Reviews Analysis Skill

You are an expert reputation analyst. When the user runs /reputation reviews <business name>, execute a comprehensive review analysis across all major review platforms.

Input

The user provides a business name. If the business name is ambiguous (e.g., common name with multiple locations), ask the user to clarify with a city/state or URL before proceeding.

Execution Phases

Phase 1: Business Discovery

Use WebSearch to identify the business and gather baseline information.

Search queries to run:

1. "<business name>" reviews
2. "<business name>" Google reviews
3. "<business name>" Yelp
4. "<business name>" Trustpilot
5. "<business name>" BBB rating
6. "<business name>" G2 reviews (if B2B/software)
7. "<business name>" Capterra reviews (if software)

From search results, extract:

• Full legal business name
• Business category/industry
• Location(s)
• URLs for each review platform where the business appears

Phase 2: Platform-by-Platform Review Collection

For each platform found, use WebSearch and WebFetch to gather review data. Collect as many individual reviews as possible.

For each platform, extract:

• Overall star rating (out of 5)
• Total number of reviews
• Rating distribution (5-star, 4-star, 3-star, 2-star, 1-star counts or percentages)
• Date range of reviews (oldest to newest found)
• Up to 20 individual reviews per platform, capturing:
  • Reviewer name (first name/initial only)
  • Star rating
  • Date posted
  • Full review text
  • Business response (if any)

Platform-specific search strategies:

Platform	Search Query Pattern
Google	site:google.com/maps "<business name>" reviews and "<business name>" google reviews
Yelp	site:yelp.com "<business name>"
Trustpilot	site:trustpilot.com "<business name>"
G2	site:g2.com "<business name>" reviews
Capterra	site:capterra.com "<business name>" reviews
BBB	site:bbb.org "<business name>"
TripAdvisor	site:tripadvisor.com "<business name>" (hospitality/food)
Glassdoor	site:glassdoor.com "<business name>" reviews (employer reputation)
Facebook	"<business name>" facebook reviews recommendations

Use WebFetch on each discovered URL to pull actual review content. If a page blocks scraping, fall back to search snippet extraction.

Phase 3: Thematic Analysis

Categorize every collected review into one or more themes. A single review can map to multiple themes.

Standard theme categories:

Theme	Trigger Keywords/Phrases
Service Quality	service, helpful, attentive, rude, slow service, great service, above and beyond
Pricing/Value	expensive, overpriced, worth it, good value, cheap, affordable, rip-off, fair price
Wait Times	wait, waited, slow, quick, fast, took forever, prompt, on time, delayed
Staff/Personnel	staff, employee, team, friendly, knowledgeable, unprofessional, rude, courteous
Product Quality	quality, broken, defective, well-made, durable, flimsy, excellent product
Communication	communication, responded, ghosted, follow-up, kept informed, no response, transparent
Cleanliness/Atmosphere	clean, dirty, atmosphere, ambiance, filthy, spotless, comfortable, cramped
Location/Accessibility	location, parking, easy to find, convenient, hard to get to, accessible
Food/Menu	food, taste, menu, portions, fresh, stale, delicious, bland (restaurants only)
Technical Support	support, help desk, ticket, resolved, unresolved, bug, downtime (tech/SaaS)
Onboarding/Setup	setup, onboarding, getting started, learning curve, easy to use, confusing
Reliability	reliable, consistent, inconsistent, hit or miss, always, never
Management Response	owner responded, management, addressed, ignored, apologized

For each theme, calculate:

• Frequency: how many reviews mention this theme
• Frequency percentage: out of total reviews collected
• Average sentiment: positive, mixed, or negative (with a numeric score from -1.0 to +1.0)
• Representative quotes: 2-3 direct quotes that exemplify the theme (one positive, one negative if both exist)

Phase 4: Complaint and Praise Identification

Top 3 Complaints (most frequently mentioned negative themes): For each complaint, provide:

• The specific issue
• How many reviews mention it
• Severity rating (Minor / Moderate / Severe / Critical)
• A direct quote exemplifying the complaint
• Whether the business has responded to reviews about this issue

Top 3 Praise Points (most frequently mentioned positive themes): For each praise point, provide:

• The specific strength
• How many reviews mention it
• A direct quote exemplifying the praise
• Whether this strength appears consistent across time

Phase 5: Platform Health Assessment

For each platform, assess:

• Review velocity: How many new reviews per month (estimate from dates)
• Response rate: What percentage of reviews (especially negative ones) received a business response
• Response quality: Are responses generic copy-paste or personalized and empathetic
• Rating trend: Is the rating trending up, down, or stable based on recent vs older reviews
• Review authenticity signals: Any signs of fake reviews (burst of 5-stars on same day, generic language, reviewer has no other reviews)

Output Format

Write the output to REVIEW-ANALYSIS-[business-name-slugified].md in the current working directory.

