A case in data job classifications

I work for a state government cabinet agency. We have a statewide set of job classifications that our own agency positions must fit into. Until 2019, my state lacked a data analyst class that all agencies could use. Prior to that time, we had to utilize existing job classifications and just make-do with what we had. A couple years ago, the state finally created a new analyst job class, but made several mistakes. This is somewhat forgivable, as data analytics is a young field and is constantly molding into its final form. It’s still a cautionary tale of opportunity, though. Let’s start with the most important opportunity areas and learn from this state’s mistakes:

First, the job class consists of a single position: a ‘data analytics specialist’ role at a salary starting below $60,000. The pay structures were extremely out of scale with the general economy in 2019, as they are today. The state offers a step system, where employees get (for the most part) guaranteed salary increases annually through year seven or eight. But do fresh 21-year old data science graduates care about making $80,000 in seven years, which is still drastically less than the rest of the industry. And if a fresh graduate with zero experience wouldn’t jump into that situation, what makes you think a mid-career professional with years of experience in data science would want to do the same? Profoundly unappealing salary: check.

Second, the fact that this single position contains both data analyst and data engineer responsibilities is depressing. A couple major, critical issues here: It’s laughable to expect to hire a competent data engineer for under $60,000. The combination job description also gives the job seeker the impression that the state doesn’t understand analytics. The practical result is that agencies had to rework legacy Database Administrator (DBA) job classes to get a broken system working. Perception of cluelessness: check.

Finally, the title itself: ‘data analytics specialist’ is just bizarre. No one in the field calls themselves that. I can only think that maybe, as the position is a union one, in making the position seem small and (frankly) pathetic, that they can hide some management duties in a position that they cannot advertise as management. No one in their right mind, who has gone to school for 4-6 years for data science, who has racked up hundreds of thousands in school debt, is going to jump at being a ‘data analytics specialist’ for $59,000 and be expected to do ETLs, mastermind and deploy a modern agency BI platform, build visuals, advise state agency directors, and do everything else under the sun. Not reflective of reality: check.

If you are a large organization, you should avoid this situation and do the following instead:

1. Create separate data analyst and data engineer positions. These are two entirely different roles. Unless you’re a tiny company and are expecting a single employee to hunch over a spreadsheet all day (and you actually want to deploy a modern platform) don’t think that you’re going to have a single all-in-one employee. Get off of that whole “I just need to pay someone a little more to be a ‘data scientist’ and then I don’t need an analyst and an engineer” fairy tale. That’s magical thinking.
2. If you’re a larger organization, you’re going to feel pressure to just have IT do the job. Resist that pressure. If want to properly use your data to drive efficiency and increase revenue through data-driven insight, you’re going to limit your own success by tucking it into IT. In nearly any IT department, a data engineer will just get used as a spare DBA and the data analyst will become a spare business process analyst every time another major project is given to IT.
3. Pay the appropriate industry salary for the right talent. Again, avoid magical thinking: expecting that you can hire a guy (or gal) who slipped Tableau onto his resume, call him a “data scientist,” and expect data scientist and data engineering output is not realistic. That is, unless you want to a) pay for the subsequent education he’ll need, and b) remain extremely patient.

In the spirit of this blog remaining weird, I’m going to put some words into Chihiro’s* mouth: “Don’t set your department up for failure. Do it right. Structure yourself for success.”

* It’s a character (known as the Ultimate Programmer) from a niche video game series called Danganronpa. It’s like a mix between Squid Game and Battle Royale/ Hunger Games**

** Hunger Games is a complete ripoff of Battle Royale. Don’t @ me.