5 Agile Leadership Practices Where CIOs Can Help Data Scientists

I've compiled a number of posts on data landfills and other bad data practices and have made a commitment, at least on Social, Agile, and Transformation to begin providing solutions.

I've always felt that other disciplines would benefit from well established technology practices. Agile practices have enabled software development teams to find sponsors, prioritize work, change the culture, insure work gets done, and market their accomplishments. I think data scientists face similar challenges and should benefit from many agile practices that have helped transform software development organizations.

I say elements, because while software development is often a collaborative practice performed by teams, that isn't always the case in data analytics work. Data scientists may not be in the same organization (team) and are often working individually or in pairs on different analytics. So while many agile practices are relevant to data science work, they have to be adapted to the nature of how this work gets done.

Also, in this post I've started with the leadership practices and might cover management practices in a follow up post. Key agile leadership practices are below:

• Sponsor work - Data scientists, data geeks, quants, data analysts, bi specialists - all go by different names in different organizations but business leaders don't always know how to best engage their capabilities or services. The CIO can lead the way by sponsoring analytics projects or drawing attention to a team or individual's capabilities. The CIO also has access to the organization and can help network departments that have high value data analytics work and are ready to partner with or hire data scientists. Sponsoring the work begins to establish an "Owner" role, similar to an agile product owner role, that can define a vision, articulate business value, and prioritize work.  


• Address the culture - Becoming a data driven organization is not just about having data scientists, it requires a commitment top down and bottom up to leverage data in decision making. This is often a culture change that requires leaders to educate the organization and find ways to align on simple practices. One of them, is to educate the organization to ask questions. Agile is not just a process - its a culture change that requires teams and organizations to think agile. 


• Establish practices for prioritization - Prioritization is a key practice for agile technology teams that have to align their efforts on a product release or development sprint to features and fixes that provide the highest business value. Data scientists face the same challenge in determining what questions to answer or analytics to prioritize. Leveraging agile practices and tools to help make the data scientists' workload transparent and establishing practices to prioritize work is a good place for CIOs to add value. 


• Review results and ask questions - Agile development teams will demo their work after the sprint and answer questions from sponsors. Data scientists would benefit by adopting a similar practice by schedule analytics reviews where they can showcase a visualization, tell a story, and suggest follow up work. CIOs can help by promoting these sessions, attending, participating, and asking good questions. 


• Get out of the away - Agile has its self organizing principles, enabling teams to have some authority around how they organize work to get things done. Data scientists also need a little bit of freedom to be who they are - scientists. Sometimes that means blazing a trail in new areas - new technologies, new data sources for example - to determine if they are useful to get a job done. Sometimes that means creating some work arounds, or creating "data processing debt" (more on this in another post, but this is the data analogy to technical debt) in order to get a job done on time.  

While I don't think these are new concepts, I haven't heard too many data scientists and their managers describe their culture or work with these practices. Similarly, while CIOs are more often consumed by Big Data platforms, I rarely hear them talk about aiding data scientists with basic practices. Common ground?