Business Intelligence (BI) or Big Data Analytics initiatives, you are perhaps looking for answers to some important questions in an effort to drive positive outcomes in the new year. What has been your Return on Investment (ROI) on BI and Big Data investments? What about Dark Data? Have you done enough to throw light on Dark Data so your colleagues can make better decisions? Is there a common understanding of dimensions and metrics across the organizations ? More importantly - How do you get everyone on the same page?">
As the year ends, it is often the time for retrospection. If you are leading Business Intelligence (BI) or Big Data Analytics initiatives, you are perhaps looking for answers to some important questions in an effort to drive positive outcomes in the new year. What has been your Return on Investment (ROI) on BI and Big Data investments? What about Dark Data? Have you done enough to throw light on Dark Data so your colleagues can make better decisions? Is there a common understanding of dimensions and metrics across the organizations ? More importantly - How do you get everyone on the same page?
BI tools and data sources have grown dramatically in every organizations. Data Visualization tools like Tableau, Birst, Power BI, Qlik etc. have found home in IT landscape of many big and small companies. This happens because of departmental decisions or a result of Mergers & Acquisitions. This also creates business silos that create inefficiencies and inconsistent decision across the enterprise.
Collecting and cataloging metadata is the first step towards harnessing the benefits from your BI investments. Many companies today understand the importance of managing a single metadata repository or data dictionary so that decisions within their organization become reliable and consistent. Investments in Master Data Management (MDM), Master Data Governance (MDG) etc. are made with the right mindset – cataloging the data elements will perhaps help getting everyone on the same page. However, the enforcement is arbitrary and is often left to Business Analysts or end-users. This creates its own problems as it is subject to various interpretations. Let’s explore how this situation can be managed effectively.
Enterprise BI Governance ensures that decisions in the entire company are rooted in one central definition of data, data authorizations and data usage. With proper implementation of Enterprise BI Governance, it should be possible to view, add or modify the data definition. But more importantly, to see all the benefits, centrally managed data dictionary should be accessible for self governance at the time of consumption . So how do we get there ?
Enterprise BI Governance starts with the creation of common data definition in a data dictionary. When a common data dictionary exists, there is no ambiguity on how a dimension like “Supplier” or “Item” is defined. There is no ambiguity on which data element represents Revenue KPI for example. Creation of authorization rules for data security are part of this phase as well. Creating the current baseline for BI landscape is an additional part of this phase so that incremental changes on the data dictionary can be made in the future as new systems are added and systems are consolidated.
Once the data dictionary, authorization rules etc. are created and centrally managed, it is important to keep this updated to reflect the changes in data definitions and usage. Searching and updating data elements in data dictionary is necessary for someone like a Data Steward or a Data Architect who is responsible for maintaining a coherent definition of all data elements.
Many tools facilitate creation and maintenance of data dictionary at the schema or data model level, but lack the ability to enforce this at the time of execution i.e when reports are consumed. Modern visualization tools allow business analysts to create their own calculated KPI on the fly. For example, Revenue KPI can be represented in a report as Sales*AveragePrice. This may not be the true definition of revenue as maintained in data dictionary. Any decision based on this KPI will be incorrect. The right enforcement of BI Governance will catch such discrepancies instantly at the time of data consumption by the end-user. This also creates an environment where BI Governance can be self-service, which further brings discipline and a culture of “trust but verify”.
What are the benefits of Enterprise wide Governance for BI and why should companies give higher priority to BI Governance Initiatives? There is an old saying – “A stich in time saves nine”. Companies have realized that decisions are inconsistent/incorrect if they lack a common definition of data. Incomplete and inconsistent decisions results in financial loss. Companies spend considerable amount of time and budget on data harmonization, data security and data privacy concerns. Having a strong governance model will help reduce these costs. Additionally, companies invest in multiple tools and licenses are underutilized. The bleeding continues in many such areas and companies are looking at ways to stop this and maximize on their existing investments.
Building an enterprise wide BI governance is step one. Enforcing the BI Governance at the time of data consumption is what is going to drive real ROI. When BI Governance initiatives are enforced by connecting data insights with the centrally managed data dictionary, you begin to bring consistency in decisions. Finally - you can get everyone on the same page with a common understanding of data.
How have you benefited from such initiatives? When business users in your company are analyzing data in reports, do they have instant contextual access to the data definition maintained in the central data dictionary ? Love to hear your thoughts and ideas on this.