Moderated tests involve a test setup in which a participant and facilitator meet--either in a physical location or remotely. The facilitator guides the participant through a series of scenarios that approximate key tasks one would perform with the product. In most cases, the participant is not free to use the product spontaneously, although this is possible.

Typically the sessions are being recorded; the facilitator take notes and engages the participant as needed. When the participant needs help because they are stuck, the facilitator suggests ways to try to resolve the issue. Typically, the facilitator does not provide the answer right away, but rather, provides hints that might point to a desirable way out. Only when the user completely gives up on completing the task, will the facilitator provide the path.

In unmoderated tests, the participant explores a prototype of the product on their own--as the name of the test suggests. The session might be recorded, but often, the participant is asked to log their experience as they go. There might be a script, or the person might just perform the tasks that are relevant to them. The benefit of such an approach is that the participant does not feel stressed about being observed. Moreover, many participants worry- despite being explicitly assured to the contrary--that their performance is being tested and that failing to perform tasks correctly might impact their job. Unmoderated testing alleviates such potential pressure, and in addition, the experience is more closely tied to the natural use pattern the user prefers, as opposed to a scripted sequence that the participant might not be as familiar with.