A Day in the Life of an Analytics Translator

Is “analytics translator” a title or a skill set? Yes.
Stephen Gossett
January 5, 2021
Updated: January 7, 2021
Stephen Gossett
January 5, 2021
Updated: January 7, 2021

Think back to early February of 2018. Drake’s “God’s Plan” is topping the pop charts, the Philadelphia Eagles are about to end their 57-year championship drought, and a group of McKinsey leads and partners have just declared the analytics translator the “new must-have role.”

It wasn’t the first time the consultancy had advocated for the position. As far back as 2015, McKinsey staffers had been urging companies to consider hiring dedicated liaisons to bridge the gap between “business users” and “hardcore data crunchers.”

But this full-throated promotion — which also noted that that the firm had established its own internal training academy for analytics translators — struck a nerve. A crush of blog posts, videos and LinkedIn groups followed within months of publication, most of which referenced the McKinsey article.

Nearly three years later, “analytics translation” hasn’t exactly surged to ubiquity, but it’s proven resilient in some enterprise circles. A Google Trends chart tracking interest in the term paints a picture of some extended flatlines but also numerous peaks, well into the present era. And we’ve spotted open analytics- or data-translator roles in 2020 for companies as diverse as Kimberly Clark, Pilot Corporation and Hartford Financial Services.

It’s proof not only of some outfits’ long-term belief in the concept, but also that the gig does manifest a formal title, not simply a skillset — even if plenty of personal employee profiles use the term as a way to jazz up and differentiate their pages, à la the “product management ninja” or “sales enablement rockstar.”

Related13 Tips for Quick, Accurate Data Wrangling

 

analytics translator h1
Image: Shutterstock

Yes, Analytics Translators Are a Real Thing

Despite the fairly explanatory title, it’s easy to wonder how a dedicated analytics translator spends their day in 2021. Alexis Gruschow got used to puzzled looks when she introduced herself as an analytics translator, a role she held at industrial technologies company Fortive from early 2019 through September of 2020. The curious faces didn’t surprise her; to this day, she’s never met another person who’s shared the title.

“Because of the nature of it being a newer title, you always get questioned, ‘What does it even mean?’” she said.

Gruschow said McKinsey’s outline from 2018 tracks pretty closely to how the job manifested in the real world. Here’s how the consulting firm explained it in 2018: “[T]ranslators play a critical role in bridging the technical expertise of data engineers and data scientists with the operational expertise of marketing, supply chain, manufacturing, risk, and other frontline managers.”

In short, the translator looks for business problems or processes that can be streamlined, brings those issues to the data team, drives buy-in from stakeholders, then works with data scientists and engineers to hammer out a solution.

That first step is key. It’s not just about digging for potential use cases, but knowing what to prioritize.

“I helped structure them so that we really have an idea of the [biggest] pain points ... and then prioritize those use cases,” she said. “What makes sense for the business to focus on now versus later down the line?”

You always get questioned, ‘What does it even mean?’”

A fair amount of value engineering — balancing cost against the function of a product or process to drive value — takes place here too.

Next, the broader data team is looped in. (Gruschow was embedded on the data science side, rather than on the business side, though it can vary by organizational preference.) Gruschow wasn’t writing code or building models herself, but she was directly involved in choosing appropriate data sets, helping gauge whether certain data sources needed to be added or scrapped, and helping make sense of the outputs.

“As they’re getting different insights in the data and different outputs from the model, that’s where I’m usually embedded with them, trying to interpret what we’re seeing,” she said. “Why might this be happening a certain way? Then connecting back with subject matter experts on the business side and trying to understand each piece.”

alexis gruschow analytics translatorGruschow (left) declined to dive into specific projects, but she pointed to an industry-standard example: analyzing the confusion matrix of a model that predicts when machinery might require maintenance. Such tables illustrate the number of true positives, false positives, true negatives and false negatives in a machine learning algorithm.

“What does a false positive mean in terms of how we might take action?” she said. “What’s the cost and impact of a false positive? You’re putting that into business context to decide whether the model is actually moving in the right direction, to be productionized. Or do we go any other route?”

Operationalizing the end solution is part technical — Gruschow would help monitor model drift over time — and a lot of advocacy, or “selling the art of the possible.”

This is where the translation aspect really takes hold, synthesizing complex end results into something stakeholders can understand and take action on.

“You’re constantly trying to break down technical concepts and figure out how to get everyone on the same path, so they’re talking the same language,” Gruschow said. “It’s critical, especially when you’re trying to adopt the insights.”

RelatedHow Do You Grow Your Data Science Business?

 

What Does It Take?

The 2018 McKinsey article lays out four broad skills that an analytics translator needs:

  • Domain knowledge
  • General technical fluency
  • Project management skills
  • An entrepreneurial spirit

Again, Gruschow’s experience tracks closely with the outline. Her industry expertise and thorough understanding of profit and loss dynamics primed her for the role more than any coding bootcamp could, she said. In fact, Gruschow doesn’t know Python, R or any other programming language — but that’s not a dealbreaker in terms of finding success in the translator role.

(The job listing for Gruschow’s old role at Fortive slots experience as a data analyst, scientist or engineer “with hands-on coding experience” as a preferred, but not required, qualification.)

“I’m not usually the one jumping into the technical weeds,” she said. The fact that translators are working so closely alongside data scientists and engineers means they aren’t really required to take weedsy, technical deep dives.

“I know different algorithms that can be used and understand a lot of that different flavor, but I’m not so far down the technical side,” she said. “Usually you either have people come up from the technical side, and then they start to develop more business understanding, or now you also see a lot of people [start] with the business understanding.”

McKinsey suggested in 2018 that analytics translators need a “general technical fluency,” rather than strict programming or modeling expertise, while at the same time adding that they often come from STEM backgrounds. Gruschow credits her mechanical engineering background for instilling an analytics bedrock.

“You just really are bringing in a lot of the business-side knowledge along with an understanding of analytics and the impact it can have.

“My [employment] history doesn’t really make too much sense when you look at it on paper. I’m all over the place.... But I will forever have that [analytical] piece with me,” she said. The common thread throughout is a desire to solve difficult problems that are best tackled with data.

Another through line? Project management. While some data and analytics experts have stressed that you shouldn’t focus too much on project management in an analytics translation context, it’s still important. It pops up in the original McKinsey outline, in Fortive’s job ad and in my conversations with Gruschow. (McKinsey did not return requests for comment for this article.) Soft skills like emotional intelligence and a knack for storytelling are indispensable, too, she added.

But perhaps most of all, it’s Gruschow’s belief in the role that shines through most. Dismiss it as a buzzy rebranding of business analysis if you like, but the framework proved advantageous in her experience. And even though dedicated roles for analytics translators tend to pop up mostly in enterprise companies, Gruschow believes they make sense for some smaller companies as well.

Ironically enough, even though Gruschow recently held the title of analytics translator and believes in the role, she too considers it more a set of skills.

“You just really are bringing in a lot of the business-side knowledge along with an understanding of analytics and the impact it can have, bridging those two different pieces,” she said. “So I think a lot of people might have played in that space before.”

“I’ve kind of played this role regardless of when I’ve held the title,” she added.

RelatedWhat Is All This Data For?

Great Companies Need Great People. That's Where We Come In.

Recruit With Us