Hiring For Big Data Analytics: The House of Analytics


If your sister were a heart surgeon you may want her to write you a prescription for your allergy pills or antibiotics for your child's ear infection, but you certainly will not ask her to see if your glaucoma is ready for surgery.  Most advanced fields have specializations that are fairly distinct in terms of skills, knowledge, and aptitude, that’s something we all take for granted.  Why is it then that job openings in the analytics sphere are so convoluted or have one-size-fits-all kind of air that reek of the hiring managers’ incompetence or ignorance?  Either they do not understand what they hope to achieve, or their organizations are just not ready for serious analytics.  Hospitals hiring general physicians would be absolutely unremarkable but if all the positions were allocated for general physicians then that would definitely raise eyebrows.  Yet, we see most job requirements in analytics looking like someone wants a heart surgeon, a gastroenterologists, and an ophthalmologist all wrapped in one person.  Such a knee-jerk manner of hiring analytics experts would at the least create inefficiency and on the other hand it may also turn out to be a massive opportunity cost with regard to potential for growth of the firm had the analytic investments been optimized.
This defeatist attitude towards hiring of analytics experts happens primarily due to lack of awareness about the field of analytics, its various segments, skills and aptitude of analysts. And that is what I would like to address with this piece. For the uninitiated, here is a typology that should be helpful in understanding the House of Analytics.

Data Hoarders: These are technologists whose lives revolve around storing data efficiently and reliably. They have little to do with analytics.

Data Gophers: They are programmers trained in data manipulation.  They can extract, transform, and load data for analysis and they respond to requirements. These people straddle technology and communicate with crunchers and modelers.  

Data Crunchers: Crunchers are reactive data analysts mostly.  They respond to requests and mainly create reports and solve structured business problems given to them by business decision makers.  .  Often times, proactive crunchers come up with insights of their own but for most part they are focused on  "What is going on" and the objectives given to them by decision-makers.

Data Modelers: Data modeling is the organization of data into tables, and accounts for access, performance and storage. Data modelers are not your typical data analysts. They work with business decision-makers and users of information in an organization to create data models that fits their requirements. They mainly make predictive models and come from the statistics/computational side of the world.  

Data Evangelists: Though short lived, they will be responsible for bridging the chasm between data and business decisions.  Data evangelists need great communication and sales skills as they have to understand the world of data in their organizations as well as manage the proclivities of business decision makers.  These people look for problems and define them for solutions.  They go beyond "what is going on" to "why is it going on".  They interface well with the data users.

Data Users: Today an investor does not need to download and analyze financial data on a spreadsheet.  All of that is readily available.  Nor does a statistician need programming training to run models or a photographer need sunlight measurements.  We are well on our way towards reducing the need for modelers & crunchers because technology will make all of that obsolete.  This means that the user will need to be better trained than they currently are about analytics.  When that happens, we would be able to say analytics has become pervasive in business. Today, however, that is not the case. Till then data users will need Data Evangelists to help them become proactive data users.

It is possible that a single person may be doing all of the above in a given organization but the roles are distinct and in some organizations each function will be an independent responsibility.  The jack-of-all-trades analytics expert model is highly inefficient and organizations will need to change to the centralized analytic team model if they are to compete on analytics. The other problem is one of aptitude.  You would not expect a professional chef to work as a computer programmer as well. Similarly, a modeler would not like to, or have the temperament to do the gopher’s job and an evangelist would feel like a fish out of water if she has to spend bulk of her time with statistical models.  Top Gophers have spent 20 years with SAS/SQL and love their life that way.  Hiring managers need to be cognizant of these differences if they are to build great analytic teams.