Trade Lifecycle Management Using TIBCO Event Fabric Platform
Sudhir Gupta
Director, Banking and Financial Practice Cognizant Technology Solutions
Deb Ghosh
Principal Architect, TIBCO Quantum Leap
Agenda
Introducing the TIBCO Event Fabric Predictive Trade Lifecycle Management
Growth of Trade Volumes and the Need to Reduce Cost Why do Trades Fail and Cost of a Fail Fail Management Process Trade Lifecycle Management using TIBCO Event Fabric Platform
Solution Architecture Key Differentiators
Introducing the TIBCO Event Fabric
The TIBCO Event Fabric Platform An executive summary
Decision problems are everywhere
Manage risk including Compliance and Regulatory requirements Perform multi-product real-time integrated Fraud/Financial Crime management Monetize the huge un-tapped opportunity in Inbound Customer Interaction by presenting the Next Best Move associated with the specific customer in Real time.
Classic decision architectures however are sub-optimal in the ability to react in (near) real-time and handle the associated challenges of scale and availability TIBCO Event Fabric uses a unique high speed Event Correlation technology that
Significantly reduce latency and helps operationalize decision platforms in real-time Scales and performs significantly better than classical technology Non-intrusive, operated by business user, extremely quick time to production
Proven to have significant impact on Revenue growth, churn control, all risk based losses and revenue leakage etc
Whats the Problem in Current Decision Architectures?
Silo Oriented evolution most infrastructure has grown in (silos)
Multiple application/ multiple channels Multiple reference masters Things get exponentially more complex with M&A is added to the mix
Huge Data Latency also called the Knowledge-Data divide
Front-end systems (channel systems) are rich in transactional capability, but poor in end-end customer knowledge and context around the specific customer interactions Back-end Data-warehouse/Business Intelligence tools have the exactly opposite problem
Closed Loop there is no easy way to take a full circle from the discovery to action Time to production a new insight takes a long time to put to production
Business Situation: The Timing of the Opportunity
Early Warning Signs
Response Latency
TIBCO has perfected a technology that is in its 3rd generation now to solve this (and other similar) problems. The technology is called
TIBCO Event Fabric Platform
Taps events from any source within the enterprise (and extended eco-system) Correlates (real-time) events with historical information to infer and predict next best move next best offer/next best service/risk level to a trade The platform is closed looped with identification of the next best opportunity (Business Situational Analytics) and closing it with full life cycle being managed Some of the biggest banks in the world have applied this platform capabilities to a variety of use cases like Predictive Next Best Offer the holy grail of Organic customer acquisition, retention and cross-sell/up-sell
TIBCO Event Fabric Architecture
Txn System Txn System Txn System Txn System
Classic Fabric
Current Generation
Traditional Use-Cases
Account Opening Channel Integration
Transaction
Business Process Management Enterprise Service Bus/SOA Manual/Point-point
First Generation (Not part of current discussion)
Operations
Non Intrusive Events
Reference Masters Analytics DWH/BI
TIBCO Event Fabric High performance Event Correlation Operational Data - Memory optimized images Business Situational Analysis Rules & Inference
Newly leveraged Use Cases
Predictive Cross-Sell/Up-sell Multi-channel Customer Experience Fraud/Compliance/AML and Ops Risk Platform Dynamic Offer Management
Predictive Trade Lifecycle Management
Growth of Trade Volumes
Ever increasing trade volumes have increased the stress on trade and settlement systems that were already under duress due to factors like
Increasing settlement attributes, new regulated exchanges, ECNs, ATSs etc
Value of Transactions cleared & Settled at DTCC 2000 1500 in $ 1000 trillion 500 0 01 02 03 04 05 06 07 Year Avg Daily Volume at National Securities Clearing Corporation
60 50
Increase in cross border trades has added to the complexity of trade and settlement cycle
Possibility of these trades failing is higher due to higher number of intermediaries and additional parameters like time zones and exchange rates Higher chance of parties having incompatible views on the terms of a trade
in m illion
40 30 20 10 0 01 02 03 04 05 06 07
Year
Source: DTCC
Cross border holdings US - Rest of the world
8 7 6 5 in $ trillion 4 3 2 1 0 Dec- 01 Dec- 06
Foreign stocks & bonds held by US Citizens US stocks & bonds held by rest of the world
Source: DTCC website
Need to Reduce Costs
Trading bonds and equities for customers is now a commodity business for Investments Banks, hence it has become imperative for them to cut costs in processing trades Estimates suggest more than a quarter of Wall Street institutions have a trade fail rate between15-20% Managing the cost of trade processing by effective fails management can be an important lever in controlling costs
Why do Trades Fail?
