Selling Data Science Projects Internally

Data Science projects are expensive and talent hungry. They require adequate data infrastructure (ADW and data lakes) that ingest and store detailed relevant data with acceptable quality for analysis and modeling purposes. In addition, the same business use cases in the same industries often require customized data preparation, analysis, and modeling approaches. Aside from these issues, a critical challenge is always selling such projects internally to business units within the company based on the value they provide. These people have deep business domain knowledge and expertise performing using established proven processes of the past. The business unit leaders in an organization may be in full support of such efforts. However, there is always a need to get buy-in and interest from other visionaries and stakeholders at different levels within a business unit for such efforts. This requires a full understanding of the business case and a good idea of the positive business impact the analytics effort could provide. Often the challenge is that a clear ROI cannot be established ahead of time. In this talk, I provide some real world examples of initiating and managing data science projects in different organizations. I will highlight the importance of an early and thorough business, data, analytics, and deployment assessment where the benefit of the project could be quantified as early as possible leveraging DS-BuDAI agile data science principles.