Reviewen Randy - Complete Documentation
Reviewen Randy - Complete Documentation
AI-Powered SEO Meta Auditor for Google Sheets
Table of Contents
- Introduction
- Getting Started
- Understanding Lil Randy vs Full Randy
- Reading Your Results
- What Randy Checks
- Scoring System
- Best Practices
- Troubleshooting
- Technical Details
Introduction
Reviewen Randy is a Google Sheets add-on that audits meta titles and descriptions for SEO quality and technical compliance. It combines code-based validation (100% accurate) with AI-powered quality assessment.
Key Features:
- ✅ Instant technical validation (length, double spaces, prohibited terms)
- ✅ AI quality review (CTAs, clarity, keyword usage)
- ✅ Batch processing (up to 50 pages in seconds)
- ✅ Australian English compliance
- ✅ Weighted scoring system
- ✅ Review history with navigation
Getting Started
Step 1: Prepare Your Sheet
Have a sheet named "export" (or configure custom name) with these columns:
URL | Title | Description | H1 |
|---|---|---|---|
https://example.com/ | Example Page - Brand | Visit Example for... | Welcome to Example |
Required columns (case-insensitive):
- URL - Full page URL
- Title - Meta title
- Description or Desc - Meta description
- H1 - Page H1 heading
Step 2: Open Randy
- Click Reviewen Randy menu in Google Sheets
- Select "Open Sidebar"
- The sidebar will appear on the right
Step 3: Run Your Review
Choose your review type:
Lil Randy (Recommended for most audits)
- Fast batch analysis (~5 seconds for 50 pages)
- Objective violations only
- Binary checks (yes/no issues)
Full Randy (For detailed feedback)
- Deep per-page review (~2 minutes for 15 pages)
- Includes writing suggestions
- Quality assessments
Understanding Lil Randy vs Full Randy
Lil Randy - Quick Batch Checker
Best for:
- Regular quality checks
- Large page sets (up to 50)
- Finding technical violations
- Pre-publish validation
What it checks:
- ✅ Missing CTAs (binary: present or not)
- ✅ Separator inconsistency
- ✅ AU spelling errors
- ✅ Code validates: length, double spaces, prohibited terms, duplicates
What it DOESN'T do:
- ❌ Subjective quality assessments
- ❌ Writing suggestions
- ❌ "Could be better" feedback
Example output:
- "Missing CTA" (binary violation)
- "Double space in title" (technical error)
- NOT: "Description could be more compelling" (subjective)
Full Randy - Comprehensive Review
Best for:
- New page creation
- High-value pages (homepage, key landing pages)
- Content improvement projects
- Learning better meta writing
What it checks:
- ✅ Everything Lil Randy checks
- ✅ CTA strength and effectiveness
- ✅ Keyword placement and usage
- ✅ Clarity and alignment
- ✅ Compelling vs generic language
- ✅ Brand voice consistency
What you get:
- Technical violations
- Quality assessments
- Specific rewrite suggestions
- Strategic recommendations
Example output:
- "Missing CTA" (violation)
- "CTA present but could be stronger - consider 'Get expert advice today' instead of 'Learn more'" (quality)
- "Description is generic - add specific benefit or unique value proposition" (suggestion)
Reading Your Results
The Results Card
Randy's Message:
- 🟢 "Looking good, mate!" = 80-100 score
- 🔵 "Not too shabby!" = 60-79 score
- 🟡 "Bit of work needed..." = 40-59 score
- 🔴 "No good, might need to read this" = 0-39 score
Top Level Issues
Common patterns affecting multiple pages:
Top Level Issues
├─ 3 pages with double spaces
├─ 5 titles exceed 700px
└─ 2 pages missing CTAs
URL by URL Issues
Individual page problems with priority colors:
🔴 High Priority (Red background)
- Double spaces
- Spelling errors / typos
- Prohibited terms
- Other errors
Fix these FIRST - they're objective mistakes that hurt credibility.
