In our digital age, we are inundated with vast amounts of user-generated content. From customer support tickets and survey responses, to social media posts and comments, anyone looking to derive insights from large volumes of content needs to spend a lot of time doing analysis and trying to make sense of it.

When I launched a new feature on YouTube comments, I wanted to understand how users were interacting with the product and what kind of comments they were posting. So, I embarked on a journey to manually and painstakingly analyze 1,000 comments. I read through each comment a dozen times, first creating a list of the prevalent themes, and then categorizing each comment into its respective theme. After a grueling 20 hours of analysis, I was able to make a simple pie chart visualizing the data.

Luckily, we're entering into an era where AI tools have revolutionized content analysis, allowing us to reclaim hours of manual and tedious work. While there's no shortage of specialized AI tools for product managers, customer support specialists, marketers and more, you might be surprised at how much you can do with just a ChatGPT Pro subscription and the knowledge of how to use it. Here’s my step-by-step guide to analyze content with AI in under 30 minutes:

Prep your data

Put your data into a spreadsheet and spend a few minutes making sure it’s clean and ready for analysis. Remember: garbage in, garbage out. For example, you might want to use only English content (which you can do with the Detect Language formula in Google Sheets). Delete any rows that look like spam, have a bunch of weird characters, and any content you otherwise want excluded. Make sure your columns are clearly labeled as well, so that the AI can interpret their semantic meaning.

Activate Advanced Data Analysis

Head on over to ChatGPT. If you haven’t upgraded to a pro subscription, now’s the time to fork over the $20 to get access to the feature you’ll need for this called Advanced Data Analysis. It’s part of GPT-4 and allows you to upload various types of content (such as documents, spreadsheets, PDFs, etc) and interact with it. Click the model toggle and select the option. If you don’t see that option, head into your Settings, then Beta features, and toggle it on.

Upload your data

Now we’re going to upload our spreadsheet into ChatGPT by clicking on the new “+” icon that’s appeared in our text input box.

Write your prompt

Using a good prompt template, write your instructions for ChatGPT. As an example, here’s the prompt I would use for my YouTube comments task:

You might decide to actually break this into three distinct steps, which will give you more opportunity to course-correct should you need it. You can try it both ways and see which generates better results for you.

Enjoy your 19.5 hours back!

You will need to iterate on the themes it comes up with, ask it for more examples to verify its work, ask it for different insights, etc. You can also ask it to explain its thinking and double-check its own work, which it's surprisingly good at. Remember, it’s a collaborative process and you’re instructing the AI and gently guiding it towards successfully completing your task, and sometimes it needs a little hand-holding. But this incredible tool will inevitably save you hours and let you get back to creative and strategic tasks that are much more fun!

Have you tried the Advanced Data Analysis function yet? Let me know in the comments!