ML-based marketing in ride-hailing app

How Data Analytics helped boost user audience and increased revenue.

Tez Data consultants set up more sophisticated Data Platform to collect all points of contact with potential customer from impression through install and to making an order. This allowed us to analyze the funnel and estimate the effect of a number of small bugs that were known but haven't attracted enough attention from the management. This resulted in +5% of install to registration ratio. We understood that ride-hailing is a new type of business model for most of the population so we tried to find special occasions when it could be offered to grow a user habit. We detected that birthdays and national holidays are 2 of the most popular drivers to use the ride-hailing app. Based on that we built AI-based personalized offers system that maximized LTV of the customer. Next we developed the robust AB-testing setup for this case that accounts for spill-over effects and detected +14% conversion to first order given the same promo budget. Given the solid proof the management of the company approved the project for implementation into production systems. At this stage Tez Data consultants analyzed the IT architecture of the client and proposed optimal solution to deploy the system. It happened to be a containerized Machine Learning model that run on the daily basis. We also implemented a monitoring system for the model not to run over budget and trained company personnel to provide inputs to the model.

This half-year project helped our client to out-grow its competitors and increased service level due to economy of scale effect.