Keyword Classification: ChatGPT or Excel

Why keyword classification?
Let’s not focus on terminology – should it be categorization vs classification? The purpose of this post is to show some practical methods for reporting or analytics. This helps digital marketers categorize and organize large amounts of text data, such as search terms, customer reviews or social media posts to identify trends and patterns. Let’s see some examples.

1. ChatGPT – Easy but not highly accurate

Using ChatGPT for text classification is a new way. The method I’m using is to ask ChatGPT to classify a bunch of keywords (over+100) based on the existing patterns provided. If you can’t suggest any candidate categories, then let the AI suggest the best category possible based on the data it was trained on.

Prompt:
Please give me a table for the following keyword classification using my given ”Keyword / Category” pattern.
Keywords:
[List your keywords here – You can paste a list of keywords from a spreadsheet or manually type them in.]
”Keyword / Category” Pattern”:
[Provide some examples for ChatGPT to classify text – You can paste a list of examples from a spreadsheet.]

Pros:
– Super simple (no code needed)
– Possible to copy the table given to your own spreadsheet
– No issue with entering over 100 keywords
Cons:
– Not accurate sometimes
– Very long prompt if a large number of keywords are requested
– Longer processing time if a large number of keywords are requested

2. Excel – More steps but highly customized

Like method 1, I have to manually provide classification patterns (column E, F), then use the excel formula ”Index/Match” (column C) to do the trick. Since I’m already in the spreadsheet, I can make this spreadsheet more functional by adding multiple formulas to display the number of keywords (Column H, I) and all corresponding keywords (Column K, L, M) under each category.

Pros:
– Highly customized and flexible (You can add more patterns to improve classification accuracy)
– More accurate
– Quick processing time
– More functional tables generated for further analysis
Cons:
– May need to add more classification patterns when you have new keywords
– More work to set up the spreadsheet but you can use this link to get a free template.

Scroll to Top