Analytics Consulting

Analytics Consulting, Prasanna Parthasarathy

Analytics Consulting

Contributed By Prasanna Parthasarathy

Currently with Infosys Technologies as Associate Consultant – Business Intelligence-Enterprise Solution Alumnus of Bharathidasan Institute of Management

Bharathidasan Institute of Management, Trichy

Analytics Consulting, Prasanna Parthasarathy

Analytics Consulting 1. Market Analysis – Opportunities, Current Size and Potential Retail industry is one of the most competitive industries - thanks to low margins it offers. In low margin businesses it is the volume of operations which plays the crucial role. In the quest to get maximum share of the consumer’s spend retailers have been using technology to chalk out their strategies. Variety of CRM solutions are one such example of the usage of technology. Undoubtedly, many retailers were able to reap early dividends owing to their Customer Loyalty Programs. However, over a period of time they have become but a standard feature across retailers and as such no longer serve as a differentiator.

In such a scenario focus is back on how to improve upon the elusive ‘Net Profit’ figure. The answer probably lies not in ‘volume selling’ but in ‘intelligent selling’ and a host of other initiatives called together as ‘Analytics consulting’.

After Finance and Manufacturing industries Retail is one the biggest spender on IT. Present size of Retail industry worldwide is XXX bn USD and with an estimated spending of Y% of revenues on IT, it turns out to be sizeable ZZ bn USD. Though data is not readily available to ascertain the pie of Analytics consulting from this, rough estimates for Business Intelligence spend provide a figure of AA mn USD. It is to be noted IT organisations can look forward to gain 10% of this in near future.

Also, please note that Analytics consulting is only the area suggested by us to begin with to be a player in much bigger area of opportunity known as Analytics Consulting.

2. Targeted Market Retailers across the world.

Bharathidasan Institute of Management, Trichy

Analytics Consulting, Prasanna Parthasarathy 3. Service Idea “The key in business is to know something that nobody else knows.” — Aristotle Onassis “To understand is to perceive patterns.” — Sir Isaiah Berlin Analytics consulting is sometimes described as the ‘Complete Thought’ involving both the ‘Left Brain’ (Logical, objective; sees the details) and the ‘Right Brain’ (Intuitive, subjective; gets the big picture).

Analytics consulting will encompass a whole gamut of solutions which aim at improving the efficiency of operations at a retailer’s end.

The benefits of Analytics consulting or Predictive Analytics extend from cost reductions to incremental revenue, from customer cross-selling and upselling to product supply chain optimization and market basket analysis.

In general, some IT applications, systems or projects are mostly cost-centric, whereas others tend to support producing incremental revenue. For example, metadata initiatives, data mart consolidation or certain kinds of data quality projects tend to support operational efficiencies and cost reductions, largely by helping IT to work smarter and do more with less. This is Business Intelligence or Data warehousing. This is the space where Data warehousing organizations across the world foocuses currently. It is true, incremental revenues

sometimes

result

from

advanced

applications

with

sophisticated

dynamics from data mart consolidation when the entire technology stack is consolidated. But by and large the result is cost reduction.

In contrast, Analytics consulting achieved by predictive analytics tend to provide benefit in generating incremental revenue. It makes sense to drill down on both dimensions - the customer as well as the product - and consider how predicative

Bharathidasan Institute of Management, Trichy

Analytics Consulting, Prasanna Parthasarathy analytics enables cost reductions as well as incremental revenue opportunities. This results in the classic two-by-two matrix depicted in the figure below.

Customer

Incremental Revenue

Product

• Cross-Selling

• Market Basket Analysis

• Up Selling

• Pricing Model

• Customer Profitability

• Product Promotion • Merchandising optimisation

Cost Reduction

• Customer Scoring / Credit worthiness • Fraud detection / detection via profiling • Customer cost profiling

• Demand planning for inventory reduction • Supply chain optimisation • Machine / system failure forecasting

Customer/Cost Reductions [Cr-CoR]: Applications include scoring customers according to a variety of behaviors, especially credit worthiness and use, and rejecting high-risk customers. Building a customer profile is on the critical path to fraud detection, and many applications in fraud detection are designed around profiling. Of course, this contributes to cost reduction. Fraud detection and reduction are growth industries where Fair Isaac (including the technology acquired in the HNC merger) has pioneered predictive profiling with Falcon Fraud Manager and analogous programs for insurance, retail, telecommunications and finances, designed and implemented around profiling customers and buying behavior. Quadstone

specializes in predicting telecommunications customer

churn and analyzing the root causes of churn to reduce it. Reduced costs of direct mail through superior targeting also play a role here.

