> For the complete documentation index, see [llms.txt](https://help.gleantap.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://help.gleantap.com/reputation/keywords-and-sentiment.md). # Keywords & Sentiment The Reputation module doesn't just collect reviews — it analyzes them. This page covers how sentiment is calculated, how keywords are extracted, and how to use both to understand what customers are really saying. For the basics of the Reputation module (dashboard, review responses), see Reputation Dashboard and Reviews. *** ## Sentiment Every review is classified as **positive**, **neutral**, or **negative** based on its star rating: | Star rating | Sentiment | | ----------- | ------------ | | 4-5 stars | **Positive** | | 3 stars | **Neutral** | | 0-2 stars | **Negative** | This applies to reviews from every connected source — Google, Facebook, Yelp, and internal surveys — as well as to NPS-style feedback captured through Gleantap's own review-request campaigns. **Where sentiment surfaces:** * **Reputation Dashboard** — counts of positive/neutral/negative reviews for the selected time window and filter set. * **Reviews page** — filter by sentiment to see only positive, only negative, or only neutral feedback. * **Reputation Insights** — trend charts of sentiment over time (see [Trends over time](#trends-over-time) below). **Why sentiment matters more than raw stars:** a 3.8 average tells you the mean; sentiment splits show you the shape. A business with 50% positive and 50% neutral looks very different from one with 60% positive and 40% negative — even if both average 3.8. *** ## Keywords **Gleantap uses AI to extract keywords from the text of every review.** When a new review comes in with more than a few words of content, an AI process reads it and identifies the topics being discussed — cleanliness, staff friendliness, class variety, pricing, wait times, etc. Extracted keywords are stored and counted across all your reviews. Over time, this builds a picture of what people mention most about your business. **Where keywords surface:** * **Keywords page** in the Reputation module — a ranked list of the most-mentioned keywords from your reviews, with counts. * **Filter reviews by keyword** — narrow the reviews list to only those mentioning a specific keyword. **What makes it into the keyword list:** * Reviews with meaningful text content (very short reviews without descriptive words are skipped). * Keywords that appear across enough reviews to be statistically interesting (typically 5+ mentions before showing up in the popular list). **How to use it:** * **Spot recurring praise.** If "friendly staff" is your #1 keyword with 200+ mentions, that's your positioning — lean into it in marketing. * **Spot recurring complaints.** If "wait time" appears 40 times in negative reviews, that's an operational signal. * **Compare across time.** If "cleanliness" was your top positive keyword in Q1 but drops off in Q2, something changed. * **Compare across locations.** In multi-location accounts, filter by location to see how keyword patterns differ between clubs. *** ## Trends over time The dashboard tracks two time series: **Average rating** — a rolling 12-month average of star ratings, shown as a line chart. This tells you whether your overall reputation is improving, stable, or declining. **Review volume** — the count of reviews received per month, shown alongside the rating trend. This tells you whether people are engaging (writing reviews) or going quiet. **Reading these together:** * **Rising rating + rising volume** = healthy growth. Happy customers, happy trajectory. * **Rising rating + falling volume** = quality is up but engagement is down. Consider more review-request outreach. * **Falling rating + rising volume** = something recently went wrong; more people are writing reviews (usually to complain). Investigate specific negative reviews from the recent period. * **Falling rating + falling volume** = disengagement + dissatisfaction. Immediate operational attention warranted. *** ## Filters and slicing All Reputation dashboards support these filters: * **Location** — for multi-location accounts, narrow to a single club or compare across * **Review source** — Google, Facebook, Yelp, Internal (from Gleantap-run surveys) * **Time range** — last 7/30/90 days, or a custom range * **Star rating** — filter to specific rating bands * **Sentiment** — positive, neutral, negative * **Reply status** — replied vs. unreplied Combine filters to slice — e.g., "negative Google reviews at Location A in the last 90 days" gives you a focused list of what to address. *** ## Requesting reviews You can automate review requests via Gleantap campaigns. Two channels are supported: * **Email** — send a review-request email to customers matching a segment (e.g., recent visitors who haven't been asked yet) * **SMS** — send a review-request SMS on the same pattern Configure the request under **Campaigns** with the review-request template. Responses land in the Reputation module and count toward your sentiment and keyword analysis just like external reviews. **NPS scoring** — Gleantap calculates a Net Promoter Score using standard methodology: `(Promoter % − Detractor %) × 100`, where 4-5 stars = promoter and 0-3 = detractor. Use the NPS trend to spot subtle shifts before they show up in raw averages. *** ## Responding to reviews For reviews from **Google** and **Facebook**, you can reply directly from Gleantap and your response is posted back to the source platform. 1. Open **Reputation → Reviews**. 2. Click the review you want to respond to. 3. Click **Reply**, write your response, and send. The response appears on Google or Facebook exactly as if you'd written it there. For Yelp and internal survey responses, replies are stored inside Gleantap. For repeat-response scenarios, set up **Saved Responses** in the Reputation settings so you don't retype the same reply. See Reviews for the reply UI in detail. *** ## Common scenarios ### "Which keywords should I care about most?" Cross-reference the keyword list against sentiment. A keyword that appears often in positive reviews is your strength; a keyword that appears often in negative reviews is your operational weakness. Focus on the intersection. ### "How can I see keywords for only one location?" Set the Location filter at the top of the Reputation module before opening the Keywords view. The keyword counts will reflect only reviews for that location. ### "The AI extracted a keyword that doesn't make sense" Keyword extraction is heuristic and can occasionally surface irrelevant terms — especially for short reviews with unusual phrasing. If a specific keyword is misleading you, mentally filter it out; there's no manual keyword-exclusion list in the UI today. If a keyword type is consistently unhelpful, email with examples. ### "I want to see only reviews that mention 'friendly staff'" Open Reviews, use the keyword filter (if available) or search within the reviews list for that phrase. Combine with the positive-sentiment filter to see the reviews you'd want to highlight in marketing. ### "Our NPS dropped last month — how do I find out why?" Filter to reviews from the last month with **negative** sentiment. Look at the top keywords in that filtered set — those are the topics driving the drop. Cross-check against recent operational changes (a new staff member, a schedule change, a price increase). ### "Can I export the keyword list or the reviews for analysis?" Some views support export via a menu action. If you need a specific format, contact . *** ## Related pages * Setting up Reputation — connecting Google, Facebook, Yelp * Dashboard — the main Reputation view * Reviews — responding to reviews * Settings — Saved Responses and other configuration * Ratings & Reviews — the account-level reputation report --- # Agent Instructions This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com. ## Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter: ``` GET https://help.gleantap.com/reputation/keywords-and-sentiment.md?ask=&goal= ``` `ask` is the immediate question: it should be specific, self-contained, and written in natural language. `goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.