Data Scientist, Experimentation

About Tripadvisor The Tripadvisor Group connects people to experiences worth sharing, and aims to be the world’s most trusted source for travel and experiences. We leverage our brands, technology, and capabilities to connect our global audience with partners through rich content, travel guidance, and two-sided marketplaces for experiences, accommodations, restaurants, and other travel categories. The subsidiaries of Tripadvisor, Inc. (Nasdaq: TRIP), include a portfolio of travel brands and businesses, including Tripadvisor, Viator, and TheFork. At Tripadvisor experiences, the only thing we love more than travel is data. We slice it, we dice it, and we use it to empower our decision making. What You will do: As a Data Scientist on our experimentation team you will work with product teams across Viator to measure whether the changes they ship actually worked, defining the metrics that matter and making sure the decisions that follow are sound. Part of your contribution is making experimentation easier for the teams you support, through reusable tooling, clear documentation and consistent practice. Viator sells experiences that travellers often book once a year, in a marketplace where supply is finite and shared, so you will be learning to navigate problems where the obvious analysis can give the wrong answer, with support from senior practitioners who have done it before. You will: Design and analyse experiments end to end with the product teams you support, working directly with the Product Managers and Engineers who will act on the results. Turn product questions into testable hypotheses with pre-registered primary metrics, appropriate guardrails, and an honest view of what the available traffic can and cannot detect before the experiment starts. Define and instrument feature-level metrics, understanding how metric choice affects sensitivity, interpretation and the decision a team is trying to make. Investigate results properly, checking assignment integrity, exposure and data quality before conclusions are drawn, and treating a surprising result as something to diagnose rather than announce. Build reusable queries, tooling and templates and apply our shared protocols, so teams can run good experiments faster and with less rework. Communicate findings clearly to technical and non-technical audiences, including inconclusive and negative results, with a recommendation attached rather than a table of numbers. Look beyond whether a change worked to why it worked, and flag when a result may not hold for other users, markets or time periods. Carry out analysis beyond experiments, including opportunity sizing, funnel and behavioural analysis, and observational measurement where randomisation isn't possible. Document hypotheses, designs, outcomes and decisions so results remain comparable and the organisation compounds what it learns. Grow your own depth in experimentation and causal inference, with review and…