Lean Analytics: Sprinting data science

How might we apply Scrum principles to lean analytics? Similar to a product backlog, we start out with a hypotheses pool that we prioritise. During hypothesis selection we create an experiment by defining the metrics and scope, the duration of the data analysis sprint, the level of statistical analysis, integration, and finally the approach for implementing a pilot. While we run the experiment we construct an audience, extract data and perform statistical analysis. This creates the foundation for reports, recommendations and calls for possible action.
In the team we need a data engineer, a data analyst, a data scientist, and a team facilitator.
Tools are Google Analytics, optimizely, tableau and languages such as R and Python.