Traditionally companies have used software like SAS or SPSS for their analytic workloads and even for data loading and data munging. While these tools have been dominant in the data science domain for a while, open source tools, are becoming more mature, have deeper statistical and machine learning libraries, and the talent pool is increasing significantly.
What has complicated the use of commercial tools is that the expense is often directly related to data volumes and compute. For every organization that means an exponential increase in cost.