11
In a data study our team at Nectiv performed earlier this month, we reviewed 8,500+ prompts to analyze the search queries that ChatGPT uses. The goal was to better understand exactly how ChatGPT utilizes it's search functionality. In the study (linked in comments), we analyzed how frequently it searches, how it differs by industry, how long queries and more.
However, one of the most useful ways we analyzed the data was to understand exactly what ChatGPT is searching for. This gives us insights into what terms SEOs and marketers should be optimizing their content for as ChatGPT is using them when performing web searches.
In the study, we found that ChatGPT uses the following terms the most:
1. Reviews (702 instance)
It looks for reviews of certain products or platforms (best cheap perfumes 2025 reviews, best SEO software reviews).
2. 2025 (257 instances)
ChatGPT looks for fresh information about the topic using the current year as a guide (button down dress trends 2025, best men's cologne 2025).
3. Free (236 instances)
The model will often try to find users free solutions to what they're looking for (free school bus routing software, best free janitorial software). In general, ChatGPT seems inclined to provide users with recommendations that are free/lower cost by default.
4. Features (204 instances)
ChatGPT looks for the top features of a given product or solution that users are asking about to understand what makes a good recommendation (features to look for in a trail camera, SoFi credit card features and review 2025).
5. Comparison (156 instances)
Finally, ChatGPT will often look for articles and content that directly compare multiple products against each other (Chase credit card comparison, termite control products comparison).
So if you're looking to better optimize your content for ChatGPT specifically, these could be great terms to utilize in key areas like your titles, headings etc. ChatGPT performs searches around these terms and the content that's optimized for these is more likely to be included in the actual output.
Link to original research: https://nectivdigital.com/new-data-study-wha...hind-the-scenes/
Author: «Chris Long»
Other articles:
