Please let's stop talking about data as the “new oil”. The industry has really done that for long enough now. However, the fact remains: more than ever, data provides companies with important insights and knowledge to successfully drive their business forward in areas such as service and customer management, product development, marketing and internal process optimization. Even if the reality is often different: According to a CXO survey by Statista and diconium, more than every second company in Germany is still not exploiting its data potential satisfactorily. There are many reasons for this. In addition to a shortage of skilled workers (45%), it is primarily technical hurdles (39%) and a lack of data availability (36%) that are slowing down their data projects. At its core, however, it is almost always about the organization of data management.
For a long time, a centrally organized data architecture was considered the ideal way to make data usable company-wide. It is clear that large companies in particular have many departments and employees who work with even more suppliers and partners. In order to avoid the formation of silos and thus a faltering flow of information, the input from all data sources is traditionally bundled in a central solution. So far, so good. The problem is that all company-wide information is managed by just one data team and organized in the central solution. However, these data specialists do not have the expertise of the individual departments to know exactly what technical content is involved and who should use it for analyses and in what context. As a result, not only agility suffers, but also the quality of the integrated data. The strategic answer to centralized data teams is the decentralized “data mesh” approach.
Instead of just one data team that has to take care of all data sources at the same time, the data mesh approach uses cross-functional teams that are only responsible for their specifically assigned business domain. This democratization of data management therefore shifts responsibility for data to the specialist departments that know their data best. In future, all data products will be created in collaboration with them. What is important here is that the organization of the data architecture in the decentralized approach is based on networking instead of centrality. This avoids data silos from the outset. In addition to the principle of domain ownership in the individual specialist teams, the most important pillars for successful data mesh deployment are the basic attitude of “data as a product” for the transfer of product thinking to data, a self-service data platform for greater efficiency through standardization and federated data governance for added trust and security.
How suitable the data mesh approach is for a company ultimately depends on various factors. These include the degree of complexity, company size, business integration, number of teams, data sources, data use cases and data fragmentation. One thing is clear: there is no universal solution for maximum data success. However, the data mesh approach offers numerous advantages such as the elimination of bottlenecks in centralized approaches and greater flexibility. Implementation requires customization to the specific needs of the organization and is a gradual, long-term process that can take months to years. Particular attention should be paid to the topic of data governance to ensure that the distribution of responsibilities is implemented without compromising the integrity, quality and security of the data. Data literacy, i.e. the right data competence, is fundamental for all data owners. We at diconium are happy to support you and your company in organizing your data management. Simply get in touch with us!
You can find more information on the topic of data mesh, including lots of helpful insights into the four basic principles and how to get started with more efficient data management, in our latest guide “Data mesh”, which is available to download free of charge.