Scoring Methodology

Exactly how ratings, sentiment, NPS and CSAT are calculated and what they do and do not mean.

Draft for legal review. This policy describes how the platform works today. It has not yet been approved by counsel and is not legal advice.

0.1 (draft)Policy version

Not yet in force. Takes effect when approved and published as finalEffective date

27 October 2026Last updated

Etiquette TechnologiesOperator

Scope

This page defines every score shown on Etiquette Pulse AI. Star ratings, AI or rule-based sentiment, NPS and CSAT are different measures and are not converted into one another.

Google rating

The rating and rating count on a business page are fetched live from Google Places and belong to Google. They are shown with attribution. We do not edit them.

Review sentiment on a business page

Up to five review excerpts returned by Google are read by our own text rules (a word-based method, not an external AI service). Each excerpt is labelled positive, neutral or negative using fixed thresholds. The share is shown only when enough excerpts are available. Google chooses which excerpts to return, so this is a small sample, not all reviews.

City sentiment score

For each city we sample well-known businesses (restaurants, hotels, hospitals and supermarkets) from public Google data and read the review excerpts as above. Score = 100 x (positive + half of neutral) / all excerpts read. The city Google rating is an average over the sampled businesses with at least 20 ratings, weighted by the logarithm of each business's rating count. This is an indicator, not CSAT, and not a ranking of cities' overall customer experience.

NPS

NPS = % promoters - % detractors, from survey answers of 0 to 10 (promoters 9-10, passives 7-8, detractors 0-6). It is shown only from genuine survey responses, with the number of responses and the period.

CSAT

CSAT = (satisfied responses / total valid responses) x 100. The definition of a satisfied response is shown beside the score. Shown only from genuine survey responses.

Missing data

If there is not enough real data, we say so and show no score. We never fill gaps with estimates or sample numbers.

Comparability

Scores gathered with different methods or sample sizes are not directly comparable. Minimum response thresholds will be published here once finalised.

Contact or complaint channel

Write to contact@etiquette.ltd. See also Grievance Redressal.