by Ron Menich, VP, Advanced Analytics & Data Science
At Catalina, Ron Menich currently leads a team of 15 data scientists who personalize ads and coupons, recommend optimal audiences, statistically measure media effectiveness, do in-flight optimization and contextual targeting, and perform other marketing science activities. During his storied career, he has applied his natural mathematical abilities (in 2nd grade, his teacher sent him upstairs to see the 8th grade science teacher because he had written something on relativity theory!) to several businesses.
“Living at the Interface of Math and Business” has been a theme throughout my professional career. After having worked in perishable asset revenue management and supply chain demand forecasting before joining Catalina in 2018, I can reflect that marketing is an incredibly rich problem area for the application of machine learning, data science, and marketing science.
While earlier in my career I often worked on big mathematical programs, the portion of the marketing space which Catalina inhabits tends to require more predictive analytics models and less large-scale optimizations for the solution of the business problems at hand. (That being said, we know of course that large-scale global optimization methods underpin the various machine learning methods we employ.)
One aspect of working in the marketing realm is that the degree of technological flux is very high relative to the revenue management and supply chain forecasting environments in which I previously worked. Marketing is constantly changing in response to new means of digital media delivery such as out-of-home screens and connected TV; in response to rapidly escalating user privacy concerns; and in response to market restructurings between CPGs, retailers, and agencies, such as the retail media network genesis currently underway. This flux leads to the need for a data science team to handle a wider set of problems than is typical in other corporate environments.
Living at the interface of math and business is a challenging and rewarding place to be. To me, it’s like solving a puzzle. There’s a business problem to be solved. There are deadlines. There are algorithmic possibilities and computational costs. These different puzzle pieces have to come together to form a comprehensive solution.
For a majority of my career, I have been involved in business areas in which users have manually managed some business process for many years, and automation and analytical recommendations are slowly replacing those manual methods. When automating previously manual processes, the long-term benefits of doing so might be clear, but in the near-term, change is difficult and often unwelcome to those who have built a career around those manual methods. Always remember to stress that the focus of these systems is not to reduce headcount, but to liberate users from drudgery, enabling them to focus on higher-value and more-rewarding activities.
Living at the interface of math and business also brings one face to face with many people in large corporations who are decision-makers or otherwise impact the success of data science applications, but who themselves have no analytics background, or who might have hated mathematics when they were in high school, and perhaps even resent you for your expertise.
It’s important to remember that folks who are not trained in analytical methods often have huge reservoirs of business knowledge and gut feel by which they can instantly examine the recommendation coming out of a big system and see flaws or concerns which you might have missed. So I have a great deal of respect for users, and the perspective they bring to the table as a result of years of experience running manual business processes. Ideally, users and decision scientists should always be partners who work together harmoniously.
I do believe that inefficiencies benefit no one in the long run, and that optimizing business processes is beneficial to many different stakeholders. And in a market economy, if a particular company maintains their historical inefficiencies, then they are at risk that some other, more efficient company will ultimately be more successful.
Net, anyone considering a career living at the interface of math and business needs to work through its challenges and enjoy its rewards.
*Adapted from an AI Workshop presentation by Ron Menich to 3,000+ attendees at the 2021 AiChef Conference. To see his full remarks, visit https://bit.ly/3rcKFyC.