Card sorting reveals how users organize and label information, guiding the design of classification systems that align with their mental models. This paper focuses on the closed card sorting variation, in which users assign content items to a predefined set of categories. Specifically, this work investigates the minimum number of participants needed for reliable closed card sort results, using an e-commerce website as a case study. First, 191 participants sorted 45 cards into 7 predefined categories. Next, statistical analysis was conducted using a resampling procedure and two analysis methods that calculated the correlation between the data from all the participants and data from 100 random samples for each possible sample size, ranging from N=1 to N=191 with a step of 1. It was found that a sample size of 8-12 participants is sufficient to obtain reliable data from closed card sorts, with further increases in the number of participants proving to be more costly than insightful.