From Twilio to Propel: Building Real-Time Customer-Facing Analytics at Scale

In this talk, we will delve into the complexities and nuances of building real-time customer-facing analytics at Twilio, which were designed for handling data at an immense scale. We'll explore why achieving sub-second latencies and ensuring sub-minute data freshness are not just desirable but a requirement in customer-facing analytics. I'll walk you through the specific architecture we implemented at Twilio, dissecting its benefits and shortcomings to provide a comprehensive understanding of what works and what doesn't in such high-demand scenarios. This journey at Twilio was instrumental in inspiring the creation of Propel – a platform designed from the ground up to meet these stringent requirements. We'll discuss the technical lessons learned in the process of building at Twilio that influenced our architectural choices at Propel, especially how it caters to a diverse array of use cases in customer-facing analytics.

This session is for technical leaders looking to learn the nuances of deploying real-time analytics products that customers depend on while operating them at a massive scale.