Creators and users of data have incongruent goals, or to phrase it in a less judgmental way: everyone is sufficiently occupied with their own objectives.
A data domain almost always cuts across the organizational structure of a company, especially when corporate structures emerge. This simple reasoning explains why startups are often successful with data products for a long time—they don’t need Data Mesh approaches. A relevant data need is identified and implemented immediately. Startups work almost exclusively with clearly defined and well-formulated products, which they sufficiently understand and therefore always prioritize.
Corporations, on the other hand, need to organize themselves differently to obtain the data required for a successful product. Due to established structures and a simultaneous focus on many topics, acquiring the necessary data becomes challenging. This is where Data Mesh comes into play—it translates the mindset of a data product manager into a corporate structure, both in terms of expertise and technology.
A central element of Data Mesh, data ownership, primarily means assigning responsibility to the data creator to consistently adhere to a data agreement. But Data Mesh also involves aligning data requirements with product relevance. Concepts such as Data Lakes or the indiscriminate collection of “useless” data without utilizing it lose priority. Instead, data requirements must be negotiated. The “art” lies in requesting the right data, doing the right things, and articulating what truly matters.
A decisive factor here is the process of prioritizing topics: either a feature creates a USP (Unique Selling Proposition) for the product, generating revenue, or the data asset is expanded to enrich the data product later. Exploring data that could become assets is another possibility. The final case involves gaining insights, which is equally important. Striking the right balance among these topics is fundamental to success.
Data Mesh approaches are needed to create a data product in an environment where it would otherwise be nearly impossible, due to operational incentive systems posing obstacles.
Data Mesh provides a few principles as answers for building successful data products. It acts as guardrails, while the implementation itself remains the responsibility of the company.
sol4data offers concrete guidance tailored to your environment.
