Data source
Every figure in a Rate My Ratings report comes from public reviews on Google Maps. Reviews are public, attributed to the business that received them, and presented in aggregate.
We collect reviews using established scraping tools that respect Google's public surface. We do not authenticate as a Google user, we do not collect anything that requires a login, and we do not pay for elevated API access. Every byte we read is what an unauthenticated visitor to the same Google Maps page would see.
What we include
For every business in the competitor set, we collect the following:
- Business name, address, category, phone number where public, and website where listed.
- Average star rating and total review count, as Google displays them.
- Every public review with text, star rating (1–5), posted date, and language.
- The business's reply to the review, where one was posted, with its date.
The competitor set is defined by category and a 3–5 mile radius from your postcode, with a minimum of 12 comparable businesses. The competitor list appears in full on page 03 of every report.
What we exclude
We exclude, on principle:
- Reviewer personal data beyond the public attribution shown on the review (display name only — never email, phone, or any private contact).
- Anything behind a login. No Google account, no authenticated profiles, no private business dashboards.
- Reviews flagged or removed by Google between collection and publication, even if our snapshot pre-dates the removal.
- Reviews from accounts with no other public activity and a single review on the business in question, where the pattern fits a known cluster of suspected fakes. Removed reviews are noted in the methodology page of the delivered report.
Sample size
A typical urban report covers between 20 and 80 competing businesses and between 200 and 2,000 reviews. The minimum we consider statistically meaningful is 12 businesses with at least 10 reviews each.
| Market type | Businesses | Reviews |
|---|---|---|
| Single-postcode small market | 12–25 | 150–500 |
| Mid-density urban category | 25–60 | 500–1,500 |
| Dense regional category | 60–120 | 1,500–4,000 |
If your market falls below the minimum, we tell you before charging. Roughly one in twenty requested markets is below the threshold; in those cases we suggest a wider radius or refund any deposit.
Freshness
Each report is built on a snapshot taken at the start of the work. The cover page records the exact date of collection. We typically deliver within two working days of collection, which means the data in the delivered report is between two and four days old.
We do not refresh the snapshot after delivery. If you would like the report re-run a year later — to see how your numbers have moved — you commission a new report. There is no subscription and no automatic re-delivery.
Bias and limits
Public review data has known biases. We don't pretend otherwise; the methodology page of every delivered report flags the same limits, in the same language. The most important to be aware of:
- Reviews skew to the extremes. Customers leave reviews when they're delighted or annoyed; the middle is under-represented. The distribution analysis on page 07 reflects this and frames the comparison accordingly.
- Volume varies sharply by category. Restaurants attract many more reviews per customer than, say, accountants. Cross-category comparisons are not meaningful and we do not run them.
- Non-English reviews are translated for analysis. The original is preserved; the analysis runs on a machine translation. Where translation confidence is low, the review is excluded from theme analysis and noted.
- Themes are identified by structured read of every review, not by keyword count. Each review is tagged against a fixed taxonomy of themes built for the category, with a human spot-check on a 10% sample of every report.
How we analyse
A report has two halves: a calculated half, drawn directly from the reviews, and a written half, generated by a language model from a representative sample of review text plus the calculated stats. Every report's Appendix section repeats this in the same language.
Calculated figures
Ratings, volume, distribution, per-business sentiment, aspect-level sentiment, dominant-emotion counts, competitor benchmark, top phrases, and the competitive map are all computed from the reviews directly. Sentiment is scored automatically — positive, neutral, or negative — at both the review level and the aspect level. Consistency reflects how tightly a business's ratings cluster around its average; higher numbers mean a more uniform experience.
Written paragraphs
The executive summary, market-structure narrative, sentiment themes, customer-language tags, aspect derivation, the synthesised customer persona and its Q&A, the persona NPS reading, the SWOT, the strategic priorities, and the recommended actions are written by a language model. Each generation runs on a representative sample of review text together with the calculated stats; any section may be omitted when the call fails or the data is insufficient.
The synthesised persona
Every report includes a Persona section: a portrait of the market's composite customer, written from a representative sample of reviews across all businesses in the market. It is not a real person, a forecast, or a statistical aggregate; it is the review base's centre of gravity, rendered as a single voice.
The persona also answers a short interview and the standard 0–10 recommendation question. The NPS shown is the persona's own answer to that question — not a directly-measured survey result, and not a claim about your real customer base.
The Strategy section
The Strategy section contains a SWOT, a list of strategic priorities scored for impact and urgency, and a set of recommended-action cards with timeline, key activities, success metrics, estimated resources, and risk mitigation. These are written by a language model from the calculated stats and the themes it surfaced.
Treat them as a structured starting point for an internal conversation. They are grounded in the data we collected, but they are not bespoke advice and they are not a substitute for your judgement.
What we do not do
The list below is non-negotiable. If we ever change any of these, this page changes first.
- We do not contact reviewers, ever, under any circumstance.
- We do not write fake reviews, suppress real reviews, or advise on either.
- We do not sell data on to third parties. The reviews collected for your report are used to produce your report and are then retained only so we can answer questions about it.
- We do not respond to reviews on your behalf.
- We do not run reputation-management or review-removal services. If that's what you need, we'll point you elsewhere — but it isn't us.
If you want to verify any specific figure or interpretation in a delivered report, email info@ratemyratings.com. We answer methodology questions before, during, and after a purchase.