Data Scientist – Subscription Data Enablement

Our paywall is a model. Come and make it smarter. Schibsted | Oslo, Bergen or Stockholm (hybrid) We built our own dynamic paywall and offer engine. Now we need to make it better, understand what works, and bring it to more brands. We're looking for a Data Scientist who enjoys building models, testing ideas and finding out whether the results actually make a difference. You'll work alongside experienced data scientists, analysts and engineers, with opportunities to take ownership of your work and grow your skills along the way. Why this work mattersOur democracies depend on independent journalism, and that is our business. Independent journalism needs readers who choose to pay for it. Our job in Subscription Data Enablement is to understand those readers well enough to offer them the right thing at the right time, and to leave them alone when we shouldn't bother them. Direct relationships with readers and digital profitability are how Schibsted measures success. Your models sit right in the middle of both. Who we areWe are the data and analytics team for Schibsted's Subscription business in Norway and Sweden: analysts, data engineers, an ML engineer and data scientists. We work with brands such as VG, Aftenposten, Aftonbladet, Svenska Dagbladet, Bergens Tidende and E24. Our models already cover churn, renewal, segmentation and engagement. The dynamic paywall and offer engine is the newest addition to the family. It is also a good test of how we work. The Dynamic Paywall project has many moving parts and competing priorities, and it needs people from data, product, tech, business and editorial teams in the same room. A consultant who recently worked with us said that, despite technically being a consultant, they "never really felt like one". They described a team where everyone brings different expertise, ideas get challenged, debates are healthy, and we don't always agree, but there is "always mutual respect" and a shared goal of finding the best way forward. We want you to feel the same way about us a year from now. What you will do• Make our paywall smarter. Develop and improve the models behind our dynamic paywall and offers: who sees which offer, and when. Propensity, recommender and offer optimisation problems are all in scope. • Find out what works. Design, run and analyse experiments with our brands, and report the results as they are, including what they don’t prove. Uplift modelling and causal inference are welcome tools here. • Build models that get used. Take models from notebook to production with our data scientists and engineers: feature engineering, training, evaluation, testing, and monitoring once they’re live. We work with tools such as Snowflake, MLflow and Streamlit. • Make the results understandable. Explain findings to commercial colleagues with clear analysis and dashboards where they help. • Improve how we work. Bring ideas on modelling, tooling and MLOps as more brands come on board, including how AI can make our own work easier. • Ask good questions. Including “does this actually make us money?” when someone needs to. What you bring• Python or R and SQL, used to analyse data and build models. • Hands-on experience developing and evaluating machine learning models, and an understanding of what it takes to run them in production. • A solid foundation in statistics and experimentation, including hypothesis testing and A/B tests. • The ability to explain your work to people who don’t speak ML, and the curiosity to learn from people who know the business better than you do. How we workWe follow a short list of team principles, and we mean them: be curious and keep learning, deliver quality together, design collaboratively, help early and ask early, share generously, have fun and stay open. In practice, that means Show n Tell and Friday data talks every other week, a monthly hack day, and an expectation that you ask for help before you’re stuck, not after. We also encourage sitting in on newsroom meetings. I