Uber was looking to launch a new range of services aimed at their business rider audience, in order to acquire new customers and ultimately drive profits.
The data science team had segmented their business customer audience and needed to understand the key similarities and differences across 5 different segments, in order to inform and inspire product development and marketing communications.
Working closely with Uber's data science and design research team, we recommended and executed the following plan.
1. Project Definition
We worked closely with Uber data science and design research to clarify the overarching business/product metrics, agree the supporting research aims, key hypotheses, research design and resource allocation across the project, inc. key deliverables.
We split the analysis phase into two parts based on the client needs: (i) investigating key hypotheses we had formed with the team, and (ii) an exploration of the data.
For the exploration, we started with basic techniques first, at a high level (e.g. across all segments), to identify key differences across segments. We gradually progressed to more sophisticated techniques, including correlation and regression analysis, where required. We recommended this 'lean' approach to most efficiently and effectively allocate resources (exploration with a big data set is resource intensive).
Rather than produce a huge, 80 slide powerpoint deck, we recommended producing a much shorter, punchier form that would both cost less and more effectively inspire decision-making. This included:
- 10 highly visual slides illuminating the key differences across the 5 segments
- One-page summaries on each segment, including a description that bought each segment to life
January 2016: results are currently being implemented and further information will follow.