Advanced analytics enables CrowdTwist to maximize channel engagement

Pythian's Data Science team helped CrowdTwist identify several predictive customer data factors to improve their models, resulting in improved customer segments and reports.


Business Needs

To better deliver on their engagement solutions, CrowdTwist wanted to improve the client experience with more visible information, a better variety of reports, and a balance between incentives given and the financial liability associated with them.


CrowdTwist engaged the Pythian Data Science team to assess the potential of their current dataset and to build a model to determine how their users clustered together and distinguished from one another. A user model was built, leveraging Facebook ‘likes’ and client IDs as distinct features that would reveal customer preferences. Further cluster analysis provided by the data science team focused on users as a means to test how to identify clusters that drive activity. Pythian’s advanced analytics and cluster analysis provided insight into the optimal spend per channel for each user, which translated into actionable outcomes for the end client.


The Pythian Data Science team was able to identify several predictive customer data factors such as the chosen target and activity level to within a standard deviation. This is significant as the range of values is very large. A refinement of the models found an optimal dollar-to-point conversion, an optimal activity-to-point conversion, as well as other ways to optimize the different loyalty programs in their portfolio. The outcome was measurably improved customer data segments and streamlined reports – resulting in a better client experience with the loyalty and engagement platform overall.