Prerequisites

• Accounts & Tools:
  • Active ChatGPT account with image input/editing (GPT-4o enabled).
  • Active Claude (or equivalent advanced language model) account.
  • Access to Google Sheets (or similar) for managing data.
• Data Sources:
  • CSV files with 300–2000 positive and negative product reviews (from Helium 10, Jungle Scout, etc.).
  • Demographic data. You can get that from ChatGPT if you don’t have any or if you are an Amazon seller, from Amazon Seller Central’s Brand Analytics (Demographics tab; select the quarter with the highest sales).
  • Product images (ideally with a white background).

Procedure

Step 1: Data Collection & Preparation

1. Collect Reviews:
   • Download 300–2000 positive and negative reviews from your own and competitor listings.
   • Save separate CSV files for positive and negative reviews.
2. Clean Data:
   • Remove unneeded columns from the CSV files to optimize token usage during analysis.
3. Gather Demographic Data:
   • From ChatGPT, prompt it to give you a demographic breakdown of your product.
   • From Seller Central, navigate to Brand Analytics’ “Demographics” tab and download a quarterly report corresponding to your peak sales period.

Step 2: Analyze Reviews for USPs & Benefits

1. Initial Upload & Prompt:

   • In Claude, upload your positive reviews CSV.

   • Use the prompt:

     pgsql
     Copy
     You are a world-class audience analyst who helps companies analyze customer data. Analyze this data. Do not answer anything, just say “Done” when ready.
     
     

2. Extract Insights:

   • Once confirmed, provide a detailed prompt to extract at least 25 USPs and Benefits, including:
     • A bullet-point list of insights.
     • A four-column table (Type, Summary, Frequency, Customer Quotes).
   • Customize context to include product and target audience details.

3. Documentation:

   • Save Claude’s output in a Google Sheet to build a customer insights library.

Step 3: Build Customer Personas

1. Combine Data:
   • Upload positive reviews, negative reviews, and demographic data CSVs to Claude.
2. Initial Analysis:
   • Use a prompt asking Claude to process the combined data and respond “Done” when finished.
3. Generate Personas:
   • Provide a detailed prompt to create three distinct customer personas with the following details:
     • Name, Age, Gender, Location, Education Level, Income Bracket
     • Professional Goals/Challenges, Psychographics, Values, Attitudes toward the product
     • Media Consumption, Goals/Motivations, and Key purchase drivers.
4. Documentation:
   • Save the generated customer personas in your centralized repository for future use.