Anneleen Rummens

Every data project runs into some version of the same question: “When will it be done?” Part of what makes it hard is that data work often reveals the work while you are doing it. You start with a plan, then learn things that change it: the source data is messier than expected, definitions do not match across teams, a join does not behave the way it should, access is missing, or the first results raise a better question than the one you started with. For many data people, that detective work is part of the appeal. It is also what makes a firm date hard to promise too early. That does not make data work entirely unpredictable by nature. But it does mean estimates need room to evolve. The easy answer is a single confident date. A more useful answer still helps people plan: it gives the best current estimate, the confidence behind it, and the assumptions that could change it. That way, when the work reveals something new, the timeline can be updated openly instead of turning every new finding into a missed commitment. This non-technical, practical session is about estimating and communicating data work without pretending the uncertainty is gone. You will learn how to separate discovery from delivery, estimate in ranges, name the assumptions behind a date, and narrow estimates as the work progresses. We will also look at how to give better status updates than “on track,” how to explain what has actually changed, and how to handle stakeholders who still want one fixed number. It’s for anyone who has had to put a timeline on data work - whether it’s a dashboard, report, pipeline, model, migration, or AI prototype- while the work was still unfolding.