Brighton Summary - October 2024

How to Optimise 3,500 Product Descriptions for eCommerce in One Day Using ChatGPT

Key Takeaways

  1. Custom GPTs for Product Descriptions:

    • Custom GPTs are personalized versions of ChatGPT trained with specific data and instructions.
    • Creating custom GPTs for each product brand improves content quality and reduces hallucinations.

  2. Creating Custom GPTs:

    • Scrape product & category descriptions and 'About Us' pages for training data.
    • Provide well-defined instructions and context to the GPT.
    • Instruct the GPT to avoid hallucinations and ask for more information when needed.

  3. Prompt Structure for Product Descriptions:

    • Include product name, brand context, and language specifications.
    • Request specific sections like product info, use cases, and target audience.
    • Format instructions (e.g., HTML, subheadings) and SEO optimization guidelines.

  4. Results of Custom GPT Usage:

    • 400 products optimized in 5 weeks.
    • 115 more words in position 1 in Google.
    • 156% more organic clicks after 5 weeks.

  5. Automating the Process with Make.com:

    • Connect apps like Google Sheets with ChatGPT to build automated workflows.
    • No coding knowledge required.
    • Achieved 3,500 product descriptions in one day or 40 descriptions in 5 minutes.

  6. Make.com Workflow:

    • Use Google Sheets for input data (URL, Product name, short info).
    • Set up ChatGPT interactions for long and short descriptions.
    • Output results to separate Google Sheets.

  7. Importing Optimized Descriptions:

    • For WordPress sites, use All Import - WooCommerce Import Add-On Pro plugin.

Action Items

  1. Set Up Custom GPTs:

    • Use Screaming Frog to scrape relevant content for training data.
    • Create custom GPTs for each product brand using ChatGPT's GPT editor.

  2. Develop Effective Prompts:

    • Structure prompts with clear instructions, context, and formatting guidelines.
    • Include examples in prompts for better results.

  3. Test and Refine:

    • Start by optimizing a small batch of products to test the custom GPT's effectiveness.
    • Analyze results and refine prompts as needed.

  4. Implement Make.com Automation:

    • Set up a Make.com account and create a new scenario.
    • Connect Google Sheets and ChatGPT modules.
    • Configure the workflow to read input data, generate descriptions, and output results.

  5. Prepare Data Sheets:

    • Create input Google Sheets with necessary product information.
    • Set up output sheets for long and short descriptions.

  6. Run and Monitor Automation:

    • Start with a small batch to ensure everything works correctly.
    • Scale up to larger product sets once confident in the process.

  7. Import Optimized Descriptions:

    • For WordPress/WooCommerce sites, set up the All Import plugin.
    • Configure import settings to match your data structure.

  8. Continuous Improvement:

    • Regularly review generated descriptions for quality and accuracy.
    • Update custom GPTs and prompts based on performance and any changes in brand guidelines.

Remember: While this process can significantly speed up product description optimization, always maintain human oversight to ensure quality and brand consistency.

For more information and detailed guides:

  • Website: katarinadahlin.com
  • LinkedIn: Linkedin.com/in/katarina-dahlin
  • Twitter: @KataDahlin
  • YouTube: Youtube.com/@KatarinaDahlinSEO
  • Presentation Deck: Speakerdeck.com/katarinadahlin/brightonseo-2024