This is part 6 of a multi-part series on the Analytics Operating Model.
In our recent blog, Data Oriented Architecture: Laying the Right Foundation, we examined the challenges of selecting an enterprise platform for big data and approaches for overcoming those challenges. As we continue this series, we dive deeper to explore each element of a sound analytics capability in an era where data is plentiful and computing is powerful, but results have not yet lived up to the hype.
We live and work in an age where every device and sensor can generate and transmit data. Companies must develop capabilities that allow them to identify and manage the right data resources. A key driver of success is an organization’s ability to manage data in a way that maximizes value through analytics data management practices. High quality, accurate data is crucial to a successful algorithm and ensures that data-driven decisions are based on facts. Further, companies must also consider how this data fits into the broader organizational value chain in which they are operating to fully extract and understand the value stored in their data.
The ability to successfully deliver analytics solutions is dependent on an organization’s ability to develop an understanding of how different data sets contextually fit into the organization by managing data as an asset and establishing an enterprise data model.
Managing Data as an Asset
To begin to understand what it means to manage data as an asset, its often best to approach the idea using traditional asset management concepts. An asset infers value and good asset management would establish a comprehensive system to protect the value of that asset. In today’s world where petabytes of data are commonplace, an organization must identify its high value data assets and input a framework to proactively manage, develop quality practices, and secure this intangible resource. Further, it requires that clear accountabilities and ownership are defined and applied for the stewardship of an organization’s data assets. Activities of this nature are best addressed by a Data Governance Framework that includes participation from the parts of the organization. These participants are either creators or consumers of the high value data assets, and they must be represented and supported by multiple layers of the organization, as depicted in Figure 1.
Figure 1: High value data assets should be supported by multiple layers of the organization
In the era of big data, the traditional approach to defining data governance priorities no longer applies. To be effective in the big data world, data governance must pivot from the idea that all data is fit to be governed, to data that has the most potential to provide value to the organization should be governed. As an example, organizations often have multiple data sets that represent the same data, which leaves teams to determine which data set is the correct source and who is accountable for the integrity of true data source. The approach to identifying data focus areas for governance must become more agile and support the development of analytic solutions that will improve business outcomes.
Establishing an Enterprise Data Model
As analytics drives the need to find and test connections between different data sets, it has become vital for organizations understand how information is organized, structured, and transformed as it moves though it’s lifecycle. Often, in large organizations the understanding of the way data is structured varies from one group to the next. In organizations where there is a lack of a well understood, common data model, it can sound like you are speaking two different languages when discussing the same data.
“…Where there is a lack of a well understood, common, data model,
it can sound like you are speaking two different languages
when discussing the same data..”
This can lead to disparities in understanding and misalignment in approaches to solve business challenges. Working to define and refine a model that represents the high value information within an organization, as well as how it is related and connected, creates a tool that can help drive alignment to a common understanding of enterprise data. This tool enables teams to come together with a common understanding of how data transforms to enable successful analytic solutions.
A key challenge to building an enterprise data catalogue for an analytics solution is storing and structuring data in an intuitive way. Building and organizing this catalogue to allow users to find and leverage data, while looking for ways to improve the organization through data-driven decision making is paramount to building cost effective analytics solutions. By leveraging the enterprise data model and data platform, data engineering teams can drive best practices in storage and catalogues for big data and data lake platforms. By utilizing an enterprise data model that is already well understood, teams can help avoid data becoming lost in big data platforms and prevent the enterprise data lake from becoming the proverbial data swamp.
Organizations that will succeed in developing an enterprise data model and managing data as an asset will be more successful out of the gate when establishing, deploying, and maturing their analytics and big data capabilities.
Click here to learn more about our Data & Analytics practice. Want to continue the conversation? Contact us at firstname.lastname@example.org, and follow the rest of this series, as we dig deeper into each element of the Analytics Operating Model.