What aspects provide a holistic view of data?

The information environment comprises Data, Processes, People, organizations, and technology associated with it. A good operating model covers all these aspects to achieve successful maturity levels.

The first question you would want to ask - "Are the processes, data, people, organizations, and technology associated with a division understood and documented?"

Enterprises often fail when they take a big bang approach to one or more of the governance services including Quality, Architecture, Stewardship and data management. The key is to have a strong business case from the divisions that would allow them to orchestrate these services. If it is a data quality assessment, then, a business case on the issues, needs, costs and value from discovering and fixing these issues is the need of the hour.

How do you define success of a data governance service?

The marketing division of an investment bank has data quality issues that are affecting the Campaign Management value chain. The sales and marketing team recognizes well, where they would have to collate the data issues – from the ETL rejects, error logs, service issues, adhoc fixes, support tickets, reporting errors, historically reported data issues from individual platforms etc.

As guided by the data governance division, the marketing division has placed a strong business case with the needs and the value from leveraging data quality assessment. But to understand the traceability from data quality assessment across technology levers to business levers and business value is not quite simple.

Though the division held some documentation including DFDs, Data Maps that could envisage the data flow and lineage, they cover for only 50% of the analysis. With the growth of channels, social media and internally generated information, the division finds it difficult to produce documentation on how, when and from where data is being acquired, updated and stored in the customer master. The division also lacks information of how this data is being applied across various enterprise business processes along with absence of agreements on data quality.

Without these definitions, the division has embarked on using the data quality services and hit a roadblock at a certain point with not being able to reap the ROI.

What has gone wrong in this scenario? Even with a mature operating model, lean processes and data governance solution, organizations are not able to fully leverage the costs and effort invested into implementing these solutions.

How to break the data Silos?

Though, Divisions hold the accountability of the data for the ease of managing it, data has always been an enterprise asset and is never division specific. Data from one division can be leveraged equally by the services of other divisions, depending on their evolving needs. The fundamental fact of looking at data to be very specific to the division is the roadblock in taking the practice forward. Data is affected by people, process and systems equally. A good operating model would envisage these aspects of understanding "how data is affected by various people (roles), processes (functions) and systems (application services)".

Does your organization promote analysis of such kind that fuels laying the right roadmap to data management and governance?

What should be the Road Ahead?

It can be quite challenging to document all the shared process and shared data across the enterprise. Several techniques can be used to classify information that needs to be prioritized, analyzed and documented.

• Enterprise classification and Data Tiering are used to identify, analyze and document the data that directly relates to business value. This is something similar to Process Value Stream Mapping to identify activities that directly relate to value creation.

• If embarking on Divisional Data quality assessment, the information that confines to the data quality issues alone should be analyzed for usage across Business processes and functions within.

It is always advisable to go with any of the above approaches. Further, have the templates and standards of documenting these findings incorporated into the organizational process assets.

Data quality assessment in investment banking - Sales and Marketing division doesn't necessarily mean that customer information is planned for and created in a "Campaign Process" but is also further updated and applied by "Account and Fund setup", "Transfer agency" processes.

While business processes across enterprise share same information, they are in fact sharing the quality problems equally. Poor quality data impacts all of these services or one key business process central to the organization. You would be able to guide your data governance division to success if you answer -

• Which business processes are impacted?
• Who are the people or organizations involved?
• Which data domains are impacted?
• Where does the data reside (Applications, Platforms, Databases involved)?
• Are there other systems associated with the same issue?

It always adds to the maturity to improve and evolve the approaches to Governance within the Data Governance Office. In fact, it would be great to see your organizational approaches in action.