🟡 Medium Priority (Yellow background)
- Title too long
- Description too long
- Separator issues
⚪ Low Priority (Gray background)
- Missing H1
- Quality suggestions
- Other recommendations
Navigation
Use the navigation bar to review past audits:
[← Previous] Review 4 of 4 [LATEST] [Next →]
- Latest review = Always shown first (highest number)
- Previous = Go back in time to older reviews
- Next = Go forward to newer reviews
Toggle H1 Issues
Click "Show H1 Issues" / "Hide H1 Issues" to filter results:
- Hide: Focus on title/description only
- Show: Include all H1 problems
What Randy Checks
Code-Based Validation (100% Accurate)
These checks are performed by code - never hallucinated, always reliable:
1. Length Check
- Title: Flags if > 700px (~65+ chars)
- Target: 55 chars / 600px
- Format: "Title too long - Current: 68 chars (705px), Recommended: 55 chars (600px)"
- Description: Flags if > 1050px (~165+ chars)
- Target: 155 chars / 920px
- Format: "Description too long - Current: 165 chars (998px), Recommended: 155 chars (920px)"
Why pixels? Different characters have different widths. "WWW" is wider than "iii". Randy counts pixels for accurate SERP display.
2. Double Spaces
- Detects: Two or more consecutive spaces
- Format: "Double space in title" or "Double space in description"
- Impact: -10 points (high severity)
3. Prohibited Terms
Exact whole-word matches (case-insensitive):
- "top notch" or "top-notch"
- "superior"
- "swift"
- "delve"
Format: Contains prohibited term: "top-notch" Impact: -10 points (high severity)
4. Duplicate Descriptions
- Flags: 100% identical descriptions across pages
- Format: "4 pages have identical descriptions: 'Barry Nilsson is a leading...'"
AI-Based Validation
5. Missing CTA (Binary)
Lil Randy: Only flags if NO call-to-action exists Full Randy: Flags missing CTAs AND suggests improvements
Valid CTAs:
- "Enquire online"
- "Contact us"
- "Get in touch"
- "Book now"
- "Call today"
❌ Lil Randy will NOT flag: "Weak CTA" or "Could be better" ✅ Full Randy WILL provide: CTA strength feedback
6. Separator Consistency
Flags if DIFFERENT separators are mixed across titles:
Valid separators:
- - (hyphen)
- – (en dash)
- — (em dash)
- | (pipe)
✅ OK: All titles use - ✅ OK: All titles use | ❌ FLAG: 5 titles use -, 3 titles use |
7. AU Spelling
Flags actual misspellings and US spelling:
Common AU vs US:
- ❌ organize → ✅ organise
- ❌ color → ✅ colour
- ❌ center → ✅ centre
8. Quality Assessments (Full Randy Only)
- Keyword placement and usage
- Clarity and alignment with page intent
- Compelling vs generic language
- Brand voice consistency
- Location mention appropriateness
Scoring System
Randy uses weighted scoring based on issue severity:
Point Deductions
High Severity (-10 points each):
- Double spaces
- Prohibited terms
- Spelling errors
- Typos
Medium Severity (-5 points each):
- Title too long (>700px)
- Description too long (>1050px)
- Separator inconsistency
Low Severity (-3 points each):
- Missing H1
- Templated descriptions
- Quality suggestions
Score Ranges
Score | Message | Action |
|---|---|---|
90-100 | Looking good, mate! | ✅ Publish-ready |
80-89 | Looking good, mate! | ✅ Minor tweaks optional |
60-79 | Not too shabby! | 🔵 Review medium issues |
40-59 | Bit of work needed... | 🟡 Fix before publishing |
0-39 | No good, might need help | 🔴 Major revisions needed |
Score Examples
Example 1: Perfect Score
- 0 issues = 100/100 ✅
Example 2: One Double Space
- 1 double space = 90/100 (-10 points)
Example 3: Mixed Issues
- 1 typo (-10) + 2 length issues (-10) + 1 H1 (-3) = 77/100
Example 4: Critical Errors
- 3 double spaces (-30) + 2 prohibited terms (-20) = 50/100
Best Practices
When to Use Lil Randy
✅ Use for:
- Regular batch audits
- Pre-publish checks
- Finding technical violations
- Large page sets (20-50 pages)
- Quick quality validation
⏱️ Processing time: 3-8 seconds
When to Use Full Randy
✅ Use for:
- New page creation
- Homepage and key landing pages
- Content improvement projects
- Learning better meta writing
- Smaller focused sets (5-15 pages)
⏱️ Processing time: 30 seconds - 2 minutes
Meta Writing Tips
Titles:
- Keep ~55 chars / 600px target
- Use one separator consistently
- Include primary keyword once
- Match page intent
- Format: Keyword/Service – Brand or Brand – What You Do
Descriptions:
- Keep 130-155 chars / 920px target
- Include keyword naturally once
- Add clear CTA
- Include USP/benefit
- Be specific, not generic
Example - Good:
Title: Family Law Solicitors Melbourne – Smith Legal
Desc: Expert family law advice for divorce, custody, and property settlements in Melbourne. Enquire online for a confidential consultation.
