Sei sulla pagina 1di 14

THE FUTURE OF

BUSINESS ANALYTICS

WILL TOO MUCH DATA KILL THE CORPORATION?


Executive Summary 2

Journey of Analytics - Introduction 3


TABLE OF CONTENTS
CXO Outlook 4

Business Benefits 6

Implementation Techniques in Analytics 8

Implementation Challenges 9

Roadmap for Analytics 12

References 13

About the Authors 13

The information contained in this document is proprietary. Copyright © 2011 Capgemini. All rights reserved. Page 1
EXECUTIVE SUMMARY
Operational scope of organizations has
increased from local to global with
liberalization of economies and
emergence of communication media. Its
performance has become a complex
function of external factors linked to
multiple decision makers and
stakeholders. Firms have included
futuristic approach in their working to
accommodate them. Relevant
information needs to be filtered from
valuable information & extensive use of
past data should be made to take future
decisions. Business analytics is playing
a crucial role in this context
Analytics provides widespread
applications across various sectors and
multiple user hierarchies, since it
provides tangible business benefits like
higher return on data related investment,
real time intelligence & risk & fraud
mitigation for data sensitive applications.
The increased adoption of SaaS, use of
in-database analytics; cloud computing
and open source will lower entry-level
barriers for analytics deployment.
Adoption of mobile devices will increase
for analytics applications and extent of
packaged implementation will go up.
Collaborative decision environments will
drive analytics investment, particularly
those that link with collaboration and
social networking functions. Newer
analytical methods in the areas of text
analytics, survival mining, time series
mining, net-lift modeling, and data
visualization will grow, but regulatory
and privacy constraints will have to be
dealt with as analytics moves out of IT
domain as a mainstream operation.

The information contained in this document is proprietary. Copyright © 2011 Capgemini. All rights reserved. Page 2
JOURNEY OF ANALYTICS - INTRODUCTION
As economies have opened up, horizons have widened for organizations to establish global
presence. This has been accompanied by increase in sources of data with the advent of
internet, search engines and social networking channels. Impact of external factors like
competition, regulation, global trade and emergence of multiple stakeholders has increased
business complexity and organization’s performance is dependent on them.
Data availability is no longer a differentiator for decision making since everyone has access to it.
While comprehensive & high quality data generates more knowledge, not all of it is relevant
business information that could be put to use. Efficient filtering of the data is necessary for
better decision making.
Traditionally organizations have been using business intelligence for knowledge creation. If it
were a species, we can envision it feeding on raw data, digesting that raw data into information
and then information into knowledge, and producing beautiful blooms as visual representations
of that knowledge.
Business intelligence has evolved beyond knowledge creation of what has happened in the past
to analytics and forecasting what might happen in the future. Business analytics has emerged
as a popular technique to provide sustainable competitive advantage to the have’s over have’s
not. Technology has added data mining, which helps distinguish between relevant and irrelevant
information, so that people don’t have to make that decision. Predictive models and other
forecasting tools have evolved to evaluate any number of possible future outcomes and guide
business users towards the optimal decision.

With changing times, analysts are morphing into consultants who may be responsible for
framing decisions, process redesign, communication and education programs, and change
management in addition to the traditional data gathering and analysis functions.

The information contained in this document is proprietary. Copyright © 2011 Capgemini. All rights reserved. Page 3
BusinessWeek launched a research program in April 2009 to determine the attitudes and
opinions of C-level executives at leading large and mid size companies spread across sectors
with regard to the use and value of business analytics.
 Executives are looking to derive greater value from existing customer relationships and work
on customer retention
 Majority feel that customer retention and management have become more difficult as
consumers are being forced to better manage their spending and amplify their savings in a
highly constrained environment
 An agreement has emerged over the fact that business analytics can have a significant
impact on customer service improvements, customer retention, and expanding existing
customer relationships
 The respondents believe that business analytics enables companies to develop agile
strategies that allow them to adapt to changing customer behavior and achieve business
goals

CXO OUTLOOK
Question:
Please indicate how much focus your organization will place on the following
programs/initiatives in 2009. Use scale where: 1 = Not an area of focus for 2009; 5 = Key
area of focus for 2009 (NET IMPACT = 4,5)
Customer loyalty/retention 91%

Expanding existing customer relationships 87%

Profitability improvement 79%

Customer service improvements 76%

New customer acquisition 74%

Question:
How has your company’s focus on these areas changed over the last 12 months?

