Hiring data science teams? --Here's some advice to help keep them.

Okay, so I've moved into a data scientist role for the last five years and after a few companies I've gotta write this advice piece about hiring/assembling together a data science team.

a, data science is a multi-skill race. Use any of the multiple visualizations online, or better yet consult them and create own picking what's relevant to you.(If only to share between the executive team).

b, If you can't figure out what's relevant to you, first hire someone who can figure it out.(like say a chief data-driven-decision officer ???).

c, Think about the incentives you're setting your data scientists and how they align with your organizational goals for the data science role. For ex: in the above case finding out who can teach you about the data science way, don't look for someone who can give great presentations about "data science", but someone who takes a bit of "skin in the game".

d, If you don't have a data warehouse/pipeline etc, please re-write your job description, (instead of a mish-mash copy-pasta that usually gets put together) and think about the core skills vs satellite/supporting skillls vs nice-to-have skills.

e, If you do have reasonably settled data warehouse,pipeline, and engineering teams, try to create a brain-storming within the team about what skillset is missing in the team, and then filter out which of those help the business. (Important to do this in the 2 different stages/steps and not combine together. Effectiveness of brain-storming depends on it.)

f, Think about (long-term) motivations of the data scientists themselves, in general it's a well-paid market now, so money may only be a temporary motivation. Figure out how the long-term goal/motivation and the incentives/offers/compensation you make get them to align with the business/organizations well-being.

0, Yes you got that right, before all of these steps there's a important thought. What or why or how do you think data science can help your business? Be super-honest with yourself about how sure are of your answer, and if you're not sure, try to find people who can bridge that business-to-data-science-applications gap first*.

* -- I won't try to go into detail about this, as I consider myself as someone trying to be that person rather than being that person already.