I need your help with a curious project I am working on.
I’m going to mine 250,000 eBay feedback comments and look for some common themes among them and see what can we learn from what’s been left so that we can make our businesses better.
Mining the Data
We take positive feedback as granted and tend to focus on only the bad side, well we would they are the ones that stick in our immediate memories.
But… what really makes buyers leave positive feedback? and crucially what can we learn by looking at a large number of feedback comments that have been left over a variety of businesses? Are there common themes, is one business excelling where another is not?
I’ve already worked out that I need to split them across users that have 3rd party software tools and ones that don’t and a good mix across different eBay categories to obtain a wide set of sample data and also no older than 6 months.
The reason why I’m curious about this project is because I’m really interested in what are the common comments left and from the sample set of data, is there such a thing as a “Perfect eBay Feedback Comment” using the top 10 or so words and if it has left in the sample set?
I’m also really curious how as a nation how polite we are using words such as “thank you” and “thanks” and what makes a buyer use emotional words like “love” and “excited”. How important delivery time is and what percentage references to “fast delivery”, “next day” and to try and apply some numbers to the words used such as “comms”, “communications” and what relationship there is between feedback comments left that use these terms for sellers that use and do not use advanced tools like CA or ESP and then to spin them against the keyword “service”.
What impact does free shipping have on feedback comments that are left, does item value change the tone of the feedback comments left and what, if any difference eTRS has between the common keywords for praise for sellers that have eTRS and sellers that do not.
I’ve never seen this done publicly before and I’m going to build out a few theories beforehand and then test to see if they are true or not, but that’s where I need your help:
- Is there a theme or comparison that I have not thought of above yet?
- Is there a way of comparing the comment or the contents of the comment to another attribute like eTRS or similar that I have not thought of?
- What do you think I’ll find?
Let me know in the comments box below: