Within product: this makes sense when the product is tightly related to data and when the organization relies on data primarily for feature testing and other product analytics use cases.
Centralized: This is arguably the most straightforward team structure to implement and a go-to for companies taking the first steps to become more data-driven. Both the team and the architecture are centralized here.
There might be, in this case, the existence of “specialists,” often Data Analysts or Analytics Engineers within each business function that have enough technical skills to communicate and autonomously work with the data team while also having a business background.