Succeeding with Data Governance – Comparing models of Data Governance

Part 1 of a multipart series 

Many companies struggle with implementing successful data governance programs. The key to success is understanding where you are, documenting where you want to be, identifying the gaps and steps required to achieve your ultimate goal. This multi-part series is designed to provide a practical guide to executing a best in class governance program. Our first post examines different governance models that will allow an organization to baseline and chart a rewarding path forward.    

Building an effective data government / governance structure requires answers to three key foundational questions:

(1)  What kind of data regime (E.g. democracy) do we want in the future

(2)  Who will be in charge of the effort, and

(3)  Baselining the type of data regime we have in the organization today

First, we must baseline the types of data regimes and the definition of data democracy. In our view, there are four types of data regimes:

Data Anarchy

Data Anarchy is most prevalent in organizations who have refused to acknowledge the value of data. Such organizations often have very little in terms of data platforms or AI systems. They often lack any systemic or automated security for data. Such organizations often use Excel, Power Point, and email as their tools for data access and information dissemination. The friction of data access is very high, the security of data due to the chaos of anarchy is the lowest in such organizations. They also score very high in Data poverty and lack of data security

A Data Monarchy or Data Dictatorship  

An organization where Data is tightly controlled, often indicates a system controlled by a Data Monarch or Data Feudal Lord or Data Dictator. Data Dictators are often managing extremely tight and often fragile controls due to issues with the presence, provenance, and security of data under their control. Access to data is often centrally controlled in a data-vault or data-keep and access if often is based on political or bureaucratic position within the organization. The maturity of data security is often low. The security keys to the data kingdom, for certain lines of business (LOBs) reside with one or two individuals. Data poverty is very high in this model while, the value expressed from data is often untapped in such organizations.

Data Oligarchy or Data Technocracy

Information is concentrated in the hands of one or more Oligarchs or Technocrats who can exploit it for their own benefit. Technocrats or oligarchs often succeed by providing tremendous value to the organization thereby subverting the influence & control of data monarchs or data dictators. While data is monetized in pockets in this model, most of the organization is often still under the thumb of a data monarch who sparingly provides data for community use.  The maturity of data security is often low. The security keys to the data kingdom, for certain lines of business (LOBs) reside with a handful of individuals.  The disparity of data access is high although some units under oligarchs and technocrats have found a way to access data and express value express value from it. Data power is concentrated in the hands of oligarchs and Technocrats leading to a higher median access to data although the disparity of data access is often highest compounded by higher data security risks in this model.

Data Democracy or a Data Marketplace

A mature data organization is able to create a data marketplace where the business community has timely, secure and equitable access to data. Business Users in the lines of business are empowered to “own” the data bound by security parameters that empower by providing the security of automated compliance. In this model, data poverty is lowest, data disparity is lowest, data security is very high and the value from data is often richest in such organizations. Data democracies are prevalent in data educated organizations and lead to data driven cultures.

The following is our summary comparing different data regimes 

The next question – who will oversee the effort – seems relatively straightforward until you look at roles and responsibilities. Ideally, the organization will have a Chief Data Offer (CDO) and that individual will be responsible for the Data Governance and Democratization effort. The CDO role is complex and must be viewed from multiple lenses:

1. Chief Data & Analytics Officer: Oversight of data management, data science, and analytics
2. Data Entrepreneur: Monetize data either by selling it directly, using it in data- and analytics-based products and services
3. Data Developer: Lead AI/data driven development of key applications or infrastructural capabilities
4. Data Defender: Lead data safety initiatives and manage regulator relationships
5. Data Architect: Drive data modernization and AI based approaches to data integration
6. Data Governor: Establish data governance and manage sub domain data verticals supporting business lines
7. Data Ethicist: Leads safeguards, ethics of data management and oversees who controls data

It is imperative that an organization understand the role of a CDO. The CDO role is beyond the scope of this article but I highly encourage you to read the following article regarding CDOs for valuable insights inti the CDO role.   https://www.dhirubhai.net/pulse/how-many-cdos-does-corporation-need-goutham-belliappa/

The key to establishing a go forward vision is to understand where one is on the scale.  Once we understand where we are and where we want to go – we need to address fundamental questions or concerns that will arise. Specifically:

• Our business partners will drown in data if I give them a data democracy. 
• Our business partners have varying skills and capabilities. To be successful, individuals will need retooling which will be expensive and time consuming. 

As with any democracy, freedom is not absolute. For example, I cannot legally drive 100 MPH on most highways. My freedom is balanced by the overall welfare of society. It is the same concept with a true data democracy. Acknowledge and build for this scenario.

Lastly, we must recognize that our data democracy is only stable and will thrive if our business partners have trust in the data.  In a data driven culture, trust can be assigned a numerical score. It is unbiased. For our business to trust that data, we must exhibit the following attributes:

• Accuracy
• Integrity
• Completeness
• Uniqueness
• Validity 
• Timeliness

Our data democracy exists on trust. If we do not trust the data, enforcement of a data framework (meta data management, data governance, etc.) and changing our mindset to an outcome oriented organization will become impossible. Organizations without a high level of data trust will have the title of a data democracy but in reality they will be one of the other three data governments. Remember, most individuals never question the hue of the green line, they question the direction of the line. Questioning the direction of the line is fundamentally doubting the integrity of the data. Trust is everything.  

Building a true governance / functioning data democracy is a multi-step process.  

Please contact us if you have any thoughts or questions. The second part of the article will dive into – different models and techniques for achieving a Data Democracy and Data Marketplace.

Goutham Belliappa: VP for Cloud Data and AI Engineering – Capgemini 

Scott Siegel: Head of Data Governance and Data Strategy – Capgemini