How to Optimize Social Media Ads with ChatGPT

optimize social media ads with chatgpt


ChatGPT has quickly become an essential tool for copywriting in marketing. Not so many people know about its other superpowers, though.

The chatbot can also analyze data, for example. This turned out to be super useful in the performance optimization of social media ads.

We recently tested it, and the results were astonishing!

Context and results

We experimented in the following context:

  • A large-scale project with a huge number of social ads. For such big volumes, standard optimization usually takes us days.
  • The compared ads targeted the same audience under the same conditions.
  • We tested it with Facebook ads, but we believe it would have the same impact on ads on other social media platforms too.

Now that you got the big picture, let’s get to the results:

  • CTR increased by 124%
  • ROAS increased by 58%
  • Conversions increased by 257%
  • Optimization took us hours instead of days

Let’s show you how you can easily optimize ads with ChatGPT. All you’ll need are 2 things:

  1. Data Analysis GPT
  2. An active campaign with at least 1 well-performing ad

What is Data Analysis GPT?

The Data Analysis GPT is part of the numerous custom GPTs that OpenAI released in November 2023. Currently, they are available in ChatGPT Plus.

The GPTs are customizable models programmed to perform specific tasks. You can either use the ones made by OpenAI or build your own.

The Data Analysis GPT in particular can analyze and visualize various types of data.

How to use ChatGPT to optimize ads?

Disclaimer: Don’t forget to add some common sense to anything ChatGPT says. Double-check and validate its answers. Also keep in mind that the more precise data you input, the more accurate results it will produce.

With this in mind, let’s break it down into 3 simple steps:

Step 1:

Compile a file with all ads from the campaign. Make sure to:

  • Include ad titles, copies, creatives, etc
  • Point out specifically which ads perform well and which don’t
  • Provide data like CTR, engagement rate, reach, impressions, etc

Step 2:

Upload the file into Data Analysis custom GPT. Ask it to do the following:

  • Analyze the ads (titles, copies, creatives, etc)
  • Tell you why certain ads work well, while others don’t
  • Provide directions on how to optimize the bad-performing ads

Pro tip: Include context about the business and target audience in your prompt.

Step 3:

Optimize bad-performing ads and/or create new ones. Make sure to:

  • Implement the insights from Data Analysis GPT’s
  • Monitor performance and compare results
  • Adjust again if needed


Try it out and let us know how it worked for you.

✉️ Subscribe to our newsletter to get tips like these straight to your inbox. ⬇️

Want to hear from some of our top digital marketers or directly from our CEO?