Data science can pull meaningful, actionable insights out of your data that can benefit your business in a big way – when it’s done correctly. Technology is obviously a major component of data science, but people and processes are equally important. If the people and processes in the “data supply chain” aren’t operating effectively, no amount of technology will give you the insights you need.
Our client, a large health insurance provider, had data science teams embedded in each business unit, as well as a shared services team. But the efficiency and effectiveness of the teams wasn’t what the company had hoped for. The result was lost time, lost opportunities, and frustration in both the business units and teams. It was clear that the company wasn’t using data science to its fullest advantage, but it wasn’t clear why or how they could correct it.
The company purchased multiple leading data science technologies, yet still faced difficulties implementing successful data science solutions effectively and efficiently. They engaged Prolifics to help identify the root issues and how to achieve better outcomes for both their business and data science teams.
As the first course of action, we all agreed that a step back was needed to evaluate and document the current issues in greater detail. Together, we found:
At the business units –
- The business units often did not have – or could not generate – ideas for data science use cases, or the ideas were not suitable for data science methods.
- When the business units did generate use cases, many times they were not “mature.” That is, the availability and/or quality of the necessary data was poor, or the case itself would not generate an acceptable return on investment (ROI).
- The data science teams tried to clean up the use cases and fill in the necessary gaps – taking time away from actual data science analysis.
- Sometimes weeks were spent before it was finally determined that certain use cases were not feasible or were not cost-justified.
At the data science teams –
- There was no consistent methodology to identify, prioritize, develop and implement data science solutions.
- Many data science teams managed data science projects using a Waterfall-centric approach, meaning projects done in a step-by-step, linear fashion.
- Many data science teams took business requirements and conducted research and development with little or no collaboration with the business unit. This led to solutions that the business units often considered ineffective to address their original requirement.
- Data science executives saw the value of using an Agile approach (a more flexible, iterative process than Waterfall) but had no clear idea of how to apply Agile in data science projects.
Together with the client, we determined the need to establish comprehensive processes for both the business units and the data science teams to correct these issues. Our team applied a methodology to optimize the work of the business units, the data scientists, their processes, and the technologies to come up with good use cases that brought value to the organization. This holistic approach included:
- Holding ideation / visionary workshops that trained the business units to flesh out ideas for analysis.
- Building a use case maturity assessment program with the business units. The assessment helps the units judge the readiness of their ideas and related data for action, so they could present the data science teams with mature case studies.
- Working with the data science teams to develop a more Agile approach to their model development.
- Operationalizing the data science models so that when new data goes in, the models return real-time recommendations to the business units.
- Developing governance for data and model usage, to ensure the consistency and quality of model recommendations.
- Developing a process to monitor and measure data science projects.
Now the client has an end-to-end, Agile data science process that encompasses the business units, the data science teams and operation teams – and includes the ability to monitor and measure key performance indicators (KPIs). The business units feel empowered. They’ve been equipped with ideation and assessment tools that generate feasible analytic projects that positively impact the business. The data science teams are now much more efficient and effective, focusing exclusively on analytics and generating iterative models more quickly. One team reported that the Prolifics-implemented Agile approach has freed up 30 percent of their time.
Would you like to know more about the latest trends in data science? Prolifics’ data science experts help companies enact effective data science programs and processes through our holistic approach. We are vendor agnostic and work with a number of leading data science technology providers. We also offer our own solutions – most notably Data Science in a Box, a fully managed, turnkey offering that works across multiple technology platforms. Data Science in a Box creates clean, structured, and governed data from all the sources across your platforms. Take a few minutes to read about our Data Science in a Box solution, visit our data and analytics services pages or email us at firstname.lastname@example.org.