how i managed to Develop a Analytics story for services about 4 years back. Contains
Maturity Model, Business Potential, Services Structures Areas that analytics can be applied to
20150108 create time stamp
Analytics growing as a business mandate.
Data is Growing Performance Gap Widens Capability Gap Exists..
4.4x
2.7x
2.4x
2.4x
2x
Investment in Data and
Analytics
Top Performer Bottom Performer
Sources: IBM Breakaway Now with Business
Analytics and Optimization
17%
42%
28%
10%
USE OF DATA BY BUSINESS*
75% or more 50-74%
25-49% 0-24%
++ There is a skill gap
60% executives say they “have more
information than we can effectively
use”** [IBM Report] .
McKinsey Report on Big Data estimates
50-60% gap in the supply of deep
analytical talent; equaling 140,000 to
190,000 unfilled positions.
40% growth in global data
annually
Globally 2.5 quintillion bytes of
data per day
90 % of the data in the world
today has been created in the last
two years alone.
Customer Transactions
Customer records through device
ubiquity and better data mgmt..
1
Customer Interactions
Social Unstructure, semantics..
20B events / Day – Facebook
2
Machine Interactions
Logs sensors intelligence on all
equipment
3
IBM Report Global Business Analytics
market size is pegged around $105 billion
and growing at CAGR 8%.
Shifting Priorities for
Management in Analytics..
Potential for applying Analytics to Business
Based on areas explored with verticals.. During BPVM
ThemesFinance &
Accounting
GRC
CRM
Service&
Warranty
Vertical
Solutions
Worldwide
financial services
OpRisk and GRC
technology market
will grow to $2
billion by 2013 at a
compound annual
growth rate of
6.5%.
The global
financial data
analytics market
size has been
potentially
estimated at $5
billion
The global
warranty
management
technology market
will represent
more than $1.1bn
in 2012, compared
to $715m in 2007
Worldwide CRM Applications Market
Forecast to Reach $18.2 Billion in
2011, Up 11% from 2010
In 10 years,
leveraging big data
in the health
industry could
capture $300
billion annually.
Potential increase
in retailers’ OM
from big data
could be 60%
High
%-age of spend directed
towards Analytics
Sources:
1 - Prithvijit Roy: New financial analytics hub;
2 - Chartis Research;
3 – IDC; 4 – Datamonitor;
5 - McKinsey BigData report,
1
2 3 4
5
Low
Value Chain of Analytics in Business.
CRITICAL
BUSINESS
KPIs
DATA
MANAGEMEN
T
PROCESS
CHANGES
Strategic
Themes
Volume,
Variety,
Velocity
Actions from
Insights /
Foresights
Business Analytics
VISUALIZATIO
N
Real time / In
Process
ANALYTICS
APPLICATION
S
Insights &
Foresights
Analytics Value to Business
Business
outcome
Operations
Transformation
Insights Data
•Customer Insight
•Digital Marketing
•Pricing / Risk
•Product Design
•Service / Operations
•BI / Dashboards
•Manual Operations
•Embedded Analytics
•CEP / Rules Engines
•RT Integration
•Analysis / Methods
•Prediction / Data Mining
•Machine Learning
•Sample vs Large Data
•Parameterized and NON
•Data Sources { External,
Unstructured }
•Data Integration {ETL}
•Data Lineage {Metadata}
•Data Preparation {Index,
Search}
•Customer Segmentation,
Behavior based models in
all industry
•Price Sensitivity analysis
•NPD / Molecule research
in Pharma
•Risk in BFSI
•Driving Digital Initiatives
like Mobile
•Triaging / Routing in
Contact centers
•Running a Analytics KPO
that provides insights for
Operations
•Methods like
Segmentation, Regression
based scoring,
• Sensitivity Scenarios ,
What-if
•Text and media mining
capabilities [ PCA ]
•Semantic Search
•70% of the effort is spelt
out in Data
•External sources, public
and paid..
•Text, media processing /
Index
Analytics Services Maturity Model
ALIGNED INTEGRATED OPTIMIZINGFRAGMENTED
DATAANALYTICSVISUALIZATIONPROCESS[ACTIONS]
SCALE / STRUCTURE
SOURCE / RETRIEVE
CONFIG - CONTROL
INTERACTION
ALGORITHM
MODELING
DESIGN
EXECUTE
MANAGE
PRESENTATION
STRUCTURE
Simple 2-Dimensional Graphs and
reports including Types of Visuals
supported?
