Getty Images


How self-service BI capabilities improve data use

Organizations can be more efficient problem solvers and enable users with self-service BI capabilities that bring more data and tools to their fingertips.

Insight-driven businesses can use data more effectively than their competitors when they've operationalized the data and enabled self-service business intelligence capabilities.

While self-service reporting and analytics are not new concepts, much has changed, ranging from the types of data available to the nature of self-service itself.

"What changed is moving from descriptive analytics to predictive and prescriptive analytics, powered by advanced analytics, data science and machine learning," said Dan Simion, VP of AI and analytics at global professional services company Capgemini. "What needs to happen moving forward is trying to be as nimble as possible when self-service reporting happens, so organizations can react to business needs as quickly as possible."

An important advancement is the ability to interact with data as opposed to passively consuming a report or a dashboard. Instead of pointing and clicking, users can utilize the natural language querying capabilities resident in augmented analytics platforms, without knowing a query language like SQL.

Better still, the answers to their questions often include automatically generated data visualizations and narrations explaining what the data visualizations mean, which helps facilitate a common understanding about what the data is saying.

Meanwhile, data engineering teams have been building data pipelines that ensure the right data gets to the right place at the right time for contextual decision-making, while including important enterprise guardrails such as data governance and compliance.

Augmented analytics platforms have also enabled self-service data preparation, enabling users to choose their data sources and combine them as necessary to answer the query. It's also possible to automatically generate insights or dynamic data stories based on a person's context, such as their role and the past queries they've run.

The leaders in the space -- Microsoft Power BI, Qlik, Sisense, Tableau, ThoughtSpot and others -- all provide augmented capabilities so organizations can democratize analytics among a wider group of users.

How to enable self-service

There are two general approaches to enabling self-service BI capabilities. One is to make a departmental tool buy and, with IT's help, get it up and running.

The other approach is to have the data experts -- data scientists, data analysts and data engineers -- provide the masses with augmented analytics capabilities. The benefit of this approach is the visibility into how people are using analytics in the organization and mitigating data-related risks such as using personally identifiable information in a noncompliant way.

Large organizations tend to have a center of excellence (COE) complemented by data scientists and data analysts who are assigned to specific departments or lines of business. That way, the organization can ensure best practices across the organization while catering to the specific needs of operating units.

Meanwhile, as companies continue to move deeper into the cloud, they're running analytics and machine learning to enhance their self-service capabilities there as well.

"For example, COEs can stand up dedicated discovery environments for each business group to enable end-to-end discovery of BI capabilities or data analysis," said Sree Majji, senior vice president at digital engineering professional services company Apexon (recently merged with Infostretch). "Also, purpose-built platforms can be instantiated for [a] fully governed versus partially governed BI environment."

The democratization of analytics enables business users to get answers to their questions faster than relying on IT. It also enables data scientists and data analysts to focus on the harder problems that require specialized expertise.

Leading vendors' approach to self-service

Sisense promotes the hub-and-spoke model in which experts from the data team enable citizen data scientists as described above. That way, baked-in policies and best practices are built into the company's data culture by virtue of having a center of excellence.

At the edge -- in business units or departments -- there are specialists, such as data scientists, and analysts who have domain knowledge. This approach strikes a balance between the business's desire to optimize BI within operating units and the need for controls and standards such as governance. Specifically, this platform provides several features including the following: 

  • visual-based data modeling and dashboard creation;
  • interactive visualizations with advanced filtering and drill pathways;
  • AI-driven data investigation; and
  • integration with workflows and business applications.

The Microsoft Power BI team promotes managed self-service which, like the hub-and-spoke approach, divides responsibilities, albeit somewhat differently. In this model, IT retains data ownership and the business units can run their own reports. The approach recognizes the dual-speed nature of modern organizations in which business professionals are moving quickly and IT is moving at a slower space.

Like Sisense, Power BI provides a drag-and-drop interface and AI assistance for those who are not data experts. Enterprise BI capabilities allow data scientists and data analysts to do more sophisticated kinds of analysis. It also provides enterprise-grade controls. Customers can combine self-service and enterprise BI for business agility and safer data use.

Qlik also balances the needs of business and IT with its platform, Qlik Sense. The company outlines five essential self-service BI capabilities, which of course its platform delivers. They are the following:

  • balancing the need for data governance with timely and accurate decision-making;
  • easily integrating data sources;
  • the ability to quickly share insights with stakeholders;
  • the ability to create apps and reports on demand; and
  • providing mobile access to analytics. 

Truth be told, augmented analytics platforms are more alike than different for competitive reasons. Often the choice between or among platforms boils down to the comfort level with the individual vendors.

Good BI architectures minimize manual intervention, secure data at rest and in motion, and reduce overhead with data movement.
Sree MajjiApexon

Start with the business case

Ease of use drives a lot of BI platform decision-making because businesses want to empower more employees with data-driven insights. When operating units purchase their own analytics platforms, they're usually solving a specific problem. Over time the company may end up with multiple platforms or multiple instances of the same platform.

Other considerations should include how well the platform scales, how it fits into the company's current and planned architecture, and whether the platform has adequate security, governance and compliance controls. The major platforms offer these capabilities and are in many ways providing similar capabilities, so the choice of vendors may depend on the buyer's relative comfort with the vendor's team.

Cost and IT impacts are other considerations

Understanding the cost of ownership and architectural impacts for any BI solution is key for business leaders. Total cost of ownership includes software licensing hardware costs, development cost and the cost of maintenance. This can run five to 10 times more than the amount of the initial vendor sales pitch, Majji said.

"Good BI architectures minimize manual intervention, secure data at rest and in motion, and reduce overhead with data movement," Majji said. "Business leaders should be cognizant of these implications before acquiring any BI solution on their own."

If BI tools already exist in an organization, stay current with the changing landscape and have a technology evaluation framework to periodically assess the tools a company already has versus what's available, recommended Majji. That includes identifying gaps in the current tools based on the desired business capabilities. All tools should have a divest, maintain, or invest assessment and evaluation technology framework.

"If the tools are flexible enough and can effectively resolve new challenges on the business side, then they are good enough to remain in place," said Simion. "Do the tools accommodate real-time or near-real-time data or not? Examples like that can help organizations evaluate where they stand and make those tough decisions [about] whether to make a change."

Dig Deeper on Business intelligence technology

Data Management
Content Management