary, the data file and charts. There is no fee, no account to open and no sales call. If the data ends up in your story, we ask you to name ReviewTactic as the source.
We have published two cuts of this kind already: a benchmark of 6,721 Brooklyn restaurants and a study of 369,688 negative Google reviews across 15,954 businesses. The Desk is the same work, done for your city, category or beat. This page is how to ask.
What data can you ask for?
Pick the one that fits the story, or tell us the question and we will suggest the cut. Turnaround is in business days from the day we accept the brief.
- 1 to 2 business days
Quick stat check
One figure or one table answering one question. "What is the median Google rating of a Phoenix dentist?" "What share of Portland cafés with 100+ reviews have never replied to a review?" You get the number, a one-paragraph method note and the citation line.
- 3 to 5 business days
City, county, state or country benchmark
For one geography and up to three business categories: the rating distribution, review-count distribution, medians and quartiles, owner response rate and the unanswered-review backlog (among businesses with 100+ reviews), the share of businesses under 4.0 stars, and the year-over-year trend where our snapshots allow it. Delivered as a two-page summary, a CSV and two to four charts.
- 5 to 7 business days
Category deep-dive with complaint themes
Everything in the benchmark, plus what people write: the top complaint themes and top praise themes, how sentiment moved over time where our stored samples allow it, how complaints differ between one-star and three-star reviews, and short anonymized excerpts that illustrate each theme.
- 5 to 7 business days
Brand or chain snapshot
A named multi-location brand, or several side by side: per-location and aggregate rating, review volume, response rate, complaint themes from the reviews we hold and the share of negative reviews that match a Google content policy. Before we accept, we check whether the brand is a ReviewTactic customer. If it is, we say so and decline that part (see Disclosure below).
- 5 to 7 business days
Suspicious-review scan
For a geography or a category: the share of negative reviews that match one of Google's published content-policy categories (spam, off-topic, conflict of interest, harassment, personal information and others), broken down by category and star band. The classifier and categories are the ones in our published violation study; the sample is the stored reviews we hold for the businesses you ask about, and the method note says so. A flagged review is a review that matches the wording of a Google policy; it is not a removed review, and we never report removal rates.
- within 2 business days
An analyst on the record
A 20-minute call or an emailed Q&A with a ReviewTactic co-founder on what the numbers show, what they cannot show, and the published research around them. Quotable, with our disclosure attached.
Need a chart you can embed? Add one business day to any of the above and we will send a web-sized PNG with the source line in the footer, plus the CSV behind it.
What is in the dataset?
1,000,000+
local business profiles in the benchmark database
420,000+
restaurants among them
Coverage
The United States; Canada; Europe (EU and EEA countries, the United Kingdom and Switzerland); Australia and New Zealand; Asia (Japan, Singapore, Thailand, Vietnam, Malaysia and Indonesia).
-
United States and Canada
Canada follows the same data rules as Europe
-
Europe
EU and EEA countries, the United Kingdom and Switzerland
-
Australia and New Zealand
All Australian states and territories, and New Zealand
-
Asia
Japan, Singapore, Thailand, Vietnam, Malaysia and Indonesia, under the same data rules as Europe
See all 42 countries
- United States
- Canada
- Austria
- Belgium
- Bulgaria
- Croatia
- Cyprus
- Czechia
- Denmark
- Estonia
- Finland
- France
- Germany
- Greece
- Hungary
- Ireland
- Italy
- Latvia
- Lithuania
- Luxembourg
- Malta
- Netherlands
- Poland
- Portugal
- Romania
- Slovakia
- Slovenia
- Spain
- Sweden
- Iceland
- Liechtenstein
- Norway
- United Kingdom
- Switzerland
- Australia
- New Zealand
- Japan
- Singapore
- Thailand
- Vietnam
- Malaysia
- Indonesia
- Per business
- Star rating, review count, category, location, owner response rate (businesses with 100+ reviews), review recency. Sentiment, complaint themes and policy-category flags come from review text we already hold, not from a fresh pull for your request.
