Business Intelligence modernization

Business Intelligence (BI) has evolved through a long way. Enterprises have been leveraging analytics to derive insights and to make decisions for ages. In the modern world, the data, the tools and the analytics that are needed to support the business on decision making are continuously evolving. This leaves the business to re-imagine their analytics platform and that demands the IT to find new ways of building solutions. This white paper talks about Embedded Analytics that enterprises can build modern, future-proof analytics solutions.

Embedded Analytics

The era of making decisions based on the insights provided by single tool is almost gone. The business’ problems are much wider and deeper than ever before. The insights are hiding inside variety of data and sources. The analytics tools are also evolving continuously to be able to deal with ever-changing data eco-system. However, there is no one-size-fits-all answers for BI analytics.

One of the Forrester article says that a single, on-premises, enterprise Business Intelligence (BI) platform, based on a single enterprise data warehouse and a single data integration platform, is an unattainable dream – and a thing of the past. And it continues that the customer and market demands require multiple Agile BI options to deliver insights when the business demands them, not when the technology team is ready to deliver.

The era of making decisions based on the insights provided by single tool is almost gone. The business’ problems are much wider and deeper than ever before. The insights are hiding inside variety of data and sources. The analytics tools are also evolving continuously to be able to deal with ever-changing data eco-system. However, there is no one-size-fits-all answers for BI analytics.

My view of Embedded Analytics is the ability to embed right set of BI Analytics tools within business applications, portals or commercial software products, seamlessly, with no or less foot-print of the tools to emphasis on organizations’ process flow and branding. A possible architecture to achieve Embedded Analytics has been discussed in this whitepaper.

Buy or Build

There is an emerging class of embedded-analytics-as-a-service providers that make embedded analysis of customer data a line of business all its own. Choosing from such COTS products or building own application is a decision to be made by individual organizations based on the below key factors:

• Analytical needs
• Size and nature of the organization
• Current and future need of scalability
• Customization required

If COTS tools can provide wide range of adaptability, high degree of customization, faster-deployments and speed-to-market that would be preferable. And, organizations that has well-defined branding through UI/UX standards might need to go for building their own tool.

Architecture

Most of the modern-day BI tools comes with rich set of APIs to provide IT teams the ability to customize the tools to suit their needs. Using the APIs, the tools allows to create and maintain the BI content, customizing the experience and automating the tasks.

Organizations can build ‘Services’ to communicate to such BI tools to interact with the APIs. The communication can be established between the business applications and the BI tools and the response back from BI tools can be used to determine the workflow of the business applications or decision making.

Tools provide APIs that can be integrated using languages like Java, .Net etc. Tools provide such flexibility so that the organizations can choose their own tool sets to suit their IT eco system. Business applications can interact with the BI tools thru the APIs with the help of ‘Services’ which performs the following key activities

• Establish communication between the business application and BI Tools (APIs)
• Performs the operations like create BI content, modify and delete
• Execute the BI content and gets the results back to the business application
• Provides the status of the BI content back to the business application

Each BI tool has its own set of APIs. For Example: IBM Cognos has APIs that can be used in Java or .Net. The APIs can create a report, execute reports, produce output in various formats (html, excel, pdf), modify a report meta (adding a new column etc.), delete a report when it is not required/used etc. Services (micro services based architecture can be adopted) are designed to perform the above mentioned activities. Business applications can invoke the services depends on the need. For ex: When a customer wanted to see the revenue for the current year while he is looking into the revenue targets through his business application, he can click a button to get more details about the revenue for the current year. This action can be used to invoke the service to execute a pre-built report or to create a report on-the-fly based on the current user’s needs, execute the report and do one of the following

• get the actual report back to the application so that the application displays the report
• the service can just bring the data so that the application can display the data in the required format

The latter option provides an abstraction layer between the application and the tool. With the latter option, if the IT wanted to switch to a different tool for various reasons, that can be done with minimal impact to the business application. This ensures the tool-agnostic architecture approach.

If business has variety of reporting needs like, operational reporting, dashboards, in-memory analytics etc., appropriate tool sets can be selected to perform different types of analytics. Each tool can be controlled using their own APIs and also can be abstracted from the business application.

From a user’s standpoint, the business application provides a complete, consistent business experience throughout the business application. User cannot feel the difference in experience while they are in the application or using the BI tool content.

Benefits

• Rich, contextual and consistent user experience for the end user throughout the business application
• Business can influence their customers through their strong branding and that can be compromised by the BI tools
• Employees can make better decisions by embedding analytics insights within the applications they use on a day-to-day basis
• External Users, partners, vendors can get relevant insights that increases the value of relationship
• Customization can help the IT teams to make the right choice in the tool sets/technologies that makes the maintenance easier, quicker and economical
• Tool-agnostic – does not have to depend or limited by on individual tools costs, capabilities, road map etc.

Challenges

• Initial integration of the eco system may be a challenging as there are many moving parts
• Building an architecture that supports the rich analytics may not be a small investment
• Expertise required to build services using APIs of individual tools

Conclusion

The BI landscape is evolving continuously. Business is expecting self-supporting models, quicker turnarounds in building new capabilities and faster time to market. IT team has to choose right architecture to meet business’ expectations. Embedded Analytics model is a solution with flexibility and adaptability and a fit for product managers as the Embedded solution matches both what the business needs and the end user customers need. This architecture can be considered while organizations modernizing their BI eco system.

More articles here

Follow us here