Customer loyalty/retention 69% 28% 3%

Expanding existing customer relationships 67% 30% 3%

Profitability improvement 56% 39% 5%

Customer service improvements 52% 43% 5%

New customer acquisition 52% 34% 14%

Increased Stayed the same Decreased

Question:
How much impact do you believe a business analytics approach would have on the
following areas? Use scale where: 1 = No Impact; 5 = Significant impact
(NET IMPACT = 4,5)
Customer service improvements 72%

Customer loyalty/retention 68%

Expanding existing customer relationships 67%

Pricing optimization 59%

New customer acquisition 55%

The information contained in this document is proprietary. Copyright © 2011 Capgemini. All rights reserved. Page 1
Strategic objectives for deploying business analytics across
business units

Enterprise 1

Business Units: 7
Business Intelligence 7
Marketing 2

Sales 2

Predictive Fraud Detection 3

Data Analytics
Call Center 2 4
6

Customers
Predictive Core Business
Models Capacity 5
4
(e.g. , product
assembly)

Compete – Building unique competitive  Satisfy – Meeting escalating consumer


stronghold expectations
 Learn – Employment of advanced
 Grow – Incremental sales and customer
analytics
retention
 Act – Actionable business intelligence
 Enforce –Business integrity with fraud
and analytics
management
 Improve – Improvement of core
business capacity

The information contained in this document is proprietary. Copyright © 2011 Capgemini. All rights reserved. Page 5
BUSINESS BENEFITS

Higher return on the organization’s data


investment
 In the Philippines, the Bureau of Internal
Revenue used analytics to recoup $114
million in unpaid value-added taxes,
resulting in 400 percent ROI in the first
year of implementation
Hidden meaning in gathered data
 A player in the secondary-ticket market
uses SAS to develop a deeper
understanding of the needs of its
thousands of customers. By segmenting
them and catering to psychographics,
the company optimizes how frequently it
contacts the customers and improves
loyalty
Real time intelligence Customer intimacy
 A UPS manufacturer is experimenting  FLOWERS.COM changes prices and
with algorithms to adjust the order of offerings on its Web site frequently
deliveries as conditions (e.g., road because it uses analytics. It also uses
closures, extraordinary customer need) analytic software to target print and
change. online promotions with greater accuracy
Data-driven decision making & optimize its marketing, shipping,
distribution and manufacturing
 A debt purchasing firm based in the UK operations. The result: a $50 million
uses SAS to predict debt portfolio reduction in costs in FY 2009.
performance. This enables the firm to
make quicker decisions on acquiring Risk and fraud mitigation
new debt portfolios at the right prices,  Using SAS Fraud Management, part of
collect more from each portfolio and the SAS Business Analytics Framework,
grow revenues by 50 percent annually HSBC prevents, detects and manages
Visualization of assumptions in action financial crimes by scoring and
accepting or rejecting millions of
 Using analytics tool, an energy trading transactions a day in real time
company enables staff to predict what
electricity and gas purchases done
today will sell for months later when
consumers buy. Business analytics
supplies that intelligence to traders in a
cleaner, faster and more accurate way

The information contained in this document is proprietary. Copyright © 2011 Capgemini. All rights reserved. Page 6
Companies have taken initiatives in the direction of
developing analytics expertise through technology
advancements and tie-ups.
 SAP has begun shipping its High Performance Analytic
Appliance software based on BusinessObjects BI
software. This application will be primarily used for sales
pipeline forecasting, smart meter analytics for utilities
and consumer packaged goods and retail applications
such as promotion planning

 IBM has combined analytics and BI in its latest Cognos


10 application that includes collaboration and analytics
capabilities and iPad, iPhone and BlackBerry support

 Teradata has acquired Aprimo, a maker of Web-based


integrated marketing software. Teradata will use
Aprimo's cloud-based integrated marketing software to
combine business analytics with integrated marketing
solutions to enable corporations to improve and optimize
marketing performance

 Oracle has rolled out BI Applications Release 7.9.7.


Oracle Financial Analytics for SAP, a feature from this
release helps front-line managers improve financial
performance and decision-making with comprehensive,
timely and role-based information on their departments'
expenses and revenue contributions, while reducing the
complexity and costs of integrating information from SAP

The information contained in this document is proprietary. Copyright © 2011 Capgemini. All rights reserved. Page 2
IMPLEMENTATION TECHNIQUES IN ANALYTICS
Most common business functions in analytics solution
 Outcome classification
 Clustering – Customer segmentation exercise
 Associative analytics – Generally used for market basket
analysis
 Time series forecasting
 Text analytics – Exploring unstructured information to
discover patterns

