The Data Science Leaders East Coast Network recently sat down with the SVP of Analytics and Data Science at Saatchi & Saatchi Wellness to discuss his experiences as a data science leader and to understand his opinions on how leaders of this nascent function.
Originally Published by the Data Science Leaders East Coast Network on 7/13/18
Was your background in science or business and have you had support acquiring skills for the other?
My educational background was on the science side: I graduated with a degree in Mathematics, with a minor in education. I’ve always felt quite lucky to be in a role that leverages the things I learned in school on the job every day – both with respect to my mathematics degree in addition to the secondary education side. On the math side – the applications are pretty obvious – but my education minor has come to be extremely useful on the business side of my day-to-day work. The art of teaching is the art of presenting complex and new ideas to people in an easy-to-understand way that minimizes cognitive load. In data science on the marketing side – this is precisely the business we’re in.
In terms of acquiring skills on the business side – I’ve been quite lucky to have worked in roles that enabled me to gain a significant degree of exposure to how business is run – both our own business in addition to that of our clients. This is one of the benefits of ‘growing up’ working in an analytics-focused marketing services organization: in order to be successful, you have to align your work back to the business itself – and doing this successfully requires wearing a lot of hats. You have to have to ability to speak at multiple levels of sophistication within multiple domains with multiple stakeholders in order to get buy-in: before, during, and after the work is done.
Why was your data science function introduced into the business and do you feel it is delivering the value that was anticipated? Has it generated new and unexpected value?
Our goal in introducing data science as a core capability within the analytics group at SSW is to elevate and redefine the role of data within the service offering of a leading creative health & wellness agency. Just a few years ago, analytics groups within creative shops were typically limited to standard digital and sometimes integrated reporting, with some ad-hoc projects thrown in the mix from time-to-time. The central focus was on tracking and reporting on performance – ultimately answering the question “What Happened”?
Fast-forward to 2018, while integrated reporting remains (and will continue to remain) one of our key service offerings, we are beginning to see that the new role our analytics team is playing is more sophisticated, predictive in nature, and significantly more strategic. Rather than only looking through the rear-view mirror at past performance, our goal is now to leverage data and algorithms to predict future performance (what will happen) and optimize (in the mathematical sense) that performance (what is the best that could happen).
The introduction and development of this core offering has generated a significant amount of new value for both our clients, and our agency. We have shown that data science has the power to be an organic growth engine for the agency vs. a supporting capability. In particular, we have been able to develop a suite of data science oriented product offerings that enable our group to drive new business, focused on data science. This has enabled us to develop new analytics-exclusive client relationships outside of our standard agency-of record relationships. Through successful value delivery & business transformation for those clients — we have seen success to data in converting these relationships into long-standing full service agency of record engagements.
As a leader, what do you find are some of the biggest challenges working in a relatively new function in organizations that are not traditionally data-driven?
Within a business that is not traditionally data-driven (advertising), working for clients that are not traditionally data-driven (healthcare) – I believe the biggest challenge is that data science tends to bring individuals and businesses outside of their comfort zone. For the uninitiated – familiarity the nuts and bolts of data science and machine learning are mysterious. Gaining alignment on data science initiatives can be a significant challenge when more traditional business stakeholders are taken far outside of their comfort zones. To this end – I have found that some of the most successful data science projects are those that start small. By starting small, data scientists have the ability to prove value quickly (and build comfort with business stake holders along the way through the conscious use of language, good storytelling, and data visualization). I believe that by focusing on small projects first – and actively working to empower business stakeholders along the way through the development of intuition for data science – data science professionals can ultimately transform traditionally non-data driven organizations.
What advice would you give to aspiring data science leaders that you have had to learn the hard way?
I would answer this two-fold:
#1) Focus on the technical fundamentals while in school and develop an insatiable appetite to continue to learn while on the job. In order to be successful, fundamental and technical skills are a critical component to any viable analytics professional. Analytics & data science require mastery of foundational skills across a wide variety of academic and professional disciplines including mathematics, statistics, economics, psychology, sociology, marketing strategy, information science, digital development, data visualization, and computer science. In order to be successful – you will need to continue to hone and develop skills in all of these areas throughout your career.
#2) Remember that the technical foundation is only a set of tools for the real task at hand: the key is in maintaining a crystal-clear focus on the client’s business. The folks that are most successful in marketing analytics are the ones who have superior data and algorithms chops at the core – but also realize that the true art is in applying that set of tools to client business in a clear, strategic, actionable way.
After all, nobody would be impressed with an artist just because that artist has the absolute best set of brushes that money can buy. The value is in the painting that comes from the artists mind, heart, and soul via those implements. In the end, the world will never know or care about the brushes used to paint the painting, but the painting will live on.
This is how successful data science professionals should view tools like machine learning, statistical models, and complex SQL. These are all only tools – a means to an end. The art of analytics is rooted in the insight, impact, and unseen opportunities generated from the well-crafted applications of tools.
As the data science landscape rapidly evolves and grows, what do you think is the next big leadership hurdle that data scientists should be planning for? What do you envisage data science to have the potential to achieve?
Unless they work for an organization that is driven by highly-customized models (i.e., the business is the algorithm): data scientists working in more traditional businesses should quickly be planning for much of their day-to-day model development work to be largely automated. Over the last few years, we have seen significant advances in the automated machine learning space that I believe will transform the role of today’s data scientist. I believe that model training & development at scale will be largely an automated effort. To that end – there are components of data science work that will not be easily automated: in particular data engineering, data strategy, and data science consulting. All of these require a significant amount of domain expertise in business & enterprise IT knowledge – and a healthy dose of soft skills. I believe that data science professionals that focus on growing their comfort and knowledge in these areas will continue to be successful in the years to come – and I fear that those that don’t may soon find themselves automated out of a job.