- Platforms
- Google plus 100+ review platforms monitored, including TripAdvisor, Yelp, Facebook, Booking.com and OpenTable. Google Maps is the deepest source; most custom cuts are Google-only unless you ask otherwise.
- Published samples
- 6,721 Brooklyn restaurants benchmarked (June 2026 data); 369,688 negative reviews across 15,954 business profiles classified against Google's policies (measured 6 September 2026).
- Methods you can quote
- The Reputation Index (a composite /100 of health, competitive position, reply quality and engagement), sentiment over time, topic heatmaps, competitor benchmarking, geo-grid local rank, and policy-violation detection.
- What is not in it
- Anything about individual reviewers. See "What we will not provide".
- Try it yourself
- The free review audit runs a live benchmark for any single business, so you can see the kind of fields a cut is built from.
How does a request work?
-
File a brief
Use the form below, or email partner@reviewtactic.com. Tell us the outlet, the story, the question, the geography, the categories or brands, and your deadline. Two minutes.
-
We reply within one business day
With yes, yes-with-changes, or no and why. If we say yes, you get a delivery date and the slot week.
-
We pull and write
A summary with the headline numbers and the limits, the CSV, the charts, a method note and the citation line. Every figure traces to a saved query we can re-run, and a co-founder signs it off.
-
You decide
Use it, don't use it, ask for a second cut, or ask for a quote. If it runs, please name us as the source and send us the link so we can log it.
Example brief
A city food reporter asks for the median Google rating and the owner response rate of restaurants in their metro, split by neighborhood, with a deadline two weeks out (the response rate covers restaurants with 100+ reviews, and the method note says so). That is a benchmark, 3 to 5 business days, delivered as a two-page summary, a CSV and charts, with the method note and the citation line.
What do we ask in return?
We ask you to name ReviewTactic as the source when you use the data. That is a request, not a condition: we do not ask for coverage, approval of your wording, or a link as a term of getting the data.
Copy-paste wording that works in most house styles:
-
In text
According to ReviewTactic, a review intelligence platform that analyzed [N] Google reviews of [M] [category] businesses in [place] in [month year], ...
-
Shorter
..., according to data from ReviewTactic, a review intelligence platform.
-
Chart or table footer
Source: ReviewTactic analysis of [N] Google reviews, [M] businesses, [place], [month year].
-
Formal citation
ReviewTactic ([year]). [Title of the analysis]. Analysis of [N] reviews across [M] business profiles. reviewtactic.com/for-journalists/
If your outlet links to sources, a link to reviewtactic.com or to the report the data came from is welcome; a nofollow link is fine, and so is no link. Please keep the limits next to the number where space allows, for example "flagged, not removed", "Google reviews only" or "businesses that ran a scan", so readers see what the figure does and does not mean.
How many requests can we take?
The Desk takes 3 a week, 100 in total, through 30 June 2027. One brief equals one story, however many cuts it needs. Quick stat checks are light and count as half a slot. Intake closes on 30 June 2027 or at the 100th brief, whichever comes first.
- Slots this week
- 3 of 3
- Briefs accepted
- 0 of 100
- Intake closes
- 30 June 2027
If the week is full we tell you the next open week rather than saying no; a confirmed deadline moves you up. Slots do not roll over. The counter at the top of this page is updated every Monday.
How is the data produced, and what are its limits?
Where the data comes from
Publicly displayed reviews and business profile details, collected through ReviewTactic's monitoring of Google and 100+ review platforms, held in a benchmark database of 1,000,000+ business profiles across the United States, Canada, Europe, Australia, New Zealand and Asia. Ratings, counts, categories and locations cover the database. Sentiment, complaint themes and policy flags run on review text we already hold: stored negative-review samples for about 460,000 businesses (roughly 20 reviews each, April to September 2026) and our scan dataset of 369,688 negative reviews from businesses that ran a scan (self-selected). Owner response rates are measured on businesses with 100 or more reviews and a website, from a sample of about 250 reviews per business in Google's default order, April to August 2026.
