Data-driven marketing analytics courses are integral to modern business management degrees in universities, yet many graduates focus solely on single, separated data analysis techniques during their learning process, hindering effective integration and practical performance. This study proposes that employing the Fishbowl method, which divides students into "fish" or "observers" to facilitate active problem-solving and analytical reflection, can effectively empower students to augment their learning and performance in marketing analysis by strengthening their metacognition. This research also explores the moderating effects of task complexity and students' divergent thinking. Two field experiments (41 Cohort 22/23 students in Study 1; 39 Cohort 23/24 students in Study 2) were implemented. The results revealed that the Fishbowl method significantly enhances students' metacognition, which affects their task-solving performance. Furthermore, students with higher (lower) divergent thinking perform better and are better suited to the observer (fish) roles. This moderating effect was strengthened when the task complexity was high. This study bridges the use of the Fishbowl method with the enhancement of metacognition in the context of marketing analytics courses. Appropriate utilization of the Fishbowl method during marketing analytics courses, along with grouping students based on their thinking traits, can significantly enhance learning effectiveness and performance.