TDWI | Training & Research | Business Intelligence, Analytics, Big Data, Data Warehousing

Does Your Management See Data Science as an Experiment?

Does management appreciate your data science team? A group of smart people isn't enough. You have to showcase your successes.

• By Jill Dyché
• March 2, 2017

I have an abiding love of all things Monty Python. Lately I've been laughing at inappropriate times remembering some scenes in the fantastic irreverent comedy, Life of Brian. (In this political climate it bears a repeat viewing.) In one scene a group of Judeans sits at a table complaining about the Romans.

Answering the question, "What have [the Romans] ever given us?" a group answer their leader Reg (played by John Cleese):

• "The aqueduct?"
• "Sanitation?"
• "And the roads!"
• "Irrigation."
• "Medicine."
• "Education."
• "And the wine!"

This scene came to mind recently when I read a letter from Craig, who leads a team of data scientists.

Hi Jill,

I suspect you've seen this problem before, but it feels unique to me. I've been leading a team of data scientists at a major retailer, and I can safely claim that we have been a success.

We've reached a petabyte of data from both inside and outside the company, and it's enriched our understanding of our customers' behaviors as well as the efficiencies of our business processes. Our insights into shelf space management have resulted in near-real-time product placement and pricing decisions. I can even quantify how much my team has saved the company, and our contribution to revenues. (We are data people, after all! )

My boss recently made an offhand comment that has me bothered. He said he didn't know "how much longer the company can afford our science experiment." At first I thought he was talking about a particular project -- a machine learning prototype in our supply chain division.

Then it dawned on me that he was talking about my team! I haven't told anyone this, but now I feel completely unrecognized and underappreciated. If he doesn't see our value by now, when will he? Do you have any advice for getting him to understand our benefit?

-- Craig, New York

Sorry, Craig, your boss sounds like Reg, taking the Romans for granted: "Apart from the sanitation, the medicine, education, wine, public order, irrigation, roads, the fresh water system, and public health, what have the Romans ever done for us?"

Impressively -- and, perhaps, presciently -- you seem to have been keeping track of your team's contributions along the way. My question back to you is: Have you shared this information? A team of smart people isn't enough. You have to showcase your successes. You've earned the right, Craig.

I suggest doing two things, pronto:

1. Document Your Measures

Your email says you can quantify cost savings and revenue contributions. Have you written those down and shared them? No executive is going to ignore hard ROI from an advanced analytics group. If you're like most managers, you're assuming that more people are paying attention to you than actually are.

If you haven't yet determined actual metrics for your team, you have an opportunity to define them. Include tried and true key performance indicators and cost-savings numbers that are so important to margin-sensitive retailers.

Then throw in some nontraditional metrics that can give a team of data scientists a halo effect. For instance, innovating solutions that allow the company to sunset legacy systems and enliven insular business processes is a skill often unique to data science teams and should be celebrated.

2: Share Your Results with Executives

Your boss might be clinging to established orthodoxies that have less relevance in the digital age, so relying on him to proselytize your team's accomplishments could be fraught with problems. Schedule a "Results Report-Out" meeting with your boss and his superiors and show some of the hard numbers you refer to earlier. Invite some of your team members to share stories of successful efforts. Use the meeting to discuss everyone's expectations about the future of your team and its work.

The risk is that this meeting becomes an argument about your numbers. Let it. (You guys are data scientists. You'll win.)

In defending your value proposition, remember to avoid jargon that can alienate executives who might otherwise be supportive. Be specific so no one conflates your team of data scientists with other analytics or innovation efforts. If you can, highlight specialized skills.

Now get out there and lead them, Craig!

Now it's your turn! Email Jill a question about analytics programs, data management, organizational issues, or culture at [email protected].