Typical reasons for trade failures
Lack of shares with counterparty Miscommunication Economic issues parties disagree on traded values Disagreement on delivery mode FOP vs. DVP Operational issues
Typical large Investment bank - Equity trading for North America
Value of trades that get cancelled Value of trades that settle post value date Value of trades that settle by the value date Weekly trading value 0 5 10 15 20 25 30 35 40 45
Weekly trading numbers
in $ billion
Cost of a Fail
Seller loses the time value of the invoice price over the interval of the fail A fail exposes both the buyer and the seller to replacement cost risk
The buyer faces the risk that the seller becomes insolvent before settlement and that the price of the security increases prior to the sellers insolvency The seller faces the risk that the buyer becomes insolvent and that the price of the security declines prior to the buyers insolvency
Prospect of a fail drive the participants to devote resources to monitoring and controlling of counterparty exposure and could, in an extreme case, lead them to limit their secondary-market trading
Fails Management Process How it works?
Trade and settlements systems provide mechanisms to operations team for monitoring trades These systems also provide the operations team with relevant data to analyze the reasons of failure and to guide them through next steps Failures are addressed by a combination of technology, manual logistics and smart-sourcing initiatives Steps to resolve an issue will vary depending on the nature and complexity of the fail Resolving a complex case is a multi-step process spanning multiple channels of communication to match the information and settle the trade
Fails Management Process What are the Challenges?
Nuances in Fails Management process are driven by a number of factors
A landscape of large number of disparate systems and the integration requirements among these systems Involvement of multiple parties across locations and geographies Trades are initiated through multiple channels, not always possible to validate the trade data with the counterparty
Some of these nuances are more pronounced in case of cross border trades
As number intermediaries involved increases so does the probability of incompatible views on assumptions and practices involving the settlement mechanism and timing Differences can be on applicable costs, taxes, risk appropriation etc
What-If?
We can detect trades that are at the highest risk of failure and as they move in the life-cycle we can continuously and in real-time predict the probability of a break BEFORE it actually breaks?
We introduce a unique solution platform to address that goal a Predictive Trade Lifecycle Management Platform powered by TIBCO Event Fabric
What TIBCO Event Fabric Platform can offer?
TIBCO Event Fabric platform based solution can address the Fails Management problem
Event probes bring in events from the trading ecosystem (trading systems, reference stores, financial markets) and are represented as a asynchronous event cloud from which IT systems have to filter out relevant information Complex Event Processing capabilities of TIBCO Business Events can be used to co-relate asynchronous events that may help predict trade failure before the trade actually fails Business Process Management capabilities allow easy follow up of the targeted cases from the Predictive platform significantly improving the accuracy of trade processing
How Complex Event Processing Works
Identify parameters on which trades need to be monitored
This is achieved by rule based filtering Rules are based on historical trends
Trades thus identified are instantiated in memory - they become part of the state machine
Trades in state machine/ memory are monitored This would be a a subset of all trades
Another independent event occurs somewhere in the ecosystem
It gets fed into TIBCO Event Fabric Platform Gets co-related with one or more trades in state machine, it is identified as a potential fail and is forwarded to the operations team for follow up
Complex Event Processing in Fails Management
Fails Management is an ideal candidate for Complex Events Processing (CEP)
Number of different parameters can impact the outcome of a trade These parameters can be dependent on number of asynchronous events Traditional BPM solutions can not be used to co-relate these asynchronous events by simply defining rules CEP solution can be used to monitor trades so as to ascertain impact of various events and their relation with various trade parameters
BPM in Fails Management
Fails Management provides an ideal landscape for a BPM solution
Improves the turn around time and accuracy Provides a clear visibility of fails issues, with granularity on fails reason to enable root cause analysis Highlights ownership of issues and provides detailed audit history on exceptions Provides automated complex allocation rules for routing exceptions by region and business group enforcing ownership Provides automated exception and trade prioritization based on transaction characteristics and timing
Typical Fails Management Scenario
Traditional flow vs. CEP solution based on TIBCO Event Fabric Platform
Solution Architecture
Conceptual Architecture
Situation Designer Exception Resolver Visual Analytics
State Machine Reasoning (Temporal/Spatial) Correlation Engine
Event Grid
Ref Data
Historical
Support (Network, DB etc)
STP (Existing)
Solution Architecture
Key Differentiators
Traditional solutions in this space are reactive with latency sometimes having a significant down-side on bottom-line
TIBCO Event Fabric Platform based solution is a pro-active PREDICTIVE solution Actions initiated in advance can alter the outcome for better having
The granularity of treatment is event and not just data This solution provides the ability to co-relate asynchronous events with historical data in real-time to predict likely events
Ordinary rules engine falls short
The platform provides powerful state machines out of the box
Provides the ability to monitor at-risk trades continuously assessing the impact of events in the ecosystem
Extremely scalable architecture
Not database-centric, uses distributed cache Runs on distributed blade class platforms
Thank You
