smart data webinar - primaticsfinancialgo.primaticsfinancial.com/rs/primatics/images/smart... ·...
TRANSCRIPT
Smart Data Webinar
Please tweet @Primatics during the
webinar using #SmartData.
Presenters:
John Lankenau- VP, Product Management,
Primatics Financial
Craig Lovell- VP, Software Development,
Primatics Financial
Banks Face Demanding Environment
Rapid changes
Tighter regulation
More complexity
Increased demands for information
2
More demands combined with today’s powerful technology means
trouble !
Many Loan Processes Are Conceptually Straightforward
• General reserve
Impairment
Stress testing
Non-accrual
3
Conceptually simple, but operationally hard. Why? !
Data!
• 27 of 28 surveyed last year said they were not satisfied with data and
data management
• 23 of 24 surveyed this year said they were not satisfied with data and
data management
• February 2015 RMA Journal survey
• Majority of U.S. institutions surveyed disagree with statement that risk data is
“clean, comprehensive, accessible and timely”
• Highest level of concern about “difficult integration among systems”
• High level of concern about “lack of integrated risk and finance reporting”
4
Despite heavy investment, data remain a challenge nearly everywhere !
Symptoms of Data Challenges that Need to be Addressed:
Inability to do desired analysis because data architecture does not support
it
Many sources for reports – inconsistent answers depending on whom you
ask
Emailing, phone calls or other manual movement of data from one function
or process to another
Multiple sources for the same data elements require significant
reconciliation efforts or cause inconsistencies
Calculations require data from unwieldy number of sources
Making data “application-ready” happens in many places
Etc…
5
Data challenges lead to inefficiency, lack of ability to do desired tasks,
lack of control !
Handy Guide to Troubled Data Projects
Never-ending
• Two years into a scheduled two-year project with little
benefit and 18 months remaining to finish
Non-adoption
• Project is done but no processes actually use it
Attic
• Data are all in one place but processes use different data
for the same thing
6
Having a successful data project requires addressing root cause of
challenges !
1
2
3
Data Challenge is Foundational with Multi-Faceted Issues
Data challenge starts at the front end of all processes
(sourcing)
• Reconciliation, timing differences, “bad” data, etc…
• Identification and segregation important parts of process
Data and applications must be synchronized
• “Application-ready” data must be accessible to applications
• All applications must link to reporting data mart in native way
Back end reporting also a data challenge
• Aggregation
• Multiple views and ad hoc analysis need to be supported
7
Addressing the data challenge successfully requires addressing
sourcing, applications, and back-end reporting holistically !
Sourcing
Reporting
Applications
Elements of a Comprehensive Data Architecture
4/13/2015 8
Integration of multiple upstream data
sources
Integration of human data
Data validation, normalization, enrichment,
and population identification
Integration with Calculation Engines
Warehouse Design and Population
Reports, Analytics, and Extracts
2
1
3
4
5
6
Pitfall #1: Conflating Operational Data with Reporting
Warehouses
4/13/2015 9
Upstream Data Sources
Reports and
Extracts
The Mother
of All
Warehouse
Concepts
• Operational Data Store (ODS) design is
completely different than a Reporting Warehouse
design
• An ODS will have multiple data sources and
multiple applications integrated
• A modern reporting warehouse is designed
and optimized for read-only reports
• The warehouse should not have to share
resources with an active ODS
• For the best dining experience, cook your data in
the kitchen (Operational Data Store) and eat it in
the dining room (Reporting Warehouse)
Validation,
Normalization,
Transformation,
etc.
Pitfall #2: Failing to Normalize
4/13/2015 10
Upstream Data Sources
• Bringing upstream data into an ODS
without normalizing means the
applications will never see a consistent
view of the data across sources
• Applications and calculation engines
should not have to deal with multiple data
formats
• Reporting across data sources becomes
extremely difficult if not impossible
• Your ODS should not be a dumping
ground for disparate formats
• A loan is a loan is a loan – represent it
that way no matter where it’s from
Operational
Data Store
(ODS)
Without
Normalization
Applications /
Calculation
Engines
?
?
?
Ugh!
Pitfall #3: Don’t Forget the Humans!
4/13/2015 11
Upstream Data Sources
• Your data architecture must integrate both
upstream systems and data from human
processes
• Spreadsheets abound – bring them into
the system
• Humans should be able to reconcile,
correct, and augment data in a controlled
and auditable fashion
• Humans need to be able to stop the train
and re-load data if they spot problems
Operational
Data Store
(ODS)
But wait, I didn’t give
you my charge-offs!
Applications /
Calculation
Engines
Pitfall #4: If You Build It, They Will Come
4/13/2015 12
• Very nice architecture for
consolidating and reporting, but no
applications integrated!
• Applications left to fend for
themselves
• Input to each application needs
to be copied from somewhere
else and transformed
• Reconciliation nightmare
• Applications have to provide
their own reporting of outputs
• Can happen when IT throws a data
warehousing party and doesn’t invite
the business
Operational
Data Store
(ODS)
Data Warehouse Reports and
Extracts
Upstream Data Sources
Applications /
Calculation
Engines
Something’s
missing
here…
Putting it All Together: Smart Data Architecture
4/13/2015 13
Integration of multiple upstream data
sources
Integration of human data
Data validation, normalization, enrichment,
and population identification
Integration with Calculation Engines
Warehouse Design and Population
Reports, Analytics, and Extracts
Operational
Data Store
(ODS)
Data Warehouse Reports and
Extracts
2
Upstream Data Sources
1
1
3
2
3
4
4 5
5
6 6
Business
Rule
Engine
Applications /
Calculation
Engines
Data User
Interface
4/13/2015 14
15
Lo
an
Data
E
nd
User
Ris
k a
nd
Fin
an
ce
Fu
ncti
on
s
Servicing
Systems
Spread-
sheets /
DB’s
People
Source Data
EVOLV Integrated Risk and Finance Solutions
1. Reserving (FAS 5, FAS 114, TDRs, Journals, CECL Prep)
2. Loan Accounting (F91, Non-accrual, SBO, PCI, Journals)
3. Stress Testing (Model Execution, Aggregation, Templates)
4. Reporting (Credit Quality, SEC, Operational, Regulatory)
5. Custom Applications (Build on the EVOLV Platform)
Finance
Perspective
Regulatory
Perspective
Risk
Perspective
Operational
Perspective
Enterprise
Warehouse
Client GL
SaaS-Based Infrastructure
EVOLV SMART Data Solution
I
m
p
o
r
t
E
x
p
o
r
t
Integration Layer
• Data Rules Engine - API
• Integration, Normalization &
Validation
• Event Identification / Tagging
Loan Data Marts / BI
• OLAP - Star Schema
• Multiple Basis / Perspectives
• Risk and Finance in One Spot
• History / “Runs” Maintained
Platform Engines and Integration
User Interface Layer
Event Engine Workflow Engine Calculation Engines Reporting Engine Sub-ledger Engine
API Layer / Developer Toolkits
“The EVOLV’d State”
16
One Platform
Unified Source Data
Integrated Reporting Data Mart
Integrated Applications
The EVOLV Difference
17
Q&A Session
Contact Us
If you would like to hear more,
please reach out to us!
703.342.0040