Through its four main pillars, Data Mesh truly moves away from the dogma of centralisation and all-technology in favor of a global approach based on federation. Data Mesh thus promises to be at the heart of company data strategies and organisations.
1- Data Mesh: the ultimate model for data-driven companies?
2- Data domains: Data Mesh gives business domains superpowers.
3- Data Mesh: data is a product
4- Data infrastructure self-service as the technological driving force behind Data Mesh
5- Data Mesh: federated governance to guarantee efficiency
6- Data Mesh: Practical examples & feedback
According to BCG, data is now like oxygen in most sectors. To take advantage of this, companies are striving to build up an indispensable foundation of capabilities, which also have a place in Data Mesh.
The ambition in the long term is to rise to the rank of ‘data masters’, to which the digital giants and a few multinationals seem to be able to aspire at the moment. While ‘some have a proven benefit from data, many are still not achieving their goals, giving leaders a growing advantage,’ notes BCG.
A federated data path and distributed responsibilities
The data maturity of companies as a whole is nevertheless progressing. But there are always obstacles in their way: the slow acculturation of businesses, the improvement of data quality, made more complex by an explosion of sources and volumes, and the hybridisation of IT systems.
Data Mesh is based on a dual model of distribution and federation. And it is clearly no coincidence that a growing number of professionals in the world of data are interested in this ‘total’ or global concept, which makes it possible to make companies data-driven ‘by design’.
With its four pillars, it offers an approach that is convergent with many of the directions companies are currently taking to reach a higher level of maturity and increase the value generated by their data strategy.
Data Mesh is based on a dual model of distribution and federation.
Taking the Data Mesh route is not a blank slate for organisations already engaged in their transformation. For the less mature, this new paradigm is even an opportunity to define a coherent, state-of-the-art strategy involving all internal stakeholders.
Data Mesh: revolution or evolution of data strategies?
A few examples illustrate the convergence between Data Mesh and the new data trends that aim to finally break free from the limits of centralisation and promote scaling up:
📌 Companies are striving to obtain the support and direct involvement of business likes so that they allocate resources to the identification of use cases and the implementation of data projects. Data Mesh advocates an orientation by data domains, which can be likened to a business division of data and uses. It gives power and autonomy back to operational staff so that they can finally control their own data destiny.
📌 The trend is towards the design of data products, which can be dashboards or machine learning algorithms. This product approach implies a different logic from that of simply delivering an application. Data Mesh conceives the data as a product and entrusts its development to the domains.
📌 The objective is to rationalise and consolidate groups around a single data platform (often taking advantage of a Cloud orientation). Data Mesh retains this concept of pooling, with emphasis on the agility and management aspects of the platform for self-service consumption.
📌 Organisations are trying to break away from top-down governance, which is decided and controlled entirely by a central team in top-down mode. Data Mesh goes beyond this by thinking of a governance (of data and products) that reconciles central and local, that is federated ‘by design’.
Data Mesh therefore does not break abruptly with the rising trends among data strategies. It draws on them to design what may appear to be the next generation of data policies.
Data Mesh therefore does not break abruptly with the rising trends among data strategies. It draws on them to design what may appear to be the next generation of data policies. Data Mesh is therefore above all a trajectory, a path, iterative. It is interesting because it proposes a global framework applicable to all companies. It is therefore not really a revolution, but it does clarify & give coherence to the many directions taken by companies. In addition, organisation into data domains enables data to be placed at the heart of the business. In the end, it is this pillar that is the most revolutionary and also the most complex to implement.
Data Mesh is not a specification, but a compass
Data Mesh is not a collection of best practices that will naturally lead to success. Nor is it a specification that should be followed to the letter. In many contexts, this would be impossible or counterproductive.
However, its approach does contain a great many elements of response. And it is above all its implementation, the quality of its execution and its appropriation by all parties that will make it a lever for internal improvement and a springboard towards the status of data champion or data-driven company.
In many companies, the Data Mesh philosophy is already taking hold, even if it is under a different name or without explicit reference to it. This is particularly true with regard to the ambition of greater involvement of the businesses and the distribution of skills and responsibilities, as outlined above.
The desire to consolidate around a single platform, often in the cloud, and the development of product logic are also points of convergence. Utopia or reality, the breakthrough embodied by Data Mesh is in fact already under way.
If Data Mesh is not a set of specifications, then it should be seen as a compass. Or better still: as a North Star. Data Mesh provides direction, and the company will create value at scale with its data with every step it takes towards it.
This article was written in collaboration with Christophe Auffray.