How Orange and Vodafone are approaching data governance

Data is one of the cornerstones of digital platform companies, and although communications service providers (CSPs) may not need to imitate the most data-driven companies such as Amazon, they will nonetheless increasingly use data to drive growth. And as they do so data governance will increase in importance.

For our report The growing importance of data governance we asked CSPs about the main objectives of their governance projects. Creating useable data for the business was, unsurprisingly, most important – especially with the ambitions of CSPs to increase their use of machine learning and automation as both require clean data sets and processes to perform well. Simplification of data is also key, along with improving accountability.

We also took a closer look at the approach Orange and Vodafone are taking.

Orange’s group-level governance priorities

Orange operates a group-level governance team which coordinates with data protection officers across operating companies. This team is working on several data governance priorities:

1. Eliminating data silos

The most important focus area in recent years has been on improving data quality and access to data stored across the multiple countries and at group level. The move to Google Cloud Platform is allowing Orange to break down silos by putting in place new data management tools to improve visibility of data availability/format as well as new access management systems.

2. Improving data quality

Orange is focusing on a better understanding of lineage, particularly around the training of models when several are manipulating the same training set, using governance and data catalog tools.

3. Enhancing data protection and sovereignty

The need to meet specific regulatory requirements, beyond GDPR (General Data Protection Regulation), within individual Orange operating companies and additional customer specific requirements in the B2B space, require Orange to create a complex framework specifying data residency requirements, on premises vs. public cloud and anonymization. Part of the solution for this complexity is new technology – for example, the use of machine learning to find personally identifiable information at scale and speed in order to anonymize and reverse encryption as needed.

4. Setting policies & principles

Orange has a set of defined “personal data principles” validated by its executive committee which the company uses to enable consistent policy decisions and creation of guidelines across operating companies. Orange is currently in the process of translating these policies into scripts within their data cloud – for access and identity management through to tooling – in order to improve compliance and move away from requiring employees to read policy documents. Executives note that this coding of policy will significantly improve compliance by removing human error, while making it easier to detect intrusions.

Vodafone’s seven pillars

UK-based Vodafone Group described seven “pillars” the company uses across its operating companies to execute its data governance program:

1. Clear definition of roles and responsibilities – ensuring ownership and accountability of data sources and the use of a business glossary so that all data users understand the data they are using. This is considered one of the most important and immediate needs.
2. Management and clear governance of data as an asset – including processes to improve quality, put data lineage in place and increase the ability for users to share data and insight about it.
3. Definition of governance levels – data classification, allowing a clear set of data access rules to be created and applied to the different teams across Vodafone. For example, financial data is defined as high risk and high value, so access is limited with various approvals required.
4. Technical policies and procedure – allowing architectural decisions around how the data is accessed, archived and deleted. For example, a taskforce was set up to assess the raw data on platforms around Vodafone, including all unstructured data (manual files and spreadsheets). A clear operating model was then created around teams across the organization (including big data, group level, technology, finance, marketing, operations and commercial management).
5. Upkeep of a data catalog – showing data source, dictionary, metadata, quality and how data is used; this promotes transparency and allows the organization to make best use of its data.
6. Data governance tools and KPIs – development of supporting capabilities such as KPIs to measure data quality, compliance and availability of data to different teams, which includes installation of systems to enable this support.
7. Deployment capabilities – development of working methods such as clear work streams that then link to a roadmap aligned with the CEO’s priorities and strategies for the year

Read the full report The growing importance of data governance to find out more.