Attention, a Digital department alone cannot carry this transformation, because it rather has the role of exploiting the data with new methods and techniques. It needs data but does not generally have the vocation of carrying data governance or creating, alone, all the necessary capacities.
Evangelize business teams to embark on the adventure
The exploitation of the knowledge obtained by the data can be greatly facilitated by IT. But to bring disruption it is essential to successfully embed the business and especially help your business to understand the importance of data and innovation, in accordance with the strategy of the general management. Next, who carries this business model disruption part depends on each company.
One of your missions is therefore also to help business teams understand that it is not enough to switch to the cloud to go from logistician to computer scientist or automotive supplier to rental of connected cars. The use of project roadmap happens to be essential there.
Data management needs structuring and modeling
We must not forget the fundamentals either. 80% of a data scientist’s time consists of correcting data (cleaning dirty or non-comparable data). It is impossible to industrialize without structuring the data. Modeling remains the first foundation of data management, and involves a homogeneous (shared) vocabulary, a knowledge of where the data are, and the choice of data according to their quality.
Without modeling, the risks are numerous:
- Redo multiple interfaces / services, and then lose agility instead of gaining it
- Create a data warehouse or a datalake very quickly unusable
- Having too long data science projects and exhausting scientists
- Complexify security
Strengthening your ability to process data deeply disrupts the IS
After having cleaned and modeled the data, it will be time to build up the performance and knowledge of the company thanks to the vision brought by the data. Data processing and analysis (structured and unstructured) will be based on data exchange capacities and API management, master data management systems, data repositories (to describe and classify data), analytics, text mining, machine learning, data processing orchestration and data visualization.
It is important to remember that these capabilities cannot be treated as an additional small point in the SI.
Conclusion
They imply a profound change in skills and perspective in the technologies to be automated. For example “classic” ERPs will allow you to automate your functions well but they will give you little freedom on the information model, the interfaces or the modalities of transformation of your process. For this, it will be necessary to go through data processing capacities (see above) and technologies such as BPM (Business Process Management) which bring data transparency and the flexibility of processes and adaptations.
