By Philip Anderson, Data Scientist

In the short- to medium-terms, no. There are numerous reasons for this, but the one that stands out to me is simply that once a company has started to use SAS, it is very difficult to move away from it. The reasons for this are again, numerous, but one that I haven’t seen receive very much attention on Quora is SAS’ “stickiness.” SAS is unlike its primary open-source competitors in that it uses a distinct, largely non-intuitive language. There are perfectly logical technical reasons for the uses of things like data steps, procedure steps, and the “macro” facility, but these types of constructs don’t exist in SAS’ more modern competitors. This may ostensibly seem like a disadvantage, but it actually helps with SAS’ ability to stick in companies, because SAS expertise is largely non-transferable. For example, if you spent 10,000 hours programming in SAS, all that this would indicate is that you are probably a highly-proficient SAS programmer, not a highly-proficient programmer in general. Sure, you will learn about databases and for loops along the way, but much of what you will learn is how these operate in relation to SAS, not how they operate in general.*

This creates an issue for companies if they choose to move away from SAS, because their most valuable analytics employees, that is, those who are both extremely productive using the current SAS infrastructure, and who have acquired essential intra-firm knowledge, are going to be challenging to retrain. This is especially true if these employees are leading teams, and are already at capacity delivering established revenue streams. In addition to this, it is much easier for companies to find individuals with 15+ years of SAS programming experience to lead such teams, based solely on legacy – SAS was the only game in town for decades – R and Python are relative newcomers.

So is SAS going to die? No – but its market share will not be likely to increase beyond where it is today. SAS has been increasing its revenue in recent years, but this is only because the entire analytics/data science space is growing rapidly. If you were to plot user growth over time, the growth rate of SAS is most likely positive, but if you compare the second derivative of that metric (the rate of that growth rate) to R and Python, it would likely lag far behind.

*This has been my experience with 5 years of SAS, 3 years of R, and 1 of Python