Static simple play out
Simple structure, numeric [ cardinal]
and non-numeric- [ Ordinal]
Internal Local Files, federated
Ad-hoc Customer opportunity
Operational Changes >
Basic Functions and statistics
User Configuration, Data Security
Structured Data with metadata
support,
Integrated data sets through DB-
DWH, SQL based retrieve
Single Iteration playout
Computational Flows
Process Maps, Kpi- Metrics
Breakdowns,
Manual Process Change / Actions
Tactical Changes – re-structure to
Business operations, processes..
Linear Functions, Regression,
Statistics,
Strategy Changes - New services
models, synthesis of business value
Integrated Partner Actions,
Automation into systems,
scenario analysis, what -if analysis,
Complex Statistics [econometrics] ,
Numerical Method, Clustering
Analysis,
System Generation-Automation ,
visual re-formation,
Compliance and traceability effort in
adding new data sources
external connectors – API,
Composite Visuals, infographics
Unstructured text, Data Scale – Size
and time
Value Chain Analysis , Benchmark
Data
New Revenue Models
Sense and response mechanisms,
Simulation, optimization,
Text & Analytics, Neural Networks,
fractals,
Actions integration - external
systems.
Storyboards, Virtual Reality
late binding – auto discovery of
structure
Access to non standard data, late
structure binding
Real time search
Data as Media like Voice, Image and
Video Bigdata Management
pivot based interaction – User self
service
Maps, Multi-dimensional Graphs,
How are Businesses acquiring Analytics
Inhouse /
Captive
Solution
Utilities
Services /
Resources
Platforms /
Tools
1. A Typical Bank would have a 1Bn USD budget
2. 80% spend inhouse and in Captive
3. 1200 Person = 600Mn $ Value / 100 Mn $ Cost
4. Slow, lethargy, internal Constraints, IPR
1. Small Boutique companies getting seeded
2. Focusing on either large platforms [ splunk ] or a
very specific Business use Case [ Mydrive ]
3. Scale issues, pricing,
1. Large resource houses, with 80% $ from staff Aug
2. Fragmented delivery, water fall, change is a
challenge , Utilisation is key , security & leakage
3. Can Scale, some can partner,
1. Best complement to Inhouse / Captive
2. Developing the foundations for the next gen,
3. Focused more on tech rather than business
4. Partner to all above entities,
Value Proposition for the Data Science Organization
Building &
Maintaining a Core
Data Platform for
Analytics: that
includes setting up of
specialized data
marts (for pricing,
reserving, etc.),
identifying internal
and external data
sources, building
connectors,
integrating with
internal core
insurance systems
and the like.
Assisting in Effort Intensive, Repetitive Non-Core
Analytical Activities that allow the client’s core
analytical team to concentrate on modeling thus
increasing core analytical bandwidth. Some
activities that vendor could take over include:
Data Cleaning
Data Aggregation and Transformation
Creating Transformed Variables
Assisting in creating transformed variables
Model Validation
Checking model accuracy
Recalibrating models and reporting results
Integration of Analytics with
Business:
Reporting Services
Integration of Results into
Core Systems
Business Process Integration
Building “Analytics as a
Service” Platform
Flexibility and Cost
Optimization with “Lab
0n Hire” Service Model
Trained Data Scientists
Onsite-Offshore model
for cost optimization
Licensing and Tool Costs
spread across multiple
projects
Multiple pricing options
including utility-based
models
1
2
3 4
Delivering Analytics Value to Business
Business
outcome
Operations
Transformation
Insights Data
SolutionsservicesToolsPlatforms
300 400 7000
wipro
Other players CTS, TCS, Big 4, musigma
TeraData
Pivotal
Opera
Cloudera
Tableau
Clikview
RevoR
Mydrive InfoChimp
70 1200 500Bank captive
Typical Analytics Practice
Strategic Eco-system Alliances
1051
Analytics [ 140 – 60 USD ]
BI [ 100 - 40 USD ]
Data / Integration [ 100 – 30 USD]
1. 80% of the business is still Staff
Augmentation
2. 80% of the business in BI / MI and
low end data services..
