What are the four principles, and what is the interplay between them?
By moving data ownership closer to a business domain, we minimize the gap between the source of data and its uses, thereby increasing data business truthfulness.
However, this can lead to data siloing, resulting in data that is not easily accessible, understandable or even usable outside of its domain team. We can break down these silos by treating data as a product: facilitate sharing by knowing our customers and their use case and ensuring our product can be found, understand and used both inside and outside the domain.
Decentralized domain teams can gain autonomy by leveraging a self-serve data platform. This reduces their dependency on other teams and their priorities.
Decentralized teams are great, but how will their data products interoperate with those from other domains and teams? Develop a system of federated governance, where global policies and standards are set by a central cross-functional body, while conforming local decisions are made and implemented at the domain level.
Applying product thinking to analytic data can help teams share high-quality datasets with their customers.
However, there is a lot of work involved in building, maintaining and publishing a high-quality data product. Much of this work is common to all data products, and we can delegate the brunt to our self-serve data platform. This enables us to increase our speed to market and hence lower our costs.
We want our analytic data to be shared, but it must also be secure. Data products rely on federated computational governance to provide keep their data secure and safe.
Finally, the self-serve data platform provide affordances for monitoring and enforcing the security policies enacted by the governance bodies.