o great with numbers? Analyze your data Or too little time to regularly dive into the available data? Then the ‘Data Analysis’ (formerly Advanced Data Analysis/Code Interpreter) options within ChatGPT are your best friend. In the paid ChatGPT-4 you now have the option to upload data files. This is perhaps the coolest addition since going live, as it allows you to analyze data and generate graphs with little knowledge.
How do you use Data Analysis?
As marketers, we regularly have to dive into large amounts of data to spot opportunities or discover trends in customer behavior. ChatGPT can now help you very well with that. If you have a raw data file of, for example, customer behavior, purchases or a customer satisfaction survey, ChatGPT can help you create visualizations of this. Think of printing tables, but also graphs for visualizing your data.
For example, using the prompt below, I segmented the data from a customer satisfaction survey by the variable ‘branch’ so that I could evaluate the sentiment per branch.
You are a data analyst. This is a customer satisfaction survey that needs to be analyzed. Clean the data if necessary. I want to see in a table based on location (column D) what the average score is, get this from column H.
With this, ChatGPT went to work to make the file readable. It removed the invalid fields and came up with the correct results below.
ChatGPT’s output of a table with average figures. Establishments are masked (by ChatGPT) for privacy.
Please note ! Always remove sensitive data first, new zealand email list 1.4 M contact leads such as customer data or other data that can reveal which company it belongs to. It remains a black box where the data goes exactly. Therefore, play it safe and stick to anonymous general data.
Interesting insights from your customer data
In the previous example I used a very specific prompt for what I wanted. This is not always the case in the case of a large customer base. Here you need help to get interesting insights from the data. If you can online marketing: a look back & trends for 2024 easily run exports of customer characteristics, the amount of purchases and average purchase values, you can easily find correlations between these characteristics. Ideal insights to continue testing on your various marketing channels.
I used a customer base of over 45,000 customers and values agb directory such as demographics and purchases, but also the first acquisition channel, interactions on different channels and email statistics (such as personal open rates). If you ask ChatGPT to extract interesting insights from your data, it will again clean the data by default and then ask you a number of questions about what kind of analysis you want to make (so it comes up with suggestions itself).
After this choice, the AI also asks which analysis method you want to use and then immediately prints out a graph. For example, I wanted to know which correlations could be found between the different data points, where ChatGPT was able to print out a correlation matrix between all these different data fields. For example, I saw that an initial interaction with TikTok, or the acquisition channel, ultimately results in a higher Customer Lifetime Value (correlation score of 0.57). Based on this insight, we