Company: Teva/Anda/Allergen/Actavis

Building a Data Science Team at South Florida’s Pharmaceutical Hub

Client: Rob Maslanski, Director Business Analytics

Teva/Anda/Allergen/Actavis are R&D-driven pharmaceutical companies operating globally. With a library of well-known pharmaceutical drugs and access to an internal global supply chain distribution network the company, which has experienced aggressive mergers and acquisitions over a short period of time has developed many convenient ways for customers to access generic drugs worldwide. Thanks to these multiple acquisitions, Teva is the world’s largest generic drug maker. Given their global reach, a demand for a data science team was imminent to match modern buyer behavior and supply chain optimization. 

What was the problem that you were trying to solve in hiring or building your team?

“As the market got tougher and more competitive there became a need for data-driven decisions. In order to compete with larger players in the pharmaceutical industry as well as smaller, niche-market-focused distributors, we had to be able to answer questions objectively.

We had a need for data-driven decisions, not decisions made from “gut instinct” or intuition. Due to multiple mergers and acquisitions, our existing systems could not communicate with one another. These disparate systems ranged from e-commerce (the first point of contact with the customer) to inventory optimization, operations, and supply-chain distribution. Each proprietary database was acting alone. In addition to the problem of siloed departments there was also a lack of data depth; leadership realized that we were only scratching the surface of reporting. For instance, 2-dimensional graphs consisted of simple profit and loss statements and vanity metrics. With over 20,000 products/SKUs on a patchwork quilt of databases, we needed an overall plan to tie it all together – and the talent to implement such an overhaul. We needed data scientists and analysts to go deeper, unify all systems, and make the data tell a story.”

What made building your team challenging that made you decide to outsource recruiting?

“Two days before Christmas I had to lay out a plan to build out a data and analytics team. When a finance guy for 20 years is asked to build a data science team, there was a flood of new technology that I had to learn. I lost a lot of sleep studying every night about analytics, correlation, Pearson coefficients, and all the stuff I forgot from college. I was in uncharted grounds. Thankfully you were there to walk me through it.”

How did you feel in the process working with us? If you’ve worked with other recruiters in the past, what was different with us?

“Octagon did not just have a role to fill; they had a team to fill. We were working with a blank slate. Each new member had to cohesively merge with the current culture of the company; yet also work together moving forward. With limited time and resources, I realized the employees I was to hire needed to have three factors. They had to be: 1) technical, 2) good analysts, 3) business minded.

The corporate recruiting team kept changing and they were not technical enough to understand the requirements.  I wanted people with a background in mathematics and analytics, not finance. I knew I had the finance background to keep them up to speed in that aspect – but I needed someone who thought like a statistician, not an analyst. Above all, I needed viable candidates – that also had to fit within the budget. Octagon was the only agency that was sending viable candidates.”

How did the people that got placed bring success to your team? And how did that make your job easier?

“Looking back, it really is a great team. It’s cool when you take a step back and look at it from a technical point of view. How it all came together, how we built that team. I think I got a really good understanding of how technical the candidates needed to be after I spent a long weekend attending TDWI, an analytics training seminar; that was a turning point because it opened up my eyes to a whole new world.”

Can you give an example?

“One data science analyst that you brought in had a Ph.D. in pharmacology and a background in data science. She was able to crunch the numbers and find out that a competitor’s promotions were adversely affecting our business. Her report found that we were losing $5 million a year when companies did a certain promo. This certainly impressed the leadership initially and we were given the go-ahead to implement similar such projects.”

What were some of the challenges faced when working with multiple databases?

“We had to move away from Qlikview and other tools and replace it with SQL on the backed and visualize the data with Tableau on the front end. Migration was certainly a challenge given factors such as naming issues, duplicate labeling, and the merging of cultures from what used to be different stakeholders. We needed to have anomaly detection, standard deviation breaks that would flag certain actions, and an early warning inventory control system.”

How does the environment look like now?

“Hints of prescriptive but more predictive. We’re slowly but surely moving towards the former.”

How was the change from “vanity metrics” to predictive/prescriptive data received by the organization?

“Some days, people hated me, other days they loved me. The data certainly tells an objective story, and the reports we created would rouse some people, especially if it was their initiative that was not the right decision initially.”

What do you have in mind for your team in the future? Where do you see “big data” headed?

“Certainly we want to move across the continuum from Descriptive to Predictive, to finally Prescriptive modeling. We’re having the conversation around headless A.I. and task automation. I think as “machine learning” moves into “machine decision-making” we will need more technically-saavy team members.”