Recently I awoke at 2 am to a very odd ‘whoosh’ sound in my bedroom. Whoosh, silence. Whoosh, silence. I tried to ignore it but the whoosh continued. Struggling awake, I realized that there was a bat in the bedroom!
I did the natural thing for 2013: Grabbed an iPhone, huddled under my down comforter, and Googled ‘how to get a bat out of the house.’ (Thankfully, my fiancé actually did the honors of escorting said bat out the window with the help of a broom, bright lights, and shouting.)
We don’t have any more bats in our house, but I remembered bats recently when a client called me.
“Help! I’m interviewing someone for a quant marketing consulting role. What should I ask to see if this person is any good?”
I spend many hours every week as the first line of defense for vetting analytical marketing talent. I’ve developed a framework I call BATS to evaluate candidates.
B: Business insight
A: Analytical heft
T: Technology fluency
S: Storytelling skills
These are often the main things I’m on the lookout for when I interview analytical marketers.
Vetting Talent Using BATS
Essentially, you want to see if a candidate can connect the ‘what’ to the ‘so what’ and the ‘now what.’ Do they have a good head for business? Can they link their analysis to implications for strategy?
Suggested interview questions for you to pose:
1. Tell me about a business project that required you to build a complex model. What was the challenge and what was your approach? (Listen to see if the candidate volunteers the forest-view as well as the tree-view.
2. To extend your first answer, let’s change parameters. If you had been working in the pharma industry rather than the technology industry, how would the implications be different? If you were analyzing offline channels rather than online channels, how would your recommendations change? (Listen to see if the candidate demonstrates breadth of business knowledge.)
3. Tell me about an insight that you pulled from data and why it mattered to the business.
4. Give me an example of a time when data didn’t jive with business logic.
Here you’re vetting the analytical/quant chops to derive insights from lots of data. Does the candidate demonstrate numeracy to the level where they understand the fundamentals of what their tools are reporting? Of course, for one role you may need someone with a deep statistical foundation and for another one a general fluency with numbers would suffice. Some starting questions are:
1. In this role you’d be working with big data sets. Can you tell me about the biggest data set you worked with and what approaches you used to tame it?
2. What did you do when your analytics tools fell short, and you needed to resort to something different?
3. Describe to me a mistake you made when architecting a predictive model and how you found it and fixed it.
Learning about the tools someone uses can help identify those who get their hands dirty with data versus those who are one step removed, managing vendors or analytic employees from a perch. On the other hand, you may observe that the candidate goes on and on about their skill with, say, Omniture, to the exclusion of discussing the business/analytics/storytelling legs of the stool. Depending on what you’re looking for, that may or may not be a dealbreaker for you.
Some suggested questions here:
1. On a scale of 1 to 10, where 1 is adding numbers and 10 is using all the advanced features, where do you put your abilities with Excel? (Note: You can substitute in SPSS or other tools)
2. Tell me about when you used a new tool for data analysis or data visualization. How did you go about learning this tool? (Note: It’s OK if someone doesn’t have immediate knowledge of all the tools you use in your business, especially if they have a track record of coming up to speed quickly with new programs.)
3. Let me tell you about a problem we are facing right now and the data set that we are working with. Which software packages would you use to work with this data set and why? What would you expect those tools to do for this problem and where would they fall short?
Here, you’re vetting the ability to craft a compelling narrative from a data-intensive project. And you’re looking for the skill to cater the ‘show and tell’ to different audiences.
You can read into storytelling ability by simply listening to how the candidate communicates. Do they spend 10 minutes going into the statistical or technical minutiae of their work when the line of questioning doesn’t warrant it? Can they communicate the bottom line of what they are doing? Do they structure and simplify their communication with frameworks or metaphors?
1. Give me the elevator pitch you’d give to your client or key internal stakeholder about what you do. Now, how would you adapt this pitch to a new PhD statistician on your team? How would you adapt it when meeting a CMO at a conference? How about when meeting a poet with no background in what you do?
2. Show me the slides that you’ve used recently to present results of an analysis. (Sign an NDA first if necessary.)
Relative Strengths Within BATS
It won’t come as news to you that few people have a perfectly-balanced BATS stool. Most candidates are lopsided. Some have the business insight chops without the quant underpinnings. Some are great at setting up and solving a problem but get stuck when they need to communicate it to a wider audience.
To get a good sense of the BATS area(s) that your candidate prioritizes and his or her relative strengths, you can ask the following:
1. If you think of the 4-legged BATS stool, rate your ability on a scale of 1 to 10 for each of the four areas.
2. In what area of BATS have you made the most progress over the past year? How did you approach this learning curve?
3. What do you want to learn in the next year? (This is an open-ended question, but the answer to it can reveal where the candidate will be most coachable.)
Using this BATS framework will help to develop a heat map of your candidates’ skills and interests.
As you get to know a candidate you can discuss each leg of the stool in more detail, depending on what is critical for the role.