Viewing entries posted in May 2017
In our latest blog post we take a look at a hot topic in BI: segmentation. One of the downsides of RCTs is they usually only identify what works on average. Although this allows us to improve outcomes overall, you can do even better if you can work out what works best for different groups of people and then tailor your approach to each group. We take a look at our hospital appointment trial and discuss both ‘classic’ segmentation and a newer approach based on machine learning.