Hey @Manuel_Stocker

TL;DR

• Make the UI as simpler as possible for everyone (you and the users).


• Have a short and sharp statement of what they need to achieve


• If you’re nearby when they run it. Ask them questions when they appears to be stuck


• If you’re not, use a third-part to record their session (maze, useberry, …) so you can analyze them later

It all depends on : what kind of test do you run, how long it is, what’s the point you’re trying to make, how many people are gonna test it and would you be there when they test it

Desiging

• 
  

  What are you trying to achieve?: Is this a proof of concept? Are you A/B testing solutions? Exploring experience for a new flow? Try to focus on where your questions rest. On what’s relevant to achieve the test and what’s not.

  
  


• 
  

  Effective UI : for everything that is no part of the test I take screenshot of the existing UI. Above that, I then can design the new interaction/flow (allow me to prototype faster and to rely on the view they actually used)

  
  


• 
  

  Simple and Sharp Statement: Whatever you’re trying to make them do, it need to be clear and simple so they won’t misunderstood/forget. If the flow is too big/long maybe cut it in several pieces (It will be easier for you to manage the different options/views and will reduce prototyping/linking error)

  
  

Running the test

• Sat with a user and watch them test the prototype: establish a simple flow and give them a clear statement of what they need to achieve. If at some point they seems to be struggling (rage click on some not clickable UI, searching for too long, …) just ask them why they’re trying to achieve, why they think this is the way, … guide them according to their answers.


• Sent to user to try on their own: Bulletproof the hell of your prototype. It need to be glitch free. If they get stuck because you forget to make a link between two views, because an option lock the sequence, … they’ll get frustrated, probably quit and you won’t be able to get the information you were first looking for.