Stop Hiring A 'Data Person

Your job post for a ‘data person’ describes three jobs. Whoever takes it will burn out doing all three.

I read a lot of these. Build and maintain pipelines. Own the data model. Create dashboards for every team. Bonus: machine learning experience.

That’s a data engineer, an analytics engineer and an analyst, with some data science on top. At a 40-person company one person can cover parts of all of them for a while. What they can’t do is all of it, on call, with every department as a stakeholder.

It usually goes like this. Month one, they fix the most visible dashboards. Month four, pipelines break and the dashboards wait. Month nine, they’re doing both badly and interviewing elsewhere.

Before you post the role, pick whatever hurts most right now. Numbers that disagree point to an analytics engineer. If data isn’t showing up, you need a data engineer. And if the company still struggles to ask useful questions, hire an analyst.

Then write down what this person won’t do. That list is what they’ll point to when the third department asks for just one more dashboard.

And if you can’t pick one because it’s all on fire, you probably need someone senior setting direction before the first hire.

What’s the longest list of responsibilities you’ve seen in one data job post?