Data Governance

Data governance is the way to make decisions about changes to data, to monitor data quality issues and respond to them and to identify the people to whom issues can be raised and escalated within the organisation.

 

It is a key aspect to Information Management and works in close combination with the other pillars of information management (see the Information Management Blueprint) such as Security, Data Quality, Data Modelling and Structure etc.

 

There are tools that claim to do data governance, and they can help, but the key aspects to data governance are organisational.

 

Benefit Examples

• Faster, better-informed decision making
• Business, Tech & Data closely aligned
• Enterprise wide initiatives can be identified, supported and monitored
• Data issues identified and remedied earlier
• More data joins up across the enterprise
• Data quality can be monitored and improved
• Architecture ‘tax’ reduced
• End to End analysis possible
• Automation becomes feasible
• Governance is cheap but it can be really effective
• Resource no longer spent joining data to support decisions which is then not trusted when the decisions are made.

 

What Data to Choose

As is often the case it is the key data items that need to be governed. There's a cost and overhead to data governance, it takes time to set up and to monitor the data so don't attempt to govern all your data, aim for the top 7-12 data objects e.g. Customer, Product, Sales, Accounts etc. The data structure/modelling - even at a high conceptual level  - will help.

Identify the key data items, the systems that they relate to (shouldn't be hard) and who is responsible for them (business and technical) that cover each of those areas.

Data Governance Roles

 

Decision Makers

 

Data Owners

Data owners have to come from the business; it is their data,  it's their business, they are the ones with real skin in the game.

 

There's a myth that the data owner has to be a single person. In simple organisations this might be the case (and in some small organisations one person might be the rightful owner of more than one pieces of data), however it's more usual in larger organisations that the key data items go across multiple systems and relate to multiple business owners.

Who is the owner of the customer? Sales? Finance? Don't Marketing and Product need to have a say?

The more we integrate across our enterprise and increase our sophisticate we become in our data management  the more we need sophistication in our data governance.

 

The other challenge is that the people who need to have the final say about their data, need to be really quite senior in the organisation. These key pieces of data are central to their business-world and their success so it's in their interest to engage but they have very limited time so we have to have a means to streamline their engagement model.

 

They need Delegates....

 

Delegates

These should also be from the business for the reasons above.

They will be less senior than the actual owners but they need to be able to represent them in the day to day decision making.

They need to have a relationship with the Data Owners such that if they need to escalate an issue they can contact them and raise emergency issues, and can keep them informed of the important issues in the normal cadence of data governance routine.

Their role is to represent the Owners and make decisions about data within their domain, without involving the actual Owners and escalating issues when the need arises.

 

Data Stewards / Data Agents

Data Stewards / Agents can be business or technical.

Their role is to monitor the data, during the course of their normal day's activity or with the help of reports or other monitoring tools.

There are usually one or two people who are at the core of managing the data in any given system. They are often already nurturing the data, they may even have a formal role that partly reflects this. 

These people are ideal Data Stewards and can act as the eyes and ears of the owners and their delegates.

They also usually have a really good understanding of the issues within their system or domain of knowledge along with some pragmatic ways to resolve them.

They should be empowered to make the straightforward data decisions.

These people are the key to making the whole thing work, it's wise to treat them as such.

 

Typical Advisors

 

The business makes the decisions but in order to do so effectively they need to take advice from a number of different quarters. These can vary but here are some typical examples.

 

Architects

Advise on the big picture of systems, strategy, policy, roadmaps etc to help ensure that any decisions made about data fit comfortably within those guardrails and are helping to drive the organisation towards it strategic objectives.

 

Data Architects

Help with the above but focus on the end to end data aspects of the decisions. They also can help to set up the governance process and evolve the governance process.

 

Business Analysts

Advise on process and provide detailed business insight on how things are done currently, any flaws and foresee the impact of changes on the immediate eco-system.

 

Data Stewards / Data Agents

As above, while these roles make day to day decisions about data they also act as the eyes and ears of the owners and their delegates and advise them on a course of action.

 

Project / Initiative Team

No organisation exists statically there are often many initiatives going on at any one time, and as an organisation becomes more of a coordinated enterprise then more of these initiatives cross system and data lines. Having representation of these relevant projects avoids duplication or at worst contradictory decisions being made about data.

 

SMEs

These can provide specialist advice about a technology, a part of the business, an industry standard or a myriad of other elements that may be relevant to a given decision.

 

Technical Delivery Team

These can provide practical insight into the technical options to support a solution,  the existing backlog and resourcing. They can provide designs and deliver them and help avoid delivery pitfalls.