Operations performed on the data


 Selection
 Transformation
 Exploration
 Modeling – Training, testing and evaluating
 Deployment – Scoring the model, embedding
models (Using PMML etc), automated decision
mgmt
 Model Management

Commonly used algorithms


 Neural networks
 Decision trees
 Clustering algorithms
 Bayesian belief networks
 Sequential analysis

The information contained in this document is proprietary. Copyright © 2011 Capgemini. All rights reserved. Page 8
IMPLEMENTATION CHALLENGES
Many companies have deployed their BI
and advanced analytics investments in
silos. When architected and implemented
separately, a complete BI solution suffers
from the following shortcomings:
Lack of common metadata
When BI and advanced analytics come from
different vendors, some metadata has to be
entered and maintained in at least two
places. Without tightly integrated metadata,
critical tasks such as impact analysis and application, or business unit. Each of these
data lineage become manual, resource- projects often tags a group of statistical
intensive efforts. This can result in analysts, data mining specialists, and data
redundant efforts, metadata duplication, and modelers. Each team may have its own set
potential errors of modeling tools, platforms, and
approaches, which they may have adopted
Constant moving of data from one earlier.
application to another
Confusion of inconsistent models across
Most of the source data for advanced diverse applications
analytics come from enterprise data
warehouses and departmental data marts, When analytical data marts are scattered,
and the results from data mining and non-integrated, and non-conformed, it is
predictive models have to refer back to difficult for data mining professionals to
these repositories for further reporting and ensure that their models are consistent with
analysis. Therefore, moving these huge those that have been developed by other
data sets from one DBMS to another always business groups. Hence, model can’t
poses a throughput challenge and puts a incorporate the highest-quality reference
strain on network traffic. data and leverage a consistent set of
company-standard metadata, dimensions,
Proliferation of disparate models, tools, schemas, and views.
and approaches
Advanced analytics initiatives address
requirements specific to a project,

The Solution: In-Database Analytics


As part of in-database analytics, developers embed application logic into database management
systems. It provides the following advantages:
1) The approach enables streamlined development, accelerated execution, improved
consistency across database applications, and decoupling of application logic from the user
interface layer.
2) In-database analytics in the EDW eliminates the need to move massive analytical data sets
between the EDW and data mining workbenches

The information contained in this document is proprietary. Copyright © 2011 Capgemini. All rights reserved. Page 9
3) By making the EDW the central governance point, in-database analytics allows
organizations to enforce cross-project enterprise-standard enterprise templates, hierarchies,
dimensions, transformation rules, and data cleansing.
4) The approach shortens the time needed to build, execute, deploy, and optimize predictive
models by accelerating data preparation, transformation, loading, clustering, segmentation,
scoring, and other functions in the EDW
5) In-database analytics eliminates or greatly reduces the need for data modelers, data mining
specialists, subject-matter experts, and power users to duplicate data preparation, modeling,
and statistical analysis tasks

BI Platform with Integrated Advanced Analytics

Advanced Analytics
Separate Bland Advanced
Functionality Integrated within BI
Analytics tools
Infrastructure

Administration Separate Common

Build and maintain in at least two Build once, leverage across the
Calculated Metrics
places platform

Data lineage analysis Customized, manual effort Out-of-the-box functionality

Data movement requirement High Low

Impact analysis Customized, manual effort Out-of-the-box functionality

ETL processes Separate Common

Metadata Separate Common

User Interface Separate Common

Version control, impact


analysis, data lineage Difficult Easier
analysis

The information contained in this document is proprietary. Copyright © 2011 Capgemini. All rights reserved. Page 10
Analytics Maturity Cycle
Maturity Level Enabling Technology Usage Scenarios
Service cloud; multiple
programming languages; fully
Level 4: Optimize pushdown of fine-
virtualized; extensible complex
grained PS/DM, DI , and ESP
Comprehensive Cloud information and streams; prebuilt
models across heterogeneous
and user-defined data
Analytics platforms, info, and streams
management models and
functions

Server grid; open API (e.g., Map


Reduce), multiple programming
languages, including PA/DM
Level 3: domain-specific (e.g.,R); diverse Execute PA/DM models
Predictive Grid persistence management efficiently against heterogeneous
approaches (e.g., HDFS); historical content across grid
Complexity

Analytics extensible info types; prebuilt


and user-defined predictive
analytics models

Server cluster; proprietary API;


multiple programming
Level 2: languages; SQL and proprietary
Scale analytics across server
Extensible Cluster extensions; analytics optimized
clusters
single-vendor RDBMS;
Analytics
extensible data types; prebuilt-
and user defined functions