Example - Bad:
Title: Smith Legal | Family Law | Melbourne | Divorce | Custody | Property | Experts
Desc: We are the best law firm in Melbourne. We provide superior service. Contact us today to learn more about our services.
Review Workflow
- Export your pages to Google Sheets
- Run Lil Randy for quick check
- Fix all red issues (double spaces, typos, prohibited terms)
- Fix medium issues (length, separators)
- Run Full Randy on key pages for quality feedback
- Iterate and improve
- Re-run Lil Randy to verify all issues resolved
Troubleshooting
"No rows provided" Error
Issue: Sheet name not found
Fix:
- Ensure sheet is named "export" (case-insensitive)
- OR set custom sheet name in Script Properties: INPUT_SHEET_NAME
"Missing input sheet" Error
Issue: Required columns not found
Fix: Ensure your sheet has columns named (case-insensitive):
- URL
- Title
- Description (or Desc)
- H1
"OPENAI_API_KEY not found" Error
Issue: API credentials not configured
Fix:
- Verify GCP_PROJECT_NUMBER is set in Script Properties
- Ensure OPENAI_API_KEY exists in Secret Manager
- Check IAM permissions are configured correctly
Results Not Showing
Issue: Empty results or "No results yet"
Fix:
- Click Refresh button
- Check if review actually ran (look for "done" status)
- Try closing and reopening sidebar
Webmaster - Permissions
You need sign out of in your browser
AI Hallucinating Issues
This shouldn't happen! Randy uses code for all binary checks.
If you see false positives:
- Check if it's in a subjective category (Full Randy quality assessment)
- Report in #seo-help with example
- We may need to move more checks to code
Technical Details
Models Used
- Lil Randy: gpt-5-nano or gpt-4o-mini
- Full Randy: gpt-5-mini or gpt-4o-mini
Set via Script Properties: OPENAI_MODEL
Configuration Options
Script Properties (optional):
Property | Default | Description |
|---|---|---|
OPENAI_MODEL | gpt-5-mini | AI model to use |
INPUT_SHEET_NAME | export | Input sheet name |
RESULT_SHEET_NAME | _ai_results | Results sheet name (hidden) |
MAX_LOCAL | 3 | Max rows for basic review |
PROMPT_MODE | json | Prompt style (json/randy) |
GCP_PROJECT_NUMBER | - | Required for Secret Manager |
Page Limits
- Lil Randy: Up to 50 pages
- Full Randy: Up to 15 pages
- Basic review: Configurable via MAX_LOCAL (default: 3)
Results Storage
Results are stored in hidden sheet _ai_results with:
- URL
- Status
- Timestamp
- Model used
- Raw AI response
- Parsed results
- Review score (weighted)
Each batch creates a new BATCH_SUMMARY_V# entry.
Review History
Randy keeps all past reviews. Navigate using:
- Previous = Older reviews
- Next = Newer reviews
- Latest badge = Most recent review
Support
Questions? Ask in #seo-help
Bug reports? Include:
- What you were doing
- Error message (if any)
- Expected vs actual behavior
- Screenshot if helpful
Feature requests? We're always improving Randy!
Changelog
Current Version: Workspace Edition with Secret Manager
Recent updates:
- ✅ Code-based validation for length, double spaces, prohibited terms
- ✅ Weighted scoring system (high/medium/low severity)
- ✅ Review history with navigation
- ✅ Improved sidebar with priority colors
- ✅ "No issues found" success state
- ✅ Lil Randy vs Full Randy separation
- ✅ Toggle H1 issues visibility