The article discusses the process of assessing whether to build or buy a data pipeline as a product manager. It highlights the author's experience in building custom ETL for their previous product, which consumed significant engineering time and delayed time-to-value for users. The author then presents a three-step process for making this decision: defining must-have requirements, comparing costs of build vs buy, and assessing trustworthiness of vendors. This assessment is crucial in the ever-expanding market of data solutions.