Resolving the tug-of-war between self-service and governed analytics-Part 2

September 14, 2021 | 3 minute read
Text Size 100%:

By guest author, Dan Vlamis-President, Vlamis Solutions

In the early, wild west days of self-service BI, the message about the benefits of centralization got lost with the zeal to “give users what they want”. Companies are now realizing the vital role that corporate IT, data governance, and centralized analytics repositories with proper metadata can play in analytics systems. For example, Oracle Analytics can define a centralized metadata repository database. This centralized repository is accessed by all users and can be enhanced (or ignored entirely) by users with user-created data sets.

In Part 1 of this blog, I reviewed the pros and cons of governed vs. self-service analytics and the tug-of-war between the two approaches. In Part 2, I am going to show you how it does not have to be a zero-sum game. You can get the best of both worlds.

The Dilemma of the Ad Hoc Approach

 Vlamis often helps clients on both sides of this tug-of-war. A recently published case study about Vlamis client Certegy demonstrates the dilemma faced by organizations that embrace the ad-hoc approach without incorporating the structure of governed analytics.

Certegy built its original Oracle Analytics Cloud (OAC) implementation mostly using data sets and data flows. This implementation did not take advantage of the RPD, modeling, or subject areas. Data sets and data flows enable a quick, simple implementation, but they lack the robustness of an analytic data model and have size limitations in the amount of data they can return.

Vlamis introduced a governed analytics infrastructure at Certegy that worked in concert with the existing ad hoc approach to achieve the following benefits:

  • Certegy can now easily share common calculations across the organization.
  • Future maintenance costs will be reduced significantly.
  • Analytics displays and filter lists return in seconds.
  • DBAs will make fewer calls to the analytics team to report problems with long-running queries.

Creating a Win-Win Situation

Based on our experience with Certegy and other clients, here are some ideas for balancing both approaches to data:

  • Prototype using user-created data sets. Let users develop the visualizations and analyses that bring them the most benefit using their own data sets, and then turn those over to IT for performance optimization.
  • Provide flexible analysis tools. By allowing users to manipulate data the way they want, you reduce their need to export the data. This enables IT to provide flexible access without the inherent security problems of exporting data.
  • Certify external data sets. Oracle has added the ability for corporate IT to “bless” external data sets. Users can bring their data into the fold for personal use, but once Oracle Analytics administrators certify it, it is discoverable when searching in that Oracle Analytics instance.
  • Add security to data set access. You can specify who (by role for instance) can access these data sets. You can take an external subject area, certify it, and say everyone with an “analyst” role can see and use it. We discuss some of these security topics in a prior blog article.
  • Establish a rubric for identifying the best candidates for the repository. Make it easier to identify whether something should be centralized or remain in the self-service realm. If a given report, analysis, or visualization has only one user, self-service BI is the way to go. As the use of it becomes more common with more users, move it into the IT-controlled repository.

The key to achieving the maximum benefit is in finding balance. Know when data needs to be governed more closely and when it can remain self-service. This is not a zero-sum game, where one method is right and the other is wrong. In the pursuit of the best, most complete analyses, both approaches should be brought to bear.

Looking Ahead

I expect future developments will make it easier than ever for the governed and self-service analytics approaches to peaceably coexist. Users can create and share their own multi-table data sets from uploaded data or web-accessible data. Combine this capability with coming developments such as a friendly web-based UI to edit the metadata repository and the lines between corporate governed analytics and self-service analytics become blurred. Organizations will control what data is governed via granting rights to groups of users rather than access to specific metadata tools. Future competition between governed and self-service analytics will become more of an act of policy than access to specific tools.

Learn more about the Oracle analytics platform. To learn more about Vlamis and how they can help with Oracle Analytics projects check here. Follow us on Twitter@OracleAnalytics and LinkedIn.


Dan Vlamis

Previous Post

8 Procurement analytics use cases to meet the moment

Roxanne Bradley | 11 min read

Next Post

Oracle Analytics Cloud on Gen 2: Features and Benefits

Oracle Analytics | 4 min read