Single server node; proprietary


Level 1: API; single programming
Accelerate analytics
language; general-purpose
Proprietary Platform RDBMS; SQL and proprietary
development leveraging prebuilt
functions
Analytics extensions; structured data;
prebuilt stored procedures

Time

Analytics provides need based solutions at multiple complexity levels. This leads to increased
scope/popularity of the technique across larger user groups with varying technical skill sets

The information contained in this document is proprietary. Copyright © 2011 Capgemini. All rights reserved. Page 11
ROADMAP FOR ANALYTICS
With a modest economic growth predicted in 2011, the use of analytics as a competitive
differentiator in selected industries will explode. 2011 marks the beginning of a multi-year cycle
where the sparks of economic growth ignite a tinderbox of technology and business forces that
are set to drive mainstream adoption of business analytics across commercial and government
sectors. The gap between analytical innovators and those who do not invest in analytics will
widen in high-profile ways. Pharmaceuticals, entertainment, airlines and baseball will be some
of the industries where the difference between innovators and laggards will begin to stand out.
The roles of marketing, sales, human resources, IT management, and finance will continue to
be transformed by the use of analytics & HR will be one of the fastest growth areas for analytics.
Going ahead, collaborative decision environments will drive investment in new BI and analytic
applications, particularly those linking collaboration and social networking functions. Newer
analytical methods in the areas of text analytics, survival mining, time series mining, net-lift
modeling, and data visualization are set to grow. Database capacity, processor speeds and
software enhancements will continue to drive even more sophisticated applications of analytics.
Majority of business intelligence functionality will be consumed via handheld devices as
enterprises and vendors will develop mobile analytic applications for specific domains in the
coming years. Many of analytic applications will use in-memory functions to add scale and
computational speed, while some will use proactive, predictive and forecasting capabilities.
Demand for stand-alone analytic products such as SAS, IBM SPSS and open-source R will also
increase. Consolidation of analytics software players will continue; entry of specialized analytics
software and service providers will accelerate and most of spending on business analytics will
go to system integrators, not software vendors. The availability of strong business-focused
analytical talent will be the greatest constraint on organizations' analytical capabilities and many
other organizations will begin to realize the need for a centralized management of analytical
capabilities. Regulatory and privacy constraints will continue to hamper growth of marketing
analytics.
It’s likely that business analytics will converge with collaboration and social software and the
future of enterprise architecture lies in Web Oriented Architecture platforms. Analytics is one of
the most fascinating emerging technologies that will provide decision support system to the
organizations over the decade.

The information contained in this document is proprietary. Copyright © 2011 Capgemini. All rights reserved. Page 12
REFERENCES
 ‘2010 Trends to Watch: Business Intelligence Technology’ – Ovum
 ‘SAP looks to fast-track business analytics’ – Ovum
 ‘Business intelligence and analytics fundamentals’ – Ovum
 ‘Getting more from your data with predictive analytics’ - Ovum
 ‘In-Database Analytics: The Heart Of The Predictive Enterprise’ – Forrester‘
 ‘ Predicts 2011: New Relationships Will Change BI and Analytics’ – Gartner
 ‘SAP Speeds Up Business Intelligence with In-Memory Analytics
http://www.ecrmguide.com/news/article.php/3915246/SAP-Speeds-Up-Business-Intelligence-with-In-
Memory-Analytics.htm
 IBM Combines Analytics, Business Intelligence in Cognos 10
 http://www.ecrmguide.com/news/article.php/3910036/IBM-Combines-Analytics-Business-Intelligence-
in-Cognos-10.htm
 ‘Seven reasons you need predictive analytics today’ – IBM website
 ‘SAS Business Analytics for IT Leaders’ – SAS website
 ‘Business Analytics for the CIO’ - SAS website
 ‘Brain trust- Enabling the confident enterprise with business analytics’ -- SAS website
 ‘The Customer You Know : Keeping, leveraging & profiting from current customers with
business analytics’ – BusinessWeek research services
 Nine Business Analytics Predictions for 2011 -
http://www.ecrmguide.com/article.php/3915626/Nine-Business-Analytics-Predictions-for-
2011.htm

ABOUT THE AUTHORS


This paper has been authored by Swanand Ranade, Arjun Vazirani , Vivek Sharma & Pratosh
Jhari of North America Sales Centre.

We the authors grant Capgemini royalty-free and full rights to use and publish the submitted
paper in any format.

The information contained in this document is proprietary. Copyright © 2011 Capgemini. All rights reserved. Page 13

Potrebbero piacerti anche