- Avoid asking leading questions and generating useless feedback
- Perfect shortcut for founders, marketers, growth hackers
- Can take less than 1 minute to generate a dozen good user research questions
You want to run a usability test on your site because, among other outcomes, a quick usability study can help validate your copy—with real humans. But you’re not sure what questions to ask in the test.
Wouldn’t it be nice for hire AI to generate user research questions for you?
Now you can!
Here’s what to do: Ask ChatGPT to give you the exact user testing questions you need. This is what it created for me when I used the AI copywriting prompts I’m about to teach you:
Here’s how to use ChatGPT to generate user research questions like those ^^ for your business or client.
Step 1: Give ChatGPT your instructions
For now we’ll just ask ChatGPT for generic questions based on what we know works.
Note: For this example I’m using ChatGPT 4 Plus (Browse with Bing Beta) because of its ability to access the web. But you can still use this first prompt with ChatGPT 3.5 or the standard GPT 4 model and it will work great.
In ChatGPT enter the following prompt:
Copy paste this prompt:
Write a series of website usability testing questions. The questions will need to help us understand whether or not our copy is hitting the mark and how we can improve it. In particular: - What website visitors are thinking when they land on the page - What their first impression of the page/brand is after looking at it for 10 seconds only - Who they think the site is for and what they think they can do there - What is the first thing they'd want to do on the page/site - What they expect to happen once they complete that action - What their reaction is after completing it - How they would go about finding out the information they need before purchasing - What questions if any they still have after browsing the site - How they would go about buying a specific product on the site and to walk us through the entire experience - What they liked/didn't like about their experience - What they would type in Google to search for a similar product/website - 3 words they would use to describe their experience Make sure the questions you write are not leading and unbiased.
ChatGPT should give you a good starting point for your questions.
Step 2: Make the user research questions it generates your own
At this point you can either take these questions and apply them to your site, or have ChatGPT do the work for you automatically. Here’s how:
In ChatGPT (Browse with Bing or Plugins using WebPilot), enter the following prompt:
Copy paste this prompt:
Now look at [website URL] and adapt the questions to this site. Then tell me what changes you made and why. When it comes to the buying journey question, make it specific to the site, coming up with a real product they should look for and buy (or get as far as possible in the purchase process before they have to enter their payment information).
Make sure you use your website’s URL in the prompt.
Within literally 1 minute from starting, ChatGPT had generated the following for me:
As you can see under Number 4, Purchasing Journey, it even found a specific product to look for on the site (that Smoky Chipotle Chicken meal sounds delicious 🤤).
And it even gave us rationale for how it adapted the usability questions to the website. So we know exactly why it did what. This gives us a good starting point for making additional changes ourselves.
Here’s the full chat if you’re curious.
I tried this prompt on the Copyhackers website. Here’s what happened.
Compared to an ecommerce like Prepped Pots, Copyhackers is all about sharing (and selling) information. This is why we need to customize our usability testing questions accordingly.
Here’s what ChatGPT produced simply replacing the Prepped Pots URL with Copyhackers’s:
The beauty of this series of prompts is that you could keep the same chat for any and all your usability testing projects. Just keep asking for a different URL to check out and use.
Time to start validating some of that copy.