6. No understanding of business problems
You really can't hire a person who reiterates all of the business in math or statistical terms to your data scientist, who then "solves" problems. Why? Because if person A knows how to do all that, he or she probably knows enough to describe that in an algorithm to a computer. Why do you need person B?
7. No familiarity with the tools of the trade
There is SAS. There is R. There is Scala. There is Python. There is Matlab and a bevy of other tools. If you don't see those on the resume, then that person probably isn't a data scientist.
8. The SAS-only syndrome
With all due respect for my friends in the Containment Area for Relocated Yankees(Cary, N.C.), it seems like all two-bit SAS developers have rebranded themselves "data scientists." But that doesn't mean they know anything about data science (aka, knowing how to read the data) except how to write SAS code.
What sort of person do you need? You need an individual with the specific skills to address the problem and augment the existing technology team: a mathematician with programming and analytics experience, business sense, and the ability to talk to CEOs and techs alike. Now, don't go chasing unicorns -- but don't settle for chumps, either.
Good luck; you're going to need it. How competitive is it to snag someone good? Try this experiment: Add "data scientist" to your LinkedIn profile and watch a million recruiters shower you with offers of riches.
Sign up for CIO Asia eNewsletters.