It’s exciting being a data scientist at a consulting company because you get to move between projects and explore different industries. Every few months, I learn about a new organization’s business needs, explore vastly different data landscapes and design a whole new analytics solution for a business issue. Planning and designing an analytics approach is by far my favorite part because it’s so hands-on. Doing the ground work in the planning stages makes or breaks a project: It means we can start with a strong project foundation and build a solution that we are proud of.
I worked on a project with a global pharma company where we were developing a solution to a problem no one had tried to solve before. The company wanted to increase the speed at which they could run clinical trials of its new drugs so that they could bring those drugs to market faster. During this engagement, I was lucky to work with senior data scientists on the QuantumBlack team who taught me everything.
Our client wanted to see if strategically selecting the hospitals where they ran the clinical trials could increase patients’ clinical trial enrolment speed. We approached this problem in two parts. First, could we predict, for a given new trial, how quickly we could expect hospitals to enroll patients? Second, could we then run an optimization algorithm to select the best pool of hospitals, that combined, would sufficiently enroll patients the fastest?
We used data to validate a number of potential hypotheses, and after many weeks of iteratively adding, testing and checking new features, we were able to accurately and predict actual hospital enrollment. Phew! We then worked on the optimization. The key was to capture the client’s constraints–e.g. getting the drug licensed in the United States–as well as potential costs.
What made this project truly exciting was the scale and effect it had. We were putting this into production, and colleagues were working on a front-end system that the client could use, all powered by our analytics.
The client now uses this tool as the backbone of their clinical trial planning process–completely changing the way their trial managers and planners work. I feel incredibly proud of what we produced.
There are two characteristics to this project that I have consistently seen throughout my time at McKinsey. First, we often take on first-time projects, which are great fun to solve. Second, we are always able to work with a great team and continually learn from each other.
This was just one part of my year. My other projects are just as exciting, and they (thankfully) keep coming!
Find a job like Paul’sAbout Paul
Paul is a data scientist on McKinsey’s QuantumBlack team based in Singapore. Originally from London, Paul earned his bachelor’s and master’s in mathematics and his PhD in technology and medicine from the Imperial College London.
For more information on McKinsey’s data science career paths, visit mckinsey.com/TechCareers.