Real-time data stream in the format of WITSML which can have frequency as low as 1 Hz is one of the best candidate to produce KPIs for the drilling operation activity. The KPIs generated from this calculation will have a relationship with other information from other data sources, known as metadata. The question is how can this KPI information be utilized for further analysis, wider/more complex analysis process which needs to be combined with metadata? An OLTP model is not the recommended model for data analytics but OLAP is. Another question is how will this data be stored in terms of the physical storage? We argue to use column-oriented for the physical storage which can perform analytical queries 10x to 30x faster than the row-oriented storage. The implementation of an OLAP model for storing KPIs data is proven to improve the performance of the analytical query significantly and combined with the implementation of column-oriented in the OLAP model improves more performance. This concludes that the implementation of OLAP with column-oriented data model can be used as the solid foundation for storing KPI data.