Along with social distancing rules during the pandemic, online shopping activities in Indonesia are increasing. With the increase in online business players, of course, the competition is also getting tougher. Creating an effective marketing strategy is something that must be done to survive in the competition. Promotional media is one of important aspects in marketing strategy that needs to be optimized. Customer characteristics need to be considered in creating attractive promotional media. Attractive promotional media are expected to increase sales and increase customer loyalty and satisfaction. This study tried to implement clustering analysis using K-means algorithm to determine customer characteristics and designing suitable marketing strategies based on customer characteristics using the marketing mix method. Before implementing K-means algorithm, Davies Bouldin Index (DBI) method was used to determine the optimum number of clusters. 236 customer data were clustered by considering 8 indicators. There results indicated 5 clusters with specific customer characteristics. Then, those results would be further analyzed to determine suitable promotion media based on the marketing mix method. The result of the study are expected can help business to achieve customer satisfaction.