The success of an experimentation programme depends upon its cultural adoption within the business. With commoditised tools on the market that lower both software and statistical barriers to randomised controlled testing, organisations have the opportunity to democratise experimentation and continuously iterate towards their north star.

This talk addresses two key questions that stretch an organisation beyond the tool itself:
– What are some of the key cultural elements and technical gotchas to consider when initiating an experimentation programme?
– What can we do when randomised control testing is not possible and yet the need to identify causal impact remains?

We will illustrate lessons learnt and opportunities pursued by way of tangible examples from YouView’s own journey with product experimentation and causal inference. You’ll take away some insights on how we have approached hypothesis testing on our UX and hear about data science techniques that extend our own toolsets far beyond the classic chi-square test.

Technical level: Introductory level/students (some technical knowledge needed)