This paper explores the data explosion phenomenon and why column-oriented database systems are able to provide the boost that is needed to gain the performance edge. The paper contrasts row-oriented systems with column-oriented systems, and then elaborates on the benefits of the columnar approach. A review of how data access latency across the different levels of the memory hierarchy can impact application performance sheds some light on why the column-oriented approach can deliver these benefits, which suggests a number of drivers for increasing analytical application performance, particularly by means of exploiting parallelism. Last, the paper suggests some factors guiding the evaluation process when considering a columnar system.