LLM Prompting Guide for SEO

LLM Prompting Guide for SEO

Essential techniques and patterns for prompting LLMs in SEO automation. Part of the Building Agentic SEO Consultants course. Use this with Module 2: Python for SEO Automation (code generation), Module 3: APIs & JSON (workflows and API prompts), and Module 4: Building SEO Agents (goal and constraint prompts).

Core Prompting Principles

1. Be Specific and Detailed

Bad: “Write Python code for SEO” Good: “Create a Python script that reads Google Search Console CSV data, calculates CTR by keyword, and exports the top 20 performing keywords to a new CSV file”

2. Include Context and Requirements

  • Specify the SEO tool (Ahrefs, SEMrush, GSC)
  • Mention data formats and column names
  • Include performance requirements
  • Specify output formats

3. Request Comments and Documentation

Always add: “Please include detailed comments explaining each step of the code”

Essential Prompt Templates

Python Script Generation

Create a Python script that:
- Reads [SEO_TOOL] data from [FILE_FORMAT]
- Performs [SPECIFIC_ANALYSIS]
- Exports results to [OUTPUT_FORMAT]
- Includes detailed comments for each step
- Handles common data issues like missing values

Data Analysis Prompts

Analyse [SEO_DATA_TYPE] data to:
- Identify [SPECIFIC_PATTERNS]
- Calculate [METRICS]
- Create [VISUALISATIONS]
- Export [FORMATTED_RESULTS]

Automation Workflow Prompts

Create an automated workflow that:
- Connects to [DATA_SOURCE]
- Processes [DATA_TYPE]
- Applies [ANALYSIS_METHODS]
- Sends [NOTIFICATIONS/REPORTS]
- Runs [SCHEDULE]

SEO-Specific Prompting Techniques

1. Keyword Analysis Prompts

"Create a Python script to analyze keyword data from Ahrefs that:
- Groups keywords by search intent
- Calculates keyword difficulty scores
- Identifies content gaps
- Exports prioritised keyword lists
- Includes data validation and error handling"

2. Technical SEO Prompts

"Build a technical SEO audit tool that:
- Crawls website URLs
- Checks for common technical issues
- Generates prioritised fix recommendations
- Creates executive summary reports
- Handles large site crawls efficiently"

3. Content Optimization Prompts

"Develop a content analysis system that:
- Analyses content performance metrics
- Identifies optimization opportunities
- Suggests content improvements
- Tracks content ROI
- Integrates with multiple data sources"

Advanced Prompting Strategies

1. Chain of Thought Prompting

"Think step by step:
1. First, identify the data requirements
2. Then, design the analysis approach
3. Next, implement the solution
4. Finally, add error handling and validation"

2. Role-Based Prompting

"You are an expert SEO data analyst. Create a comprehensive solution for [SPECIFIC_TASK] that follows industry best practices and handles edge cases."

3. Iterative Refinement

"Based on the previous code, improve it by:
- Adding [SPECIFIC_FEATURE]
- Optimizing [PERFORMANCE_ASPECT]
- Enhancing [USER_EXPERIENCE]
- Fixing [IDENTIFIED_ISSUES]"

Common Prompting Mistakes to Avoid

1. Vague Requests

  • ❌ “Make it better”
  • ✅ “Add error handling for missing data and improve performance for files over 10MB”

2. Missing Context

  • ❌ “Create a dashboard”
  • ✅ “Create a Streamlit dashboard for Google Search Console data with filters for date range, device type, and country”

3. No Output Specifications

  • ❌ “Analyse the data”
  • ✅ “Analyse the data and export a CSV with columns: keyword, clicks, impressions, CTR, and position”

SEO Tool-Specific Prompts

Google Search Console

"Create a Python script for Google Search Console data that:
- Handles the standard GSC export format
- Processes query, page, and country data
- Calculates performance metrics
- Filters by date range and device type"

Ahrefs Data

"Build an Ahrefs data processor that:
- Reads Ahrefs keyword export files
- Handles different export formats
- Calculates keyword metrics
- Identifies ranking opportunities"

SEMrush Data

"Create a SEMrush data analyser that:
- Processes SEMrush CSV exports
- Handles encoding issues
- Calculates competitive metrics
- Generates actionable insights"

Prompting for Different Outputs

Code Generation

  • Specify programming language
  • Include error handling requirements
  • Request documentation and comments
  • Mention performance considerations

Data Analysis

  • Define metrics and calculations
  • Specify visualisation requirements
  • Include data validation steps
  • Request export formats

Automation Workflows

  • Define triggers and schedules
  • Specify integration requirements
  • Include monitoring and alerting
  • Request deployment instructions

Best Practices Summary

  1. Be Specific: Include exact requirements and constraints
  2. Provide Context: Explain the SEO use case and data sources
  3. Request Documentation: Always ask for comments and explanations
  4. Include Error Handling: Specify how to handle common issues
  5. Test and Iterate: Refine prompts based on results
  6. Use Examples: Provide sample data or expected outputs
  7. Specify Formats: Include input/output format requirements

This guide supports Module 2: Python for SEO, Module 3: APIs & JSON, Module 4: Building SEO Agents, and Module 5: Data Analysis.

Published on 29 January 2025

You Might Also Like