In contrast, user experience research is based on behaviour. Behaviour turns out to be remarkably consistent from person to person. For example, if you watch 5 people approach a door, and 4 of them attempt to pull it when the door needs to be pushed, you know there’s something wrong with the design. You don’t need to randomly sample 370 people to draw this conclusion. You observe that the door has a pull handle, and it’s probably that that’s causing the problem. So you replace the pull handle with a push panel, and see if you’ve fixed the problem.
Il est possible de prendre 4 ou 5 utilisateurs types, de la même catégorie s'il vous plaît, pour évaluer rapidement les grands obstacles d'une expérience car justement, on étudie (et non teste) le comportement et la réaction qui sont généralement communes à l'ensemble des humains.
The second assumption — that you can use the person at the next desk as your user — is equally flawed. With the exception of intranets, it’s very rare that internal staff are the target market for the product you’re designing. Real users are almost certainly less tech-savvy, much less knowledgeable about the product domain and a lot less tolerant of the product’s shortcomings than internal users will be.
Eh oui, vous n'êtes pas l'un de ces utilisateurs types. Vos collègues non plus, vos clients non plus, les vendeurs non plus …
… et tant qu'à y être, vos segments de marché ne représentent pas non plus les différentes catégories d'utilisateurs. Il ne faut pas mélanger part de marchés et personnas.