Anything that cannot be verified within the confines of your development team is a potential playtesting candidate. A well prepared playtest will be confined to a particular set of questions that your team wants to answer, for example:
- Are the controls easy to understand and use?
- Do players like and care about their character?
- Is this area/mechanic/enemy/puzzle too difficult or too easy?
- Are people motivated to keep on playing?
Such questions and hypothesis are often a part of the production milestone and determine whether the game should continue as is, change its course, or even get canceled.
Testing the whole game without focus is rarely a good idea. The broader the spectrum of feedback, the harder it becomes to capture ...