Design teams often rely on precedents, but the impact of using the direct experience of a precedent on design behavior requires further investigation. To explore this impact, fifteen teams of master students of architecture participated in two design sessions: one without and the other with a previous experience of visiting an example. The think-aloud protocol was applied, and a multi-layer coding based on FBS and precedent types was proposed to measure episodic (a specific design experience) and semantic (general design principles) precedents. Furthermore, a measurement was proposed to examine how and when the teams analyzed the precedents or applied them to solve the problem. Dynamic analyses were used to study the differences between the two conditions based on the distribution of precedents and the Problem-Solution index. Statistical analysis unfolded significant differences between the two conditions. In the second design session, the number of precedents, solution generations, and analyzing precedents were significantly increased. The first design session's episodic precedent had categorical similarity with both semantic and episodic precedents of the second design session. The results also indicated that design teams preferred to analyze the visiting experience of the precedent at the outset of the design session and apply other precedents for solving the design problem until the end of the session. While in the second design session, solution generation significantly increased, more episodic precedents related to the direct experience were used for problem framing. From another perspective, these precedents were also applied more to generate design concepts rather than refining design ideations.