What are the best practices for data product management and how does it differ from software product management? Data mesh produced a fair amount of interest in treating data as a product. Treating data as a product implies some kind of product management. 

Data Generation, Structure, Release, Access, Security, Maintainability and data schema upgrade are synonym terms that match with software product lifecycle elements of requirements, design, implementation, release, maintenance and upgrades. 
The best practice for data product management is to establish and adopt optimised standard procedures in all these areas and rinse-and-repeat them in every data release cycle. This way whole data product release can be managed efficiently. 

Ensure the data is used properly and build a roadmap as to how to make it more useful and accessible.

I like the idea of app portfolio management tools to create efficiencies. It's important to continually examine the most high risk areas and look to transition them out or into different solutions.