More than ever, companies are looking to big data to help make business decisions. We’re seeing an increase in demand for data scientists, market researchers, analytic innovators and chief data officers. These roles, however, are tough to fill. It’s a niche skill set that some candidates claim to have but don’t always have the experience to back it up. Here’s how to gauge whether you’ve interviewing a true data-analytics job candidate — or an impersonator.
Check references. References may sound basic, but they are crucial. Some companies use data-analytics firms to not only gather the data, but also to package it in a way that pulls out analysis and trends. If candidates say they got the data and worked through it to package it, verify that with those who’ve worked alongside them. It’s one thing to get raw data to sift through and analyze, and it’s another to receive it already packaged, with the work done with explanations and callouts. Depending on the company’s budget, working with a consulting firm may not always be an option, so the ability to work within raw data is important.
Actual examples. Regardless of their previous role, have them share an example of how they’ve analyzed data in the past. Ask for both the written and oral presentation. You want the person who actually did the heavy lifting, versus the person who only interpreted the information.
Take-home projects. Give your candidates a case study to take home and analyze. Don’t put too many guidelines around it — just the bare minimum and see what they come up with. Beyond grammar and punctuation, look to see how they digested the data and presented it. Did they put together a PowerPoint, leave it in Excel, or submit a word doc with bulleted thoughts? Give them only a couple of days to complete the task, which also helps identify if they can meet deadlines.
On-the-spot tests. The best way to tell in real time whether or not a candidate is good at analyzing data is to present them with a data set during the interview and have them share how they would go about drawing conclusions. Have them walk you through their thoughts and explain each step. The goal or intention isn’t for it to be perfect but to gauge their process. Are they familiar with the meaningful metrics in your operation? How will they interpret the findings? What trends can they pull from it, if any? Do they discuss at all the impact the findings could have on the team or company, and do they offer any suggestions for change?
Challenge the status quo. Talk to the candidate about a flawed process, or something you did that went wrong. Do they challenge or push back on why you went about it a certain way, or suggest a different way? If they are tenacious and aggressive in the interview, they will challenge the status quo on the job.
Storytelling. If when explaining a project they worked on, candidates claim to have reduced or increased key metrics, ask why they thought it was successful and what downstream impact it had on the business. Can they explain the story, not just the end result? How well do they understand the complexities of business? Great data analytics people can tear something apart, put it back together and explain why their process of reconstruction is “better.”
Insightfulness. Regardless of the project, whether it was an in-person analysis or report from a take-home assignment, have them walk through how they got to each step. What was their thought process, and are they able to expand on how it would impact business? Data analytics is analyzing raw data and deriving trends, having great communication skills to translate the findings into layman’s terms so all understand, suggesting and implementing changes, and reflecting to see how those changes impacted business and what could have been done differently. Great data analytics people look beyond the numbers and also understand that there are different ways to interpret and implement data. They’re not black and white thinkers.
Big data is only going to grow more consequential as our businesses evolve. Finding leaders who can keep the data closely meshed to operational and strategic concerns will decide which companies will be best prepared to meet that future.
This article originally appeared on WSJ.com.