Izzy has almost 20 years of working as a data scientist
Worked in multiple industries: Academic research, Bio-tech, Gaming, Fintech, consulting and e-commerce
Loves the theory and practice of statistical and ML techniques: from hand-crafted statistical models to large scale text/image recommendation problems
Bayesian by belief but a big proponent of GSD (get s*it done) and KISS (because sometime the average/median as all you need!)
Pilots are a great way to test user response to new products or features without taking too much risk, but the small scale of most pilots means there are limits on what we can learn from them. A big shadow looms over all the KPIs we measure: Selection bias (especially for opt-in pilots). In Shopify we love pilots, and we've been working on ways to extrapolate our learnings from them so we can help make better decisions. In this talk I will review several strategies you can use to augment your learnings and how we communicate these results to stakeholders.