Once upon a time there was a lady (let’s call her Alice) who was very practical that she has her fixed deposit – FD savings in tact in the bank. The bank is very happy with her savings as it would give the bank the assurance and increase in their wealth management for greener investments.
The bank is equipped with a high-end infrastructure in analytics and sophisticated business intelligence (BI) system giving their business analysts and frontline business professionals flexibility in their quantitative analysis, trending, and forecasting.
The BI infrastructure of the bank is still using the traditional approach. Incremental and delta data changes will be extracted by the ETL, that is, Extract, Transform & Load the data to their target analytics engine. Or the approach can be ELT, that is, Extract, Load, & Transform the data to their target analytics engine. Once all the transformations are complete, the data that are once OLTP in nature now becomes OLAP and is ready for analysis and ready for usage of the bank’s downstream systems.
The service level agreement (SLA) of the bank’s IT department to the business is 12 hours. The extraction of data up to the completion of all the transformation jobs needs to complete at 6AM everyday. Previously, the SLA was 15 hours since the decision systems environment is free by 3PM. But due to ever changing demands of business analytics, the SLA is now reduced to 12 hours and the IT department is having tough time coping up because their analytics database engine is starting to suffer from performance degradation.
The bank’s business analysts does collect their data everyday but their method of analysis based on the business process is within 1 week worth of data. There wasn’t a near real time business process to analyze the behavior of account holders.
Alice’s FD account has been very good for the last 2 years. She belongs to the demographics of single, female, high income professional, very good standing on accounts, and within the age of 25 to 30.
However changes happened in Alice’s life. In less than 1 week, Alice performed the following transactions:
- Withdrawal from FD accounts
- Withdrawal from Savings accounts
- Withdrawal from Current accounts
- Applied for a personal loan
After the above withdrawals (Events 1 to 3), her demographics has actually changed to single, female, high income professional, poor standing on accounts (risk), and within the age of 25 to 30.
When Alice applied for the loan (Event 4), she was surprised that her loan was granted. But of course, she was happy.
Why did the bank approved Alice’s loan?
Simply because when the loan officer saw Alice’s account holder profile, she still belongs to the very good accounts. Though there were withdrawals from Alice’s accounts, the loan approver thinks that Alice may have some investments in the bank that is not visible to the loan officer due to the bank’s security policies.
The loan officer doesn’t know that the demographics profile analysis wasn’t updated yet.
Clearly based on the above scenario, Alice is under the risk of defaulting her payment to her loan and the bank is also under the risk for non payments.
If all the four complex events that Alice did were captured in real time and handled by various business processes or rules, the bank would have the chance to acknowledged the risk of lending the loan to Alice to avoid financial losses.
Alice story is a normal customer behavior.
However, if there are one million bank customer who performed the complex events that Alice did without the bank realizing it, it is not surprising why banks suffered financial losses and can probably attributed to the recent financial downturn that happened globally. The Great Depression of the Financial markets.
After the business realizes their gaps, financial institutions have demanded for real time analysis of events.
Previously, the OLAP is sitting in the datawarehouse and is only accessed and analyzed once a week or everyday for operational reporting, the business demands have changed that they need the events to be processed and analyzed as soon as it happened and entered in the company’s analytics system.
Analytics has evolved. As always.
Complex events processing or greatly known as CEP systems is the analysis of different levels of events that correlates to the business rules or processes.
As soon as the events have entered into the CEP stream, it is analyzed immediately based on the rules thus triggering the different actions as provided by the business logics. It provides advancement and flexibility to the company’s frontline force – business analyst and business support team.
What’s the value of having a CEP system?
(1) It provides an enhanced risk management system. The earlier alert of Alice’s complex event 4 would trigger the bank that her demographics has changed thus the loan should be disapproved.
(2) Fraud detection. When I was in Oman, I purchased the jewelry set and used my Singapore credit card in payment. In less than 5 minutes while I was still in the jewelry shop, the bank called me to verify the validity of the transaction. As CEP systems analyzed the near real time events even in seconds, considering that I was overseas and with few delays due to the network speed, etc, the complex event that I triggered changed my behavior pattern thus it is suspect for fraud detection as I have never been in middle east let alone use my credit in middle east. The good thing about it is that if it was a true fraud, the jewelry shop can hold the person who is misusing the credit card since the bank will reject the transaction. Thanks to CEP systems.
What happens to traditional OLAP models?
Well, the usage of ROLAP, MOLAP, HOLAP, etc is still going to be used by the business for their trending, slicing and other analysis for better business positioning as required by the C-levels of the company. However, the CEP systems is catered for frontline force and business support team for real time business activity monitoring.
Who offers CEP systems in the market?
There are large enterprise vendors offering CEP solutions. Below is the list to name a few and this list doesn’t follow any ranking:
So, next time when a bank calls you to verify your transaction, don’t get annoyed especially if you’re not in the mood. You have to thank your bank that it is investing in this type of technology to better service you as a valued customer. (Wink).
Good morning America. Photo taken in New York, USA. 2005.
Til next time. Au revoir.