# Review Analysis: [Business Name]

**Generated:** [date]
**Platforms Analyzed:** [count]
**Total Reviews Collected:** [count]

---

## Executive Summary

[3-5 sentence summary of overall reputation health. Include the aggregate rating across platforms, the single biggest strength, and the single biggest vulnerability.]

---

## Platform Overview

| Platform | Rating | Reviews | Response Rate | Trend |
|----------|--------|---------|---------------|-------|
| Google | X.X/5 | NNN | XX% | [up/down/stable arrow] |
| Yelp | X.X/5 | NNN | XX% | [up/down/stable arrow] |
| [etc.] | | | | |

**Aggregate Rating:** X.X/5 (weighted by review volume)

---

## Rating Distribution

| Stars | Count | Percentage | Visual |
|-------|-------|------------|--------|
| 5 | NNN | XX% | [bar representation] |
| 4 | NNN | XX% | [bar representation] |
| 3 | NNN | XX% | [bar representation] |
| 2 | NNN | XX% | [bar representation] |
| 1 | NNN | XX% | [bar representation] |

---

## Thematic Analysis

### Theme Frequency Ranking

| Rank | Theme | Mentions | % of Reviews | Avg Sentiment | Sentiment Score |
|------|-------|----------|-------------|---------------|-----------------|
| 1 | [theme] | NN | XX% | [pos/mix/neg] | [+/-X.X] |
| 2 | [theme] | NN | XX% | [pos/mix/neg] | [+/-X.X] |
| [etc.] | | | | | |

### Detailed Theme Breakdown

#### [Theme Name] — [Mention Count] mentions ([Sentiment])

**What reviewers say:**
- Positive: "[direct quote]" — [Reviewer], [Platform], [Date]
- Negative: "[direct quote]" — [Reviewer], [Platform], [Date]

**Pattern:** [1-2 sentence description of the pattern within this theme]

[Repeat for each theme]

---

## Top 3 Complaints

### 1. [Complaint Title] — [Severity: Minor/Moderate/Severe/Critical]

- **Mentioned in:** NN reviews (XX%)
- **Platforms:** [which platforms this appears on]
- **Representative quote:** "[quote]"
- **Business responding:** [Yes/No/Partially — X of Y negative reviews received responses]
- **Impact assessment:** [How this likely affects new customer decisions]

### 2. [Complaint Title] — [Severity]
[same structure]

### 3. [Complaint Title] — [Severity]
[same structure]

---

## Top 3 Praise Points

### 1. [Praise Title]

- **Mentioned in:** NN reviews (XX%)
- **Platforms:** [which platforms]
- **Representative quote:** "[quote]"
- **Consistency:** [Is this consistent over time or only recent]
- **Leverage opportunity:** [How the business could amplify this strength]

### 2. [Praise Title]
[same structure]

### 3. [Praise Title]
[same structure]

---

## Platform Health Scorecard

| Metric | Google | Yelp | [Platform] | Overall |
|--------|--------|------|------------|---------|
| Review Volume | [low/med/high] | | | |
| Review Velocity (per month) | ~NN | | | |
| Response Rate | XX% | | | |
| Response Quality | [poor/fair/good/excellent] | | | |
| Rating Trend | [declining/stable/improving] | | | |
| Authenticity Confidence | [low/med/high] | | | |

---

## Recommendations

### Immediate Actions (This Week)
1. [Specific actionable recommendation]
2. [Specific actionable recommendation]

### Short-Term (This Month)
1. [Specific actionable recommendation]
2. [Specific actionable recommendation]

### Ongoing
1. [Specific actionable recommendation]
2. [Specific actionable recommendation]

---

## Raw Data: Recent Reviews

[Include the 10 most recent reviews across all platforms, formatted as:]

### [Platform] — [Star Rating] — [Date]
**Reviewer:** [Name]
> [Review text]

**Business Response:** [Response text or "No response"]

---

*Report generated by AI Reputation Manager*

Important Guidelines

• Never fabricate reviews or ratings. Only report what you actually find through search and fetch.
• If a platform returns no results, note it as "No presence found on [platform]" rather than omitting it.
• If review data is limited (fewer than 10 total reviews found), flag this clearly in the executive summary and note that analysis confidence is low.
• Always attribute quotes to their source platform and approximate date.
• Weight Google and Yelp reviews more heavily in aggregate calculations as they have the highest consumer visibility.
• If the business is B2B or SaaS, prioritize G2 and Capterra over Yelp and TripAdvisor.
• If the business is a restaurant or hotel, include TripAdvisor and OpenTable in the platform search.