When testing usability, think of which hypothesis you want to validate and which scenarios you have to work with. Based on those, create tasks that will help you measure how well you perform or solve those. 
**Tip**, for writing up hypothesis you can think of current problems and or opportunities, for example: 

 - Ex. problem: *users don't read my articles*. Hypothesis: 1. *the articles are not discoverable* 2. *the headers of the articles are not interesting enough* 

What you would do for the usability test is to **figure out whether your hypothesis are right or wrong and whether the design** (or design changes) **may help your problem.** 

Don't ask users directly why they don't read your articles to start with. You create tasks that may be solved by reading the article. They user tho may solve the task however they please. After this you can ask about how well they feel they could solve the task, missing info, what could work better, how else they could think of solving the task, etc. Get them to try those if they want to.

If they didn't discover the thing you wanted them to test, you can follow up by asking them to do that very thing and follow up with questions: 
-why do you do it in that way?
-How well did it performed?
-Would you have tried it like this if I didn't ask you to? why?
etc.

This way you can learn more about the challenges your product may have, which expectations your users have and how to solve it. 

**There is lots of good reading here**
https://www.nngroup.com/articles/measuring-perceived-usability/

Mind that tests are different based on whether you are looking for qualitative or quantitative data. Also mind that tests are different too when performed 1:1 or online. Can you see facial expressions? those will tell you a lot. 

Give users enough time to think and don't rush them in replying to questions or lead them to tell you what you want to hear :)