Future of Business Analytics.pdf

THE FUTURE OF BUSINESS ANALYTICS WILL TOO MUCH DATA KILL THE CORPORATION?

TABLE OF CONTENTS

Executive Summary

2

Journey of Analytics - Introduction

3

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.


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 pe rformance 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 edu cation programs, and change management in addition to the traditional data gathering and analysis functions.


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 compan ies 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 = N ot an area of f ocus 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%


67%

30%

3%

Profitability improvement

56%

Customer service improvements

52%

New customer acquisition

52%

Inc reased

Stay ed the sam e

39%

5%

43%

5%

34%

14%

D ec reased

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

72%

Customer loyalty/retention

68%


67%

Pricing optimization New customer acquisition

59% 55%


Page 1

Strategic objectives for deploying business analytics across business units Enterprise

1

Business Intelligence

Data

7

Predictive Analytics 6

Predictive Models

Compete – Building unique competitive stronghold 





Grow – Incremental sales and customer retention Enforce –Business integrity with fraud management Improve – Improvement of core business capacity

Business Units:

7

Marketing

2

Sales

2

Fraud Detection

3

Call Center 2

4

Core Business Capacity (e.g. , produc t assembly)







4

s r e m 5 o t s u C

Satisfy – Meeting escalating consumer expectations Learn – Employment of advanced analytics Act – Actionable business intelligence and analytics


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 

A UPS manufacturer is experimenting with algorithms to adjust the order of deliveries as conditions (e.g., road closures, extraordinary customer need) change.

Customer intimacy 

Data-driven decision making 

A debt purchasing firm based in the UK uses SAS to predict debt portfolio performance. This enables the firm to make quicker decisions on acquiring new debt portfolios at the right prices, collect more from each portfolio and grow revenues by 50 percent annually

Visualization of assumptions in action 

Using analytics tool, an energy trading 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

FLOWERS.COM changes prices and offerings on its Web site frequently because it uses analytics. It also uses analytic software to target print and online promotions with greater accuracy & optimize its marketing, shipping, distribution and manufacturing operations. The result: a $50 million reduction in costs in FY 2009.

Risk and fraud mitigation 

Using SAS Fraud Management, part of the SAS Business Analytics Framework, HSBC prevents, detects and manages financial crimes by scoring and accepting or rejecting millions of transactions a day in real time


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


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


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 data lineage become manual, resourceintensive efforts. This can result in redundant efforts, metadata duplication, and potential errors Constant moving of data from one application to another

Most of the source data for advanced analytics come from enterprise data warehouses and departmental data marts, and the results from data mining and predictive models have to refer back to these repositories for further reporting and analysis. Therefore, moving these huge data sets from one DBMS to another always poses a throughput challenge and puts a strain on network traffic. Proliferation of disparate models, tools, and approaches

application, or business unit. Each of these projects often tags a group of statistical analysts, data mining specialists, and data modelers. Each team may have its own set of modeling tools, platforms, and approaches, which they may have adopted earlier. Confusion of inconsistent models across diverse applications

When analytical data marts are scattered, non-integrated, and non-conformed, it is difficult for data mining professionals to ensure that their models are consistent with those that have been developed by other business groups. Hence, model can’t incorporate the highest-quality reference data and leverage a consistent set of company-standard metadata, dimensions, schemas, and views.

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


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 pr eparation, modeling, and statistical analysis tasks BI Platform with Integrated Advanced Analytics

Functionality

Separate Bland Advanced Analytics tools

Advanced Analytics Integrated within BI Infrastructure

Administration

Separate

Common

Calculated Metrics

Build and maintain in at least two places

Build once, leverage across the 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 analysis

Difficult

Easier


Page 10

Analytics Maturity Cycle Maturity Level

Enabling Technology

Level 4: Comprehensive Cloud Analytics

y t i x e l p m o C

Level 3: Predic tive Grid Analytics

Level 2: Extensible Cluster Analytics

Level 1: Proprietary Platform Analytics

Service cloud; multiple programming languages; fully virtualized; extensible complex infor mation and streams; prebuilt and user-defined data management models and functions

Server grid; open API (e.g., Map Reduce), multiple pr ogramming languages, in cluding PA/DM domain-specific ( e.g.,R); diverse persistence management approaches (e.g., HDFS); extensible info types; prebuilt and user-defined predictive analytics models

Server cluster; pro prietary API; multiple programming languages; SQL and proprietary extensions; analytics optimized single-vendor RDBMS; extensible data types; prebuiltand user defined functions

Single server node; proprietary API; single programming language; general-pur pose RDBMS; SQL and proprietary extensions; structured data; prebuilt stored procedures

Usage Scenarios Optimize pushdown of finegrained PS/DM, DI , and ESP models across heteroge neous platforms, inf o, and streams

Execute PA/DM models eff iciently against heterogeneous historical content across gr id

Scale analytics across server clusters

Accelerate analytics development leveraging prebuilt functions

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


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 W eb Oriented Architecture platforms. Analytics is one of the most fascinating emerging technologies that will provide d ecision support system to the organizations over the decade.


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 ana lytics’ - 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-InMemory-Analytics.htm

 

    



IBM Combines Analytics, Business Intelligence in Cognos 10 http://www.ecrmguide.com/news/article.php/3910036/IBM-Combines-Analytics-Business-Intelligencein-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-for2011.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.


Page 13

Future of Business Analytics.pdf

Recommend Documents