When it comes to storing application data, a more or less object-relational approach is currently the norm. That means, the current state of the application is stored in the database, and whenever that state changes, the database contents gets overwritten and replaced by the new state. This does not allow us to reason about information we had in the past, e.g. "When did this user change her address?" or "Where did this user live previously?". In other words, we drop data - data that might be important for us in the future. Event sourcing circumvents this amnesia by capturing each and every event that happened throughout the life of the application, and by deriving the current state from the total volume of these events. In this talk, I will show how event sourcing works in general and which peculiarities we need to cater for when applying it to a typical node.js application.