I was intrigued by Donald Farmer’s recent tweet about ETL: “Migrating technologies is just work. Migrating people and mindsets is the real challenge.” I asked him to elaborate.
Donald is principal program manager of SQL Server Analysis Services at Microsoft. He sees “numerous” examples of migrating users who reject perfectly good methods in favor of “the Informatica way of doing it,” or the Datastage, or whatever they’re used to. Many IT workers see their career paths as not “DBA” but “Oracle DBA” or other brand, causing much extra expense in licensing and support. He writes in email, “CTOs have to battle that very hard.”
He sees this within Microsoft’s marketing analytics program. There, most recruits come from SAS — and they’re used to a much different approach.
SAS users, especially advanced users, are used to deploying a battery of statistical analysis and building a few carefully crafted models to deliver theoretically compelling results. In SQL Server, our users tend to iterate quickly and efficiently through many models, validating and testing as they go. These approaches and mindsets are very different.
We can generally answer the business questions that SAS answers, just as effectively for practical use — but SAS users will prefer the SAS methodology, and will often insist on it.
What’s a CTO to do? “The most effective technique I heard of for changing that mindset,” he writes, “was from an excellent CTO who simply made the SAS team responsible for their own budget. Two years later … well, SPSS not SQL Server, but the moral is sound.”
Dan Murray says
I’m definitely of the evolutionary model…do quickly, adapt, improve. It’s a much less costly way to go and for 95% of the real world works just fine.