Data governance: the 7 challenges of an innovation machine

1 – Involving businesses to motivate them

Customer data, industrial and commercial information, financial data, etc. : in an organization the data are numerous and are constantly increasing with digital. What they have in common: They are owned by a business function or a support function. If historically some data is shared between functions (customer data between marketing and sales, product data between studies and production, financial data between sales and DAF, etc.), today it is the complete transversality in the organization that is able to create value through innovation.

But can this transversality be decreed? Yes in part, with regulations (and related penalties) mandating global data management as part of a compliance obligation. This is the case, for example, of the GDPR, for which the company must demonstrate its control and the correct management of all personal data processed within the organization.

But to be a true source of innovation, this transversality must be acquired: the business lines must understand how sharing their data with the other players in the organization can bring them benefits, and therefore motivate them. The use of artificial intelligence provides a good example. It is by pooling all data for global governance that all companies can in return benefit from the disruptive innovations offered by AI: detection of weak signals, detailed understanding of customer behavior, etc. on the basis of stable, reliable and controlled data.

If the proximity of the data science team to companies can be an additional motivation, an evangelization and onboarding phase is essential. This is one of the roles of the CDO (Chief Data Officer) who is there to organize knowledge, promote exchanges, ensure data reliability and compliance. It is also there to involve every data “owner” in the process. Specifically, to build this shared culture of data, the CDO will have to find corporate sponsors – also at the CODIR level, set up a charter of the Data Governance approach, then organize the information, communication and induction of the various corresponding data.

2 – Reconcile the technicality of data governance and business needs

Data governance aims to know and catalog all the data of an organization, to evaluate and improve its quality and compliance, to make it available to stakeholders.

which guarantee the smooth running of the business. An extremely technical concept: data are the prerogative of the information system, while meeting the needs and challenges of the company is a priority. And it is precisely this need, this concrete use case that must remain the starting point of any project: for example, detecting future customers with an appetite for this or that product, detecting the risks of customer abandonment, etc.

Based on a defined use case, companies will select the most relevant business concepts and data dimensions with data scientists. This “Data Shopping” is achieved first through the corporate glossary (concepts and related elements) then through the data catalog, which is the concrete image of the data in real systems (applications) – and then the data sources for reuse based on the quality, validity, freshness of the data, etc. This technical part is obviously essential, it becomes more easily accessible to all actors through the company glossary.

3 – Model the data life cycle

The data is not static, it has a life. For this, a simple mapping is not enough: data governance requires modeling the entire data life cycle: creation, use, reuse, obsolescence, destruction (GDPR). Goal: To link business process modeling and data governance to save time and understand business issues.

In this context, to accelerate data governance, it will be a question of relying on the business processes already known in the company (data entry and use in the various departments), on the systems that use such data and on the risk management procedures of the company. company (for example the control of personal data).

Therefore, the business process models each of the business actors’ activities in order to concretely conceptualize the data that will then be used in the organization. For example, when setting up a credit offer in a bank, the financial advisor enters the data (CNI, Salary, family situation, medical, etc.) used for the entire process.

4 – Promote quality to improve data

Checking the data also means checking its quality. Because only good quality data at the beginning of the collection process guarantees the quality of the output use case. For example, a credit agreement or insurance cost depends on the data initially collected and used about this or that customer. It is only on this condition that the performance of the service rendered to the customer but also the innovation can be achieved, in the context of an industrialized process (put into production), reproducible and agile.

Data quality is already measured across many customer IT systems, heterogeneously and in silos. Implementing advanced data governance should make it possible to create a repository of control and quality rules. This will make it possible not to duplicate checks, to concentrate all available measures, to complete them and to define priority plans for improving the data.

5 – Facilitate the integration of rules and regulations

Regulatory and regulatory aspects are generally perceived as constraints, generating costs. However, they can also lead players to collaborate with each other, and thus represent opportunities for value creation.

However, new regulatory obligations are constantly emerging, although they are generally complementary to previous ones. With each new development, effective data governance consists in not starting over, but in capitalizing on the controls already in place, in order to identify the only necessary and complementary elements to be integrated for effective implementation.

6 – Putting long-term data governance

New markets, new offers, new automated processes … data, their acquisition, their processing are constantly evolving: if the implementation of data governance is long and complex, it is never finished and must be part of the long time.

As in any project of this type, the first use cases must allow to quickly demonstrate a real efficiency (“quick win”) to start the machine. And it is up to the CDO – through dashboards and indicators – to be able to communicate these results to its community to continue building optimal data governance, to motivate and multiply uses over time.

7 – Install a data sharing culture

In “digital native” companies, such as GAFAM or start-ups, the data culture is innate. Especially since it is generally on the data that the added value of these new leaders is determined and built. In other companies, it’s a whole state of mind that needs to be changed.

Change management is long and complex. It requires a lot of persuasive work on the part of CDOs, who have to rely on successful use cases to create the data sharing reflex (data literacy) and thereby foster innovation and deliver new competitive advantages to the company.

The role of Data Governance is therefore to promote the transformation of the company, its sustainability and its necessary renewal in the face of interruptions and market developments.

En d’autres termes, le plus grand défi des CDOs est d’éveiller les consciences pour que toutes les parties-prenantes se directing ensemble vers the innovation, la création de valeur pour assurer, à terme, la survie et le développement de l ‘agency.

Leave a Comment