Solving business problems with data. Sounds trivial these days, no? Those of you who have tried taming data into valuable solutions and integrating them into your products, know the truth. When working on a business problem in a data science prism, potential solutions often require a significant research phase, during which the team can gain confidence as to whether a solution is viable and to gather information about it. During this phase, we focus on diminishing uncertainty so we can make data-driven decisions.

As Head of Data Science, I need the freedom to explore projects without committing to results; sometimes, it is impossible to know in advance whether we can solve a given business problem and how well the solution will perform. As a PM working with these teams, my goal is to challenge them with well-defined business problems & clear value definitions, while maintaining their focus to ensure the company can leverage their work promptly with minimum risk. We have developed a work methodology that helps us balancing these seemingly contradicting interests.

In this round-table, we will present our methodology, and show-case it on examples raised by the audience or by our own experience. We will discuss together how we can refine and adapt it to the project type and to the specific team/company.