In this training session we will go through designing a good experiment in an online world, from hypothesis generation, KPIs, pre-analysis and post analysis.
The idea is to help marketers, analysts and optimisers think more like scientists and statisticians. By the end of the session, we hope that you will be able you will be able to give your experiments a fighting chance by understanding that meaningful results start with good design, why prioritisation is important, why no results can have non-commercial impact and of course, the statistical methods that underpin all of this. This is not a “”best practise”” session nor will it be the definitive guide. Ultimately a good experimenter will look adapt the learnings from this training session and learn to apply them in their team and company.
Bhav is head of analytics at MOO ltd, a business card printing and design company. He manages the site & product analytics team whose role is to bring insights to the business and aid in the decision-making process. His team oversee experiments at MOO and are responsible for helping scale the experiment programme by building a test and learn culture. In the past Bhav has managed analytics and optimisation at PhotoBox, News UK and Ladbrokes, (a Ladbrokes case study can be found on the Qubit website: https://www.qubit.com/case-studies/ladbrokes/). Bhav also runs London’s biggest (and best) conversion rate optimisation, analytics and product meetup called CRAP Talks and in his spare time likes to play chess and hang out with his two year old son.
laptop (with internet connection), notepad and pen and most importantly an open mind and willingness to learn.”
Who is it for: This session is for people who understand the importance of running experiments and are looking to improve their knowledge on experiment design and statistics. The session is aimed to help those looking to build a solid foundation.
Who is it not for: If you already have something that works for you and your company, it’s unlikely that this session will add any additional value.