The growing amount of data enables more complex business analytics. Data for business analytics is stored in databases or data warehouses. Analysts want to execute their queries under requirements in a suitable time horizon. An architectural decision has a significant influence on response times. Therefore, it is necessary not only to identify and weight analysis tasks, but also to decide on the storage architecture. The architectural design influences the query execution time up to factor 100. We present both architectures and their influence on the database workload. Classic row stores perform better on transactional analysis and column stores outperform the other in simple online analytical processing.