Our client is a US-based audio content subscription business.

They wanted an improved customer lifetime value solution that would generate robust, trusted estimates even across new markets and products.

We worked with their in-house data team to develop a Python package and a set of hierarchical Bayesian models.

The models are used to help guide decisions on optimal acquisition budgets and to plan.

Our client is a global furniture retailer.

They wanted to understand the impact of new store openings across different locations. They wanted to know the ‘cannibalisation’ impact on their existing store network and where the best new store locations are.

Our solution was a suite of models that feed a user interface. The models provide predictions for the impact of store openings. The tool enables users to easily run simulations to investigate different scenarios of store location and store attributes.

Our client is a popular UK weekly magazine, looking to better optimise their marketing plans across both offline and digital channels.

We worked together to build a Bayesian MMM R package for their in-house data team to use to run weekly model refreshes. The package includes complex optimisation functionality, data visualisation outputs and the ability to incorporate geo lift results into priors.

They have used the package to replace their previous vendor and gain control of their measurement.