(Photo by Tim Mossholder on Unsplash )

Are you (still!) having trouble hiring for data roles? Even with the recent tech layoffs, there's a competitive market for those data scientists, data engineers, and ML engineers.

You can't control the job market. But you can control how you approach it. Work through these three questions to improve your approach to data hiring:

Question 1: "What will this person work on?"

You want to be specific here. I mean, very specific. If you just hand-wave it away as "they'll do data stuff" or "I'll hire them and let them figure that out," then you're not being honest with yourself. You need to list actual projects they'll work on.

Are you coming up short on that list? Either you don't actually need to hire anyone (you can stop your search here and save money) or you need help developing concrete use cases (I can help).

Why this is important: You don't want to hire someone until you've confirmed that they'll actually have work to do. This is especially important for your company's first data hire. What happens if they don't have anything to work on when they arrive? Sure, they may find something of value to do. Or maybe they'll get bored and leave after a few months. Do you really want to take the chance?

Question 2: "What skills do we actually need, versus want?"

Yes, you may want your founding data scientist to have a PhD in statistics, be a prolific writer and public speaker, plus operate as a player-coach who can seamlessly shift between "develop strategy with the executives," "train neural networks," "implement cutting-edge research they picked up from a paper," "grow a team," and "mentor junior hires." Sure.

(I'm only slightly exaggerating a wish list someone once rattled off to me. Seriously.)

To be fair, that person certainly exists. But ...ask yourself: why does your company need that one person who embodies all of those attributes, right now?

Why this is important: This kind of wish list is a tall order. There's a slim chance you'll find such a candidate in a reasonable amount of time. Once you do, expect them to be in high demand, so you'll need to craft a very attractive package (title, role, comp) to convince them to leave their current company and join yours.

You could do that. Or you could review your list and see what this role actually requires. Pick up that job posting and sort every line item into a "want" or "need" pile. Be sure to note why. This can be a grueling exercise, but you may as well do it now. Candidates will ask you about it later.

Question 3: "Are we the problem?"

If you've interviewed a number of candidates and no one has been a fit ... it may be time to look inward.

Even though you've confirmed that they'll have work to do, and you've separated the "nice to haves" from "must-haves," there may be needless friction in the mix:

• Have you made your hiring process too much of a meat grinder? You want to evaluate candidates for skill and fit; you're not supposed to use the interview to show off how smart or tough you are.
• Does your company culture need an adjustment? If you and your team come off as unwelcoming, unpleasant, or unprofessional, expect candidates to react accordingly.
• Why not take another look at the comp? Yes, you need to pay an appropriate wage for the role. Don't forget about vacation, health benefits, and anything else that will round out the total compensation package.

Why this is important: Remember that hiring is a matter of sales. Candidates are interviewing you as much as you're interviewing them.

What next?

Don't be fooled. These questions are easy to ask, but require plenty of effort and discipline to properly answer. The payoff is that you'll improve your chances of landing that data scientist, data engineer, or ML engineer you've been waiting for.