Unraveling the Customer Mind

1,639
-1

Published on

We offer a framework for conducting market basket analysis (MBA), a powerful technique for determining retail association rules and other patterns that can help retailers increase profits and customer loyalty.

0 Comments
0 Likes
Statistics
Notes
  • Be the first to comment

  • Be the first to like this

No Downloads
Views
Total Views
1,639
On Slideshare
0
From Embeds
0
Number of Embeds
1
Actions
Shares
0
Downloads
0
Comments
0
Likes
0
Embeds 0
No embeds

No notes for slide

Unraveling the Customer Mind

  1. 1. • Cognizant 20-20 InsightsUnraveling the Customer MindA soft economy and unrelenting margin pressures require retailers tothink more proactively about what drives consumer buying preferencesand behaviors. Our unique twist on market basket analysis can helpretailers create more personalized campaigns that result in greatertransactional activity and deeper customer loyalty. Executive Summary highly variable. Thus it is important to understand how budgets are being used, what impact it has As the economy slowly recovers and a raft of on your organization/business and how to ensure new channels, products and offers emerge, some success within this changing landscape. Tradition- consumer segments (particularly the affluent) ally, retailers spend a significant amount of their are gradually spending more money at retail campaign budgets on TV, followed by print, online, stores. It is therefore imperative for retailers to catalogues, direct marketing, events and outdoor better understand consumer thinking and derive advertising. However, the pattern is shifting behavioral insights that reveal buying prefer- toward greater investments in online channels. ences and decisions that take advantage of this The retail industry is expected to be one of the rejuvenated willingness to consume. biggest spenders on online advertising, according Consumers are ever-changing. In fact, research to researchers who follow the space. (both hard data and anecdotal evidence) shows The industry is also highly competitive and they are better informed, more nimble, more fragmented globally, so it is increasingly selective and less loyal than ever before. To win important to capture and analyze every customer the hearts and minds of consumers, billions of interaction. Moreover, with the improved sophisti- dollars are spent annually on mass advertis- cation of data handling and collection technology, ing and promotional campaigns both online and retailers have access to a gold mine of POS offline. These campaigns are not targeted to consumer transactional data across multiple specific customer segments nor are they linked to channels. customers’ online and offline purchasing behavior because most retailers lack an in-depth under- The struggle for retailers is to differentiate standing of consumer preferences and purchasing themselves through their product offerings and behavior. This results in campaigns that bombard promotions to customers across all channels. consumers with offers, discounts and promotions Some smart retailers effectively deploy targeted at odds with their interests and needs. product offerings, which can generate significant revenues; poorly chosen ones, though, waste Within this changing landscape, both marketing money and opportunity. For example, grocery strategies/options and marketing budgets are cognizant 20-20 insights | december 2012
  2. 2. chain operator Kroger recently refined its direct long did it take to tend? How many items — by marketing strategy by using data from its loyalty- type — were in the basket? What was the relation- card program and sent unique coupon offerings to ship among the purchased items? How do other specific households. In the words of David Dillon, baskets compare? Kroger’s chairman-CEO, “We The reasons why most retailers refrain from a The reasons why understand and appreciate that deep-dive analysis is their preconceived notionmost retailers refrain no two customers are alike.” The that analyzing data at this level of granularity is company believes that this level from a deep-dive of promotional personalization expensive, overly time-consuming and has limited business value. analysis is their offers a path toward creating a preconceived notion direct link to its customers that Solutions Emerge no other U.S. grocery retailer that analyzing can replicate.1 Effective retailing requires an immediate response data at this level to consumer requirements, mandating extremely This white paper lays out new efficient, agile and responsive business and oper- of granularity is thinking about market basket ational processes. Store operations, merchandis- expensive, overly analysis that can help retailers ing, marketing and advertising must all perform time-consuming drive campaign effectiveness in consistently, with little room for error. Due to con- more tailored ways and generate tinually increasing pressure on margins in many and has limited greater per-customer purchases industry segments, only the best retailers are business value. and loyalty. surviving; few are thriving. In Search of Enlightenment To execute on all cylinders, retailers must gain an in-depth understanding of their operations and Since the introduction of electronic point of sale, maintain the ability to delve into the operational retailers have had at their disposal an incredible data to ask (and answer) any and all business- amount of data. The sheer volume of data, critical questions. Traditional DW/BI systems were however, obscures patterns, making it impossible designed to handle large amounts of summarized to discern customer preferences and behavior via information to address a predefined set of manual inspection. The initial challenge, tradi- questions. To be agile and adaptable, retailers tionally, has been to seek ways to leverage trans- need the ability to continually measure, track and actional data to produce business value. Most probe all aspects of their businesses to answer retailers have already figured out a way to con- new questions. solidate and aggregate their data to understand the basics of the business — what is selling, how This starts with effective market basket analysis. many units are moving and the sales amount. Figure 1 depicts a shopping cart by a typical However, few retailers are successfully analyzing consumer containing various purchased products. this data at its lowest level of granularity: the A complete list of purchases made by all customers market basket transaction. provides much more information; it describes the most important part of a retailing business — what In fact, the essence of the analysis lies in the merchandise customers are buying and when. depths of the detail. Thus, a 360-degree view The analysis uses the information about what of the customer through integrated online and customers purchase to provide insight into who offline transactional data can ensure retailers they are and why they make certain purchases. that the products they offer and promotions they run match shopper preferences and behavior Market basket analysis provides insight into the and deliver maximum return on their marketing merchandise by highlighting which products spend. The ability to link purchases to individual tend to be purchased together and which are purchasers can take this even further; for most amenable to promotion. This information is example, tailoring offers to specific customer actionable; it can be used to: segments and driving higher returns from more precisely targeted campaigns. • Understand new store layouts. • Determine which products to put on special For retailing, this would mean understanding promotions and bundles. everything possible about the sales transaction, including: What time of day did the customer • Indicate when to issue coupons. shop? How long did it take to check out? Was a The data-mining technique most closely allied loyalty card used? Who was the cashier? How with this analytical approach to market basket cognizant 20-20 insights 2
  3. 3. Market Basket Analysis Unravels Customer Purchasing Behavior In this shopping basket, the shopper placed a quart of orange juice, some bananas, dish detergent, some window cleaner and a six-pack of soda. Is soda typically purchased with Are window cleaning products bananas? Does the brand of purchased when detergent soda make a difference? and orange juice are bought together? How do the demographics of What should be in the the neighborhood affect basket but is not? what customers buy?Figure 1analysis is the automatic generation of associa- • Choosing the right set of items.tion rules. Association rules represent patterns inthe data without a specified target. As such, they • Generating rules by deciphering the counts in the co-occurrence matrix.are an example of undirected data mining.2 • Overcoming the practical limits imposed byIn a relational database, the data structure thousands or tens of thousands