Business Intelligence in the digital era
Jay Shah, ERP Practice Head, Nihilent, provides a background of how BI evolved and what can be expected from a BI solution.
Business Intelligence (BI) plays a critical role in providing key insights into the functioning of the organisation, to help improve decision-making. Many articles, some of them as recent 2013-14, lament the limited adoption of BI by organisations in the African continent. However, the last two years have seen a radical change, with companies across South Africa, Nigeria, Kenya, Tanzania, Congo recognising the benefits of BI and making a beeline for its implementation. Adoption of BI has been witnessed across industries including banking, financial services and insurance (BFSI), retail, media, process as well as discrete manufacturing amongst others, while many others have BI on their roadmap.
For those planning a BI implementation, this article provides a background of how BI evolved and what can be expected from a BI solution, says Jay Shah, ERP Practice Head, Nihilent.
We are well aware of the universal adoption of enterprise resource planning (ERP) systems. An ERP system is a transaction processing system that records business transactions and hence also categorised as OLTP (online transaction processing) systems. OLTP systems, by design, are great at capturing transactions. They are also capable of providing good operational reporting capabilities. However, when it comes to complex analytical reports, the OLTP systems hit a performance snag.
What exactly is the difference between operational reports and analytical reports, one may ask. While operational reports are typically consumed by line of business staff, are tactical, and provide status of operational aspects. Analytical reports are consumed by senior executives and report on key performance indicators thereby providing key insights for analysis and decision-making. Understandably analytical reports require data from several areas, over multiple years to be assimilated and organised to generate powerful insights.
Since the OLTP systems suffer performance issues, when it comes to analytical reporting, enter the online analytical processing (OLAP) systems. OLAP systems extract relevant data from the OLTP systems (such as ERP, CRM, SRM, HR), and re-organise it into formats that is conducive to easy analytics. Thus, these OLAP systems or the BI system, consist of three primary components. First the tool necessary to extract the data from the OLTP systems, also referred to as the source system. This category of tools is called the ETL (extract transform and load) tools. Secondly, we have a store where the data extracted from the source systems is stored and then reorganized into data models for easy analysis. Such a store is referred to as the data warehouse. And finally the third component is the visualisation tool wherein the requisite data is extracted from the data models and presented to the user in the form of visually attractive formats such as graphs, charts etc for facilitating decision making.
Today, visualisation tools have grown in sophistication and provide visually attractive graphical formats such as dashboards, drill down reports, what-if analysis, among others. This has facilitated easier identification of issues and assisted quicker decision-making. It is now possible to represent sales data, cash flow data, among others, in the form of trends. Tools further can extend the trend and incorporate logic/algorithms to arrive at future estimates.
BI systems are designed to be able to consume data from multiple systems such as ERP, CRM, SRM, HR amongst others. With the advent of Big Data, data from social media sites and others can also be fed to the BI systems. This has given rise to the field of data science with which, statistics is used to uncover patterns in such voluminous data, and design algorithms that will enable prediction of future events. A case in point is Amazon proposing you options based on your past transactions. BI systems can thus provide insight into customer behaviour, can analyse purchase spend management, turn data into actionable information for improving process efficiencies, and more. BI / analytical systems, thus, help improve decision-making processes at all levels of management.
The explosive growth of data - big data - fuelled by social media and collaborative tools has presented an opportunity to analyse the large volume of data to improve product specifications, offer better services, reach prospect better and more. Big data needs to be analysed using techniques such as natural language processing (NLP), summarised and fed into the data warehouse for analysing in individually or in conjunction with other corporate data. High performing In-memory devices are helping manage and analyse voluminous data for gains, while the in-memory computing devices help provide analytics in near-real-time. BI systems are now available in the software-as-a-service (SAAS) model, thereby reducing the cost of ownership. Of late, vendors, encouraged by the superior performance of the in-memory computing devices are combining the OLTP and OLAP solutions onto a single platform. With the promise of real-time analytics and prescriptive solutions, digital technology is driving faster realisation of corporate goals.
As visibility into corporate data is possible, BI systems also lend themselves to better governance. Measures and metrics can be defined and set for specific processes and corresponding controls can be setup within the ERP system. BI systems can provide graphical representation of the process vis-`a-vis such controls to enable effective process control, and also provide alerts in case such controls are exceeded.
BI systems are growing in their ability to provide complex analysis, by reporting on the data that exists within your systems as well as big data. BI is able to provide information in visually attractive ways, and moving from being informative to predictive to being prescriptive. But these advances also bring along with itself levels of complexity. This has led to CIOs exploring the option of outsourcing BI operations. The advent of cloud-based BI systems have further supported this process. On the other hand, considerations of confidentiality are delaying such transitions. Nevertheless, this space promises exciting times ahead.