So you thought the beginning and end of BI is the setting up of a cutting edge Data Warehouse? Well, think again. The BI landscape has undergone a sea change. Trying to articulate its various aspects and how it is evolving would fill several books. In this article, we will skim the surface and talk about the various ways in which BI is being perceived and handled in the current IT landscape.
If Traditional BI = Data Warehouse + Reporting Tools
Then
BI 2.0 = (Past + Present) Data/Analytics + Future Business Analytics + Reporting Tools
I have used this simple equation by way of demonstration: BI v2.0 is all about studying and analyzing Past, Present and Future trends using archived and real-time data feeds and then turning this data into knowledge for stronger decision-making.
It is real-time or it is of no real use:
When it comes to Traditional BI tools, the data retrieved from them, in today's parlance, is considered outdated. Intelligence is gleaned after the fact-after significant time has elapsed between the actual activity and the time when it is analyzed in either an individual or aggregated fashion. In this Internet Age, even the lapse of a day could render that data obsolete.
BAM and CEP are real-time tools; data is analyzed as the transactions are being executed. It is this optimally effective decision-making in the moment that makes BI 2.0 so attractive to new age businesses.
Use Case:
Let us now look at a common use case: the Shopping Cart. Shopping on the web is one of the most common online transactional activities today, and it is increasing year over year. For high traffic sites like eBay and Amazon, business cannot be made truly agile without the use of real-time analytics. Traditional BI just will not suffice.
So let us consider the following example:
• Acme Mart is a medium-sized discount store with brick and mortar stores throughout the country.
• It also sells merchandise through its high traffic website acmediscountstore.com.
• The transaction volumes both at the stores and online is quite large-to the extent of several million per day.
• It has a mature, established SOA infrastructure and its entire order management process is service-enabled.
Acme Mart also uses BAM (Business Activity Monitoring) to monitor services. Custom dashboards have been built to monitor sales of specific item categories and even items through sophisticated dashboards.
Contrast this with a similar business that uses Traditional BI mechanisms. It is obvious that it will not be able to match the level of agility of Acme Mart. By the time the sales data is analyzed, the opportunity to respond proactively has passed. Acme Mart is clearly more capable of reducing costs and increasing revenues via its real-time BAM-enabled BI solutions.
While CEP (Complex Event Processing) is similar to BAM, it is more of a services agnostic technology. It is sort of a real-time message aggregator that can process messages from various different sources including web services. One example of this can be applied to the current use case, along with order information, if related stock price information for the companies that produce those products are also needed in a separate dashboard then this feed that would normally come from the likes of a Bloomberg or Reuters can be combined with this Order data and displayed.
BI Technology Trends: The Past
We need not go into the details of how BI functions were being performed or are still being approached using DW-driven technology. If Business needed insights into key trends then they would follow this well beaten path:
- Setup a de-normalized DW or Data Mart
- ETL information from transactional data stores into DW
- Use reporting tools like Cognos to generate the required MIS reports.
Therefore, this was the tried and trusted method and was considered business as usual. Traditional BI is very batch-oriented in nature and not real-time. However, while real-time mechanisms may look very compelling in terms of their value proposition and ROI, it may not be feasible to implement these quickly as they are dependent on a services-based infrastructure.
BI Technology Trends: The Present
As the competitive landscape becomes more intense, the need for real-time business intelligence is almost De Rigueur in certain types of industries, in order to gain business agility and competitive advantage. With the advent of technologies such as BAM and CEP, it is possible to perform real-time and even future what if business analytics.
BAM and CEP fall under a new category called Event Driven Architecture (EDA). These tools permit BI in both synchronous and asynchronous modes.
Business Activity Monitoring (BAM) is associated mainly with Web Services. It is a dashboard driven toolset that is comprised of various design time components and a runtime component. It is used in conjunction with other runtime components of a typical SOA implementation, such as ESBs and BPEL engines.
CEP (Complex Event Processing) is similar to BAM with the difference that CEP is not tied to web services alone and is message agnostic. You can aggregate disparate messages from myriad sources and use them as part of one CEP application.CEP is usually event driven.
The online shopping segment is also benefiting a lot from these advanced BI services paradigms. The likes of Amazon, eBay and Zappos are rejoicing at the arrival of BAM and CEP. In such high traffic sites where the hits are measured in millions per day, attaining the level of agility and edge over the competition is virtually impossible without real-time analytics. Traditional BI strategies fall woefully short and are not equipped to deal with these types of business operations.
BI Technology Trends: The Future
There are several game changing trends that are in vogue in the BI field that promise to change the face of the industry. While BAM and CEP are the pre-cursors, they are fueling other paradigms that were not possible before.
Here is an example that is taken from advertising industry pundits whose central focus is always on "The Consumer": Smart Billboards that will glean biometric information based on facial features of anyone walking by. If the consumer turns out to be a young 20-something year old male, the back end system would beam back relevant information such as ads for an iPod, cell phone, energy drink, sports car or anything that fits this particular demographic.
Just 5 years ago, this idea would have seemed outlandish, but not anymore. With real-time BI delivery mechanisms at the foundation, scenarios such as this are becoming reality. With a little ingenuity and BI 2.0, the possibilities are seemingly endless.