How numbers are produced
Every delivery states the population (how many businesses and reviews), the geography and categories, the platforms, the sample definition and the date measured. Sentiment and complaint themes come from an AI classifier that stores a category and a confidence score for each review; policy flags use the categories in Google's published content policy. Each figure traces to a saved query we can re-run if you or your editor asks.
Limits we state up front, every time
- A policy flag is not a removal. Google alone decides what comes down, and it rejects many requests. Nothing we send is a removal-rate claim.
- Two kinds of sample, never mixed: a population we define from our database for your request (every business of a category in a place, as listed on Google on the measurement date), or our published scan dataset, which is self-selected toward owners who suspected a problem. The method note says which one you are looking at.
- Where a profile's reviews are sampled rather than read in full, per-business counts are floors, not totals. We do not extrapolate.
- We do not report a figure for fewer than 10 businesses or 50 reviews; small cells are merged or shown as "too small".
- Category labels come from model output; near-duplicate labels are merged and small categories are grouped as "other".
- If the database does not carry a reliable field for a split you want, we do not invent one.
- Depth varies by place within the covered regions. If we cannot cover your geography at the depth you need, we tell you at acceptance.
Corrections
If we find an error after delivery, we tell you the same day with the corrected figure and what changed.
What we will not provide
- Nothing about individual reviewers: no names, profile links, photos, Local Guide levels, review histories or cross-venue patterns for identifiable people.
- No AI-inferred personal traits (nationality, gender, visitor type) about anyone. Our product estimates these for a business owner's own customers; they are not part of what journalists receive.
- No full review dumps. Public review text is quoted only as short excerpts, without reviewer names and without third-party personal details such as staff names or health information, and only to illustrate a theme the numbers already show.
- No ranked "worst business" lists built to order against a named target. Rankings cover a whole geography or category and always carry the method note. Named businesses get the same right of reply from you that any subject would.
- No data about a ReviewTactic customer, and no cut designed to single out a customer's competitor.
- For European (EU, EEA, UK, Switzerland), Canadian and Asian profiles, aggregates only: reviewer identifiers are dropped before analysis, no reviewer-level records leave our systems, excerpts are a single sentence that cannot identify a reviewer, and case working data is deleted 12 months after delivery. A reviewer's objection is honored by excluding the record.
- No predictions, no revenue estimates for named businesses, no claims about why a particular rating moved.
Disclosure: our commercial interest
ReviewTactic provides review management and review-removal services to local businesses and has a commercial interest in the review ecosystem. For any analysis we provide, we supply data and methodology only and have no role in the story's conclusions.
What that means in practice: we do not suggest headlines or angles; we report both the share of reviews that match a policy and the share that do not; we never promise or estimate removals; we check every named brand against our customer list before accepting a brief and tell you if there is an overlap; and we will say "we don't know" rather than fill a gap. You are welcome to print the sentence above alongside the data. How we handle data in general is on our trust page.
Data we have already published
- Free to cite
Google review policy violations: 2026 data study
369,688 negative reviews across 15,954 business profiles; 24.3% matched at least one Google content policy; spam was the largest category at 41.2% of the 89,694 flagged reviews.
Read the study → - Free to cite
How 6,721 Brooklyn restaurants compare on Google reviews
Rating benchmarks, category trends and competitive positioning from Google Maps data, June 2026.
Read the report → - Free to cite
Google rating benchmarks 2026
Rating percentiles for 1,511,957 Google Business Profiles by number of reviews, October 2026: median 4.6, and a perfect 5.0 falls from 61.3% under 10 reviews to 1.9% at 500 or more.
Read the study → - Free to cite
Google review statistics
A sourced, graded compilation of consumer-behavior and platform-enforcement figures, with every source linked.
See the statistics →
Want the same cut for your city or beat? That is what the Desk is for.
Request data for your story
About two minutes. We reply within one business day. Prefer email? Write to partner@reviewtactic.com.