Bharathidasan Institute of Management, Trichy

Analytics Consulting, Prasanna Parthasarathy Customer/Incremental Revenue [Cr-IR]: Cross-selling, upselling, customer profitability - this is the most famous quadrant in the matrix and the one that has received the most attention. Unica Affinium, Group 1 Model 1 and Quadstone have established client bases in the direct marketing vertical where gaining efficiencies through targeted promotions and lift charts represents revenue incrementally earned. SAS Marketing Automation and KXEN also have applications that address the marketing opportunity. Siebel and PeopleSoft have licensed private-label versions of the ANGOSS Mining Manager technology, which builds solutions with templates and industry-specific predictive analytic content in their analytic CRM offerings.

Product/Cost Reductions [Pr-CoR]: Demand planning to reduce inventory, sourcing product based on forecast, supply chain optimization, machine/system failure analysis and prediction

Product/Incremental Revenue [Pr-IR]: Market basket analysis, sourcing based on demand forecast, revenue optimization from pricing mark downs (retail), merchandising and promotions (retail) - portfolio balancing in the capital markets is a source of incremental revenue due to arbitrage opportunities based on the prediction of default and prepayment rates as well as risk (and therefore cost) reduction.

Workbenches can be used to build applications when the retail problem is not well-defined or when there are many problems. The predictive modeling workbenches such as SAS Enterprise Miner, SPSS Clementine, IBM Intelligent Miner For Data, Insightful Miner and ANGOSS are a good fit under these circumstances - when an enterprise will confront a diversity of opportunities or when the problem is not well defined in advance - as well as when the team does not want to code directly in the underlying statistical language such as SAS, SPSS or S-Plus.

Bharathidasan Institute of Management, Trichy

Analytics Consulting, Prasanna Parthasarathy

Data Warehousing Query

&

Classic Data Mining

Reporting Statistical Analysis

functions

Predictive Analytics Prescriptive Algorithms

Static perspective

Continuous changes

Also

discontinuous

changes Describe the present and Predict the past

Predict the future

the past Assume Hypothesis

Validate Hypothesis

Invent

and

Validate

Hypothesis

Differentiators

In data warehousing, the analyst asks a question of the data set with a predefined set of conditions and qualifications, and a known output structure. This is the traditional data cube - what customers are buying or using what product or service and when and where are they doing so? Typically, the question is represented in a piece of SQL against a relational database. The business insight needed to craft the question to be answered by the data warehouse remains hidden in a black box - the analyst's head.

Data mining gives us tools with which to engage in question formulation based primarily on the "law of large numbers" of classic statistics. Predictive analytics has introduced decision trees, neural networks and other pattern matching algorithms constrained by data percolation. It is true that, in doing so, technologies such as neural networks have themselves become a black box. However, neural networks and related technologies have enabled significant progress in automating, formulating and answering questions not previously envisioned. In science, such a practice is called "hypothesis formation," where the hypothesis is treated as a question to be defined, validated and refuted or Bharathidasan Institute of Management, Trichy

Analytics Consulting, Prasanna Parthasarathy confirmed by the data. The confirmation or refutation of the hypothesis counts as knowledge in the strict sense. In neither case, data mining or predictive analytics, is a decision made. A prediction is a prediction, not a decision. The ultimate determining mark of predictive analytics (and applications) is that the prediction is inside the model.

A Case of Customer clustering in Retail (Source: Forrester Research)

The future of Business Intelligence and Data Warehousing lies in predictive analytics. Predictive analytics has successfully proliferated into applications to support customer recommendations, customer value and churn management, campaign optimization, and fraud detection. On the product side, success stories in demand planning, just-in-time inventory and market basket optimization are a staple of predictive analytics. The retail clients should focus on Analytics consulting

to get to know the customer and segment, to predict customer

behavior, and to forecast product demand and related market dynamics. Be realistic about the required complex mixture of business acumen, statistical processing and information technology support as well as the fragility of the resulting predictive model, but make no assumptions about the limits of predictive

Bharathidasan Institute of Management, Trichy

Analytics Consulting, Prasanna Parthasarathy analytics. Breakthroughs often occur in the application of the tools and methods to new commercial opportunities. Recommendations End-user enterprises should plan on using predictive analytics to gain incremental revenue

opportunities

though

superior

predictions

about

consumer

buying

behavior, product pricing and related market dynamics. Sample applications include

attrition

reduction

in

finance

and

telecommunications;

pricing

optimization in retail; fraud reduction in insurance; demand planning inventory optimization in manufacturing and consumer packaged goods (CPG); and loyalty development, upselling and cross-selling in direct marketing and customer-facing applications across diverse industries. Enterprises should look for tools that offer a wide variety of high-performance statistical functions for data preparation and analytic algorithms for predictive modeling and inference. Enterprises should take a cross-functional approach to predictive analytics. In particular, enable collaboration between business owners, IT data preparation and deployment, and statisticians in order to gain the benefits of predictive analytics in incremental revenue opportunities.

Bharathidasan Institute of Management, Trichy