In simple words, data product is

"applying product thinking to data"

Lets decipher this - Product thinking is simply problem-solving. So, a data product should be solving a specific problem, a problem of a consumer. To solve a specific problem, we need to first understand the problem and then create a solution. This is the paradigm shift in data world. In the data warehousing and data lake world, building data "assets" has always been forward and never backward facing.

If a product is solving a specific problem, then it needs to be independent technically, such product needs to evolve on its own. Such independent quantum is known as Architecture quantum.

In book Evolutionary architecture, an architecture quantum is defined as "an independently deployable component with high functional cohesion, which includes all the structural elements required for the system to function properly."

Putting it all together, So a data product contains all the technical structural components required technically that are needed to solve a specific problem of a consumer.