What is the mortal sin of software engineering? Wasting effort, time, and money building something that is never used. How does it happen? There are usually three causes: no product-market fit, overengineering, or lack of trust. Unless you run your own startup, there is little you can do about the product-market fit. Similarly, it makes no sense to discourage you from overengineering at a conference that has "buzzword" in the name ;) Let's focus on the lack of trust. How many of you have built event collection pipelines and dashboards that duplicate features of Google Analytics? What happens when your data differs from values in GA? It does not matter how long you talk about data sampling and eventual consistency. People immediately assume that there is a bug in your pipeline. In this talk, I'll show you how to build trustworthy data products. I start with defining business requirements and involving users. No, it is not an optional step. After that, I talk about ensuring quality during data ingestion and processing. I admit that it will be a strange tech talk. No buzzwords. No new toys. A little bit of machine learning. Only things that will increase your job satisfaction and let you sleep calmly at night. Nobody wants that, right? ;)