3. Large players like Wipro / TCS /
MuSigma in the range of 5000-
10000 resources
4. Lot of SME consulting Smaller
players
5. Clients are slower than the vendor..
1. Staff Augmentation in various Skill Areas
2. Partnering and COE development for clients
3. Project based Delivery – Agile Waterfall
4. Embedded Analytics in Operations and other initiatives
like Digital, mobile etc..
5. Service Transformational Analytics – CTS
6. Very weak in industry / Business domain
Industry Trend Past and Future
• Rapid directionless ops growth –
has helped ISV [+30% CAGR ]
• Bringing structured data together
• Now looking for Show and Tell + 0
consulting + More Action
• Shifting Operations to Offshore –
Captives
• Partnerships, COE, Investments,
Utilities = Value Add
• BI Sophistication has kept managers
in charm
• Integrated solutions with Digital
Initiatives
• Large Data Initiatives – Lakes,
Metadata, External Data
• IOT / more sensors, new data
• Unstructured Data, Media and
therefore Big Data
• Shift from Model to Compute
• Specific Business Use Cases
• Shift from Management to
Operations and thereby Customer
• Privacy and Security will be a big
issue
• More utilities and Plug-n-Play
What to look for..
• Deep integration with a Business
outcome [ MyDrive]
• Show and Tell / Productized
services
• Eco System Partnerships
• Non-Linear Scale in the Business
Model
• Easy to Consume, Utility, Pricing
• Ability to Partner / Co-innovate
• Future Proofing customers.
• Agile Delivery Models
• Charging and Collection Model
[RDC]
• Application potential across the
Economy [ MyDrive]
• Time to deploy and transform [
Splunk ]
Business Model Factors
Solution Capability Development
Business Value Modeling.
Analytics Program Model..
Business Value and thereby Performance Hotspots drive solutions and messages
Sales &
Marketing
Member
Mgmt & UW
Provider
Mgmt
Claims
Mgmt
Customer
Service
Medical
Mgmt
Revenue - GTM
Business Case
Account Intel
Pitch /
Proposal
Partnership /
POC
Events / ABM
Engagements
Quote
Generation
Broker
Mgmt
Campaig
n Mgmt
Market
Research
Member
Retention
1. Brand Perception / Perf
Ratio
2. Influence Ratio
3. Number of leads
4. Cost per lead
5. Medium Conversion Rate
6. Avg Premium Val
7. Days visit to purchase
8. Task Completion Rate SOLUTION
CATALOG
KEY
OUTCOMES
Key
Resources
Partnership Algorithm
Training Research LAB/ COE
Understand Business Landscape:
What value is business after? Key pain
points in decision making / operations
Leverage Internal Capability:
No duplication of work already done /
capability already in existence
In Sight of the Customer:
Develop capability through the
customer, interface, POC / Pilots
Develop Ecosystem for delivery:
Relationships with established &
emergent OEM who will drive the
market
Time Bound:
Ensure outcomes with time frame. 3
months to customer and 6 months to
pilot
Develop Systemic Solutions:
Consulting to understand customer,
quick entry, low change and capital….
1
2
3
4
5
6Data
Process
Actions
Analytics
Visualization
Capability Framework
1
2
3
Key principles
Program Status
Business Themes and Analytics COE
Marketing RoI & Growth analytics
Customer acquisition analytics
Customer retention analytics
Social media driven analytics
Customer/Employee fraud & risk
Competitive intelligence analytics
Supply chain analytics
MFG process quality & compliance
Early warning analytics
Asset Perf. Maint. & warranty
Network analytics
Service Problem Analysis
Service Logistic & Resource Alloc.
Governance, Risk & compliance
Integrated financial perf. - EPM
Store operations Analytics
Merchandising & Pricing analytics
Claims analytics
Pre-Trade Post Trade Analytics
Drug discovery analytics
Post market analytics (Pharma)
Care & Safety analytics
Care analytics
Member Retention Analytics
Smart meter analytics
Technology
Business Automation Modeling
Data
Analysis
Visuals
Process
People
Methods
Tools
Vertical
Themes
Customer
Lifecycle
Service &
Warranty
GRC
EPM/WIPM
• Product Mgrs [10]
• Clustered Solution
Themes + verticals
• Teams for Verticals
program mgmt
• Modelers &
Technologist report
in.
• Business Consulting
• Innovation &
Transformation
Client Pitch /
Engagement
• Analytics Program
Management
• Long term look at
business Automation
solutions
• Modelers
• Cluster Solution
Themes
• Exploring Analysis Tools
• Develop Models/Methods
• # Of experiments
• Play with data
• Information Technologist
• Cluster 1
• All Data Gather &
Aggregation technologies
• Solution Warranty / Scale
• Speed, Variety – API
• # Of experiments
• Manage COEEnv.
RCTG, HLS, E&U,
Insurance, Securities
Common + special
aspects.. 5PDM,
expanded slowly.
Telecom, RCTG, E&U,
Banking, Insurance
2 PDM
1 BFSI, 1 OTH
MFG, E&U, Telecom,
1 PDM ALL
BFSI
1PDM ALL
All verticals, close collab
with WCS
Systematic Modeling Approach to Persistency
Propensity
Premium
Communication Strategy
Customer Segments
Act
To neutralize
the intent
Collect
Business need
and Data
Data Integration
Demographics for
Agency Information
Product Information
Pscyhographic History
Additional Sources of Data.
Optimize Data
Data Analysis +
Imputation
Bivariate Variable
Business Objectives
Major Risks Affecting Business
Customer Segments Scope
What’s Communication Strategy
Predict
The potential
customers
Analytical Model
Monitor + Feedback
Monitoring + Reports
Input feedback from operations
to further fine tune the model.
The Generic Analytical Modeling Process
DATA
COLLECTION
Business Problem
Definition
BUSINESS
PROBLEM
DEFINITION
DATA PREPARATION MODEL
DEVELOPMENT
MODEL DEPLOYMENT
&
MAINTENANCE
Business Problem
Statement
Collect & Analyze
Business
Requirements
Define Goals And
Objectives
FEEDBACK
Define Data
Requirements
Identify Data
Sources
Unstructured,
Structured, Internal
& External
Data Cleansing
Data Aggregation
Derived Variables
Model Selection
Build Connectors &
Data Marts
Data
Transformation
Variable Selection
Modeling Alternatives
Model Building
Model Training
Model Evaluation
Pilot Implementation
Model Validation
Recalibration
Monitoring
Business Process
Integration
Business Processes
& Systems
Knowledge
Data Modeling &
Business
Data Modeling
Knowledge
Intensive
Core: Business
Knowledge Intensive
Analytical Modeling
and Business
Knowledge
10-20% of Total
Effort
20-30% of Total
Effort
25-30 % of Total
Effort
5-10% of Total Effort 20-30% of Total Effort
PHASESKEYACTIVITIES:CORE&NON-CORE
KNOWLE
DGE
COST
Reporting & BP IntegrationAnalytical Support Team
Data Integration
MODELINGINFRASTRUCTURE
Internal Data [AIG]
Enterprise
Doc Manager
Loss
Notification
System
Claim
Admin
System
Policy
Admin
System
GL/Paymen
t
Engine
Data Preparation
Dashboards
& Reports
ANALYSISTEAM
External Data
Credit Records
Social Networks
Others
Data Marts,
ETL
Mapping,
Connectors
Analytics - Structural View
Core Analytical Modeling Team
Generic Analytical Models
Segmentation
Regression
Predictive Analytics
Core Insurance Analytical Models
Capital Adequacy Models
Pricing & Rating Models
Reserving Models
Risk Transfer Mechanisms
Modeling Foundation Data Governance
Specialized Data Marts Insurance Models & Standards Data Mining Tools
Modeling Repositories & Practices
Fraud Models
OUTCOME
Interventions through Data & Analytics
Data
Data Quality &
Cleansing
Pricing & Rating
Models
Dashboards:
Events &
Triggers
External Data
Data Integration
Services
Visualization
System
Integration - AIG
Reporting
Services
Reserving
Models
KPO / BPO
Services
Monitor model
performance
Modeling Business Services
Internal Data
Specialized
Research
Services
Model Validation
Unstructured
Data
Data - Readiness
Assessments
Actuarial Data Marts: Creation and
maintenance
Capital Adequacy
Models
Risk Transfer
Mechanisms
Model
Maintenance
Services for
Market Research
Vishwanath Ramdas
Head Analytics FCC Compliance , Large MNC Bank
8 years in the industry with 17 Y experience in Business Transformation.