Why aren’t models and insights generated by many Data Science projects an instant hit with companies looking for data driven growth? They miss the Business Translator, an important role that nobody is currently recruiting for. Read on to learn about my proposal on how to make Data Science projects stick at your company and build an enduring business career.
In a recent post on one of the most trafficked sites for data scientists, a corporate recruiter wrote:
“… the difference between a good Data Scientist and a GREAT Data Scientist is often not found in their technical ability or their amazing mathematical genius. Data Science exists to provide a service to business and business is run by people. If Data Scientists cannot comfortably communicate with their non-expert colleagues and bosses, then their effectiveness is greatly reduced. They need to communicate easily with people, to understand, to interpret, to translate. “ (Matt R., BigCloud.io).
Yes, the inability to communicate with non-experts does reduce one’s effectiveness, Data Scientist or not. But doesn’t the Data Scientist have and impossibly long list of skills already that they should master in order to do well in their field?
The Data Scientists I met as a TA for the Data Science Certificate Program at University of Washington are extraordinarily smart people, with a fantastic technical skill set and the persistence required to sift through the mountains of data in front of them. But they are technical people who think, act and communicate very differently than business people.
Let's be honest, it is technology that integrates into business and not the other way around. Data Scientists could be providing a very important service, yet with few exceptions (high tech companies who make their money from data driven models... Google, Amazon, FB, LINKEDIN and such), they are at the hands of business, who pays for their value-add projects.
In my last post I wrote about the 7 key things to help the company harness Data Science power. What we are hearing from students in our Data Science Program at UW is that all too often companies are throwing ONE data scientist at mountains of data and expecting that he'll do EVERYTHING, all the digging, preparation, model generation and insights uncovering, then telling a great story that business can understand and appreciate. And if he fails to do so, then he’s not the great Data Scientist the company thought they hired. Really? He’s being held to such a high standard already!
We hear that the One Data Scientist approach is most often not working. Rather it is about the company still being on a learning curve to understand what data science can do for them.
That is why I am proposing a new approach. I think it would make more sense for companies to experiment with a slightly different model to increase their chances to get most out of their developing data science capabilities.
How about creating a data science team with specialized roles:
The team can have multiple Data Sources Specialists and Modelers depending on the complexity of the project and approved budget, yet I want to focus on the one Business Translator role and its’ crucial contribution to the success of the Data Science team.
The Business Translator works with the business to understand their needs and objectives and helps formulate the questions for the Data Science team. He is the voice of business in the process of creating and optimizing models, and helps the Data Science team focus their efforts, often contributing his opinion to early feature selection and other stages in the project... because he understands business!
Then once the model is created and insights are identified, the Business Translator packages those into a story that the business can understand and appreciate. His strength is storytelling and he leverages visualization techniques to tell business a great, actionable story.
The most relevant model results and insights I’ve experienced working with Data Science tools look cryptic and some are downright ugly. Data Scientists are thinkers who put value on the process and are not all that concerned with the aspect of the outcome and the presentation. For them Data speaks!
But business stakeholders are very different. They want and expect a story that is told in their language: easy to understand, clear, beautiful, inspiring. And above all, actionable. They are more likely to love the Data Science projects if one from their ranks, the Business Translator, creates and delivers the story.
My passion for data science and my strong business background took me right into that Business Translator space, and I think the need is huge. Yet my search did not reveal any career opportunities that resemble the Business Translator role yet. What I am hear when talking with recruiters is: you can’t be part of Data Science Team unless you are doing the Data Science work.
I strongly believe that needs to change and I want to lead that change.
There are far too few data scientists to go around. Why not match them with great business people fascinated by Data Science projects, who are already the best at storytelling and can take the “light” back to their business leaders? The Business Translator is the missing link to making Data Science stick.
Marius Marcu is a strategic innovator and a data science enthusiast who developed his thinking and skill set while at Intel, Microsoft and Quantum Corporation, where he solved tough business problems, launching global products and services that excite and delight customers.He helps businesses build Data Science practices that can really mine the gold in their data mountains and deliver results, enriching business leaders, customers and shareholders alike. Marius enjoys being the TA for the Data Science Program at University of Washington where he delivers a session called “The Business Perspective on Data Science. You can email him at [email protected].