When designing a software system it is beneficial to study and use architectural styles from literature, to ensure certain quality attributes. However, as the interpretation of literature may differ depending on the background and area of expertise of the person reading the literature, we suggest that structured discussions about different architecture candidates provides more valuable insight not only in the architectures themselves, but in peoples’ opinions of the architectures’ benefits and liabilities. In this paper, we propose a method to elicit the views of individuals concerning architecture candidates for a software system and pinpoint where discussions are needed to come to a consensus view of the architectures.