Business Intelligence: Current state & road ahead
Business Intelligence (BI): Its components and features
Business intelligence is a category of applications around all enterprise data involving tools and technologies combined with processes and procedures for gathering and storing in a persistent form and presenting back to business users using data visualization techniques thereby empowering them to business insights and in making better business decisions.
The figure above represents typical BI application architecture. It involves a broader spectrum of components and features as mentioned below:
Data Integration Tools: Integrate data from disparate sources into persistent data stores. Extract, transform and load (ETL)/capture, transform and flow (CTF), enterprise application integration (EAI), data quality improvement/enrichment and master data management (MDM) are some of the primary processes these set of tools are equipped with.
Presentation servers and management systems: Store data in a persistent form to facilitate downstream data presentation layers. ODS, data marts and data warehouse constitute this category of components. From an implementation perspective, this can involve database systems that are relational (RDBMS), multidimensional (physical cube) or column in nature along with DBA tools for managing them.
BI Applications: Provide actionable insights and knowledge discovery of the business, facilitate business decision making process and helps to monitor, manage, and control the business performance. The application generally covers the following sub-components:
Tools metadata component: An abstraction layer used by BI tools for mapping physical data models to the exposed business interpretable components for querying, reporting and analytics.
Data visualization component: Data presentation in the form of hierarchical organization of metrics, key performance indicators (KPIs), dashboards and scorecards as a rich internet application (RIA) application.
Reports, query authoring and publishing: Data presentation in the form of reports with drill-down, drill-up, drill-across and drill-through options. Strategic and tactical querying capabilities including more tactical ad hoc query generations primarily for power users. It also includes scheduling, report bursting and delivery as e-mails (with URL links and
Data/text mining components: This goes beyond the end-user driven knowledge discovery into machine learning techniques, statistical and other approaches to gain valuable insights into data. Hence, various mining techniques and models form part of the subcomponent for supervised or unsupervised learning, classification and predictive modeling.
Alerts and notifications: In case of exceptional business events that can be configurable, the users can be proactively notified of such occurrences in various formats to end-users’ devices.
Data & reports exports: Reports data be made available in various formats to the information consumer in the form of text, csv, pdf dumps for use by downstream systems.
Data governance processes: Is associated with best practices and quality control for managing, monitoring, enriching, consumption and protecting organizational information in an efficient and effective manner.
The benefits of business intelligence investment in a fluctuating economy
In the current economic downturn businesses tend to upheaval all difficulties around budgetary constraints causing BI investment seemingly out of reach for IT manager and CIO. However, there is still optimism for both categories of enterprises those that already have a matured BI infrastructure (generally large and some mid-size companies), and those that have negligible or no BI in place (small and medium-sized organizations).
For organizations having done significant investment in BI, survey results suggest that BI is actually helping them to stay afloat by uncovering the areas of inefficiencies in the business operations. As BI lays the foundation for strategies to unravel knowledge discovery, these organizations are made aware of business operations and market performance enabling them to navigate the current standoff of economic landscape more effectively than some others. BI is actually helping them to quickly identify, manage and control newer sets of KPIs to make business decisions in the changed scenario that would not have been possible without BI scorecards and dashboards.
Organizations without BI can still choose a lightweight BI solution that is fast to deploy and easy to use and leverages existing business data and IT investment. Another alternative is to go for Software as a Service (SaaS) for operational intelligence solutions as SaaS model comes with low-cost, low-risk and quick to deploy as compared to the traditional BI model. Again, with either alternative, BI can help to quickly and effectively identify trends, relationships between initiatives and results, and the performance drivers that will help capitalize in a market where less informed competitors will falter.
The road ahead for BI success
We are now at a stage where many really understand the need to have BI in their organizations. Hence, all reports by the research analysts suggest that the awareness and adoption of BI among enterprises are definitely on the rise and are going to grow manifold in the days to come. As a result of this, BI has been the key focus of most of the independent software vendors (ISVs). There will be two pronged approaches of BI maturity in the road ahead, those that are attributed to improvement/enhancements in the BI tools/technology space and increased customized industry specific BI packaged solutions. These would translate to potentially all of the following in the days to come:
- BI tools have started to offer RIA for data visualization implementing flash (Adobe Flex platform) based UI plug-ins into the browser.
- Data visualization capabilities of these tools will further mature with mashup UI content, RSS feeds as part of BI 2.0 (Web 2.0/3.0) and richer animated graphs/charts that roll with time.
- With HTML 5.0, it is expected to even get rid of the portlets or pluggable UIs while displaying in the web portal.
- Another area the BI tools have to address is the performance, scalability and availability of the BI infrastructure and flexibility in extending solutions.
- Though some of the BI tools already have support for advanced analytics and data/text mining, there will be more coverage of mining and machine learning techniques built into the tools. In addition, the implementations of the mining models are bound to become simpler to assist researchers build, test, validate, deploy, score and fine-tune the models.
- Common warehouse metamodel (CWM) by OMG (Object Management Group) will be adopted by more vendors to enable reuse of warehouse and BI metadata between different vendor tools, platforms and repositories. Similarly, some of the other initiatives like Predictive Model Markup Language (PMML) will also be widely adopted to develop mining models in one vendor’s application and use in another.
- In the database space, newer features/enhancements will continue to emerge in terms of data organization/placement, optimizer features, indexing, caching and performance in different types of databases including RDBMS, column and MOLAP (multidimensional OLAP) cubes to support VLDB systems and BI. At the same time, the stage is all set for data warehouse/database appliances to compete with the rest on overall performance in handling petabytes of data and MPP configurations.
- ETL and CTF tools will see a merger with EAI components especially for real-time data warehousing needs and research and vendor tools will continue to mature with stream ETL preferences over conventional batch and micro-batch (MBETL). In addition, tools working closely with ETL and related to MDM and data quality management will also see more maturity.
- Technologies such as RFID and near field communication (NFC) will stretch the BI capabilities to accommodate the smart devices supporting them. This also includes maturity of mobile BI.
- There will be an increasing demand for customized and industry specific BI solutions. Though there are BI applications in the areas of customer Intelligence, Business Performance Management, Supply Chain Intelligence and Marketing Analytics, to name some that are held by larger ISVs, there will be more of startups contending in this space especially to meet the demands of mid- and small-sized organizations.
- Cloud computing and on-demand services have been a buzzword today. It is a natural progression for BI to embrace the cloud. Hence, it is bound to see more of BI supporting virtualization and dynamic provisioning without compromise on scalability in the days to come.
There are many stakeholders of BI such as — enterprises that are implementing BI solutions, ISVs building tools, vendors providing packaged BI solutions and the vendors provisioning the BI services/solutions in cloud infrastructure. For each of them the priorities are different and hence call for maturity and research areas beyond the scope of this discussion. For example, security is among the utmost significance for SaaS providers, real-time BI for certain businesses, to name some.
- How Digital Platforms Are Helping Real Estate Business
- How Artificial Intelligence Is Changing Diwali Shopping
- Delving Into The ABC Of Cyber Security
- Is Artificial Intelligence The Future Of ERP Software Optimization?
- Weekly Rewind: Top 10 Stories On CXOToday (Oct 9-13)
- IoT Knowledge Gaps Exist In Consumer Product Industry
- HR Managers See CRM As An Effective Business Tool
- Why Companies Need To Invest In Chief AI Officer
- How AI Is Fueling Customer Experience Strategies
- Cloud Computing Driving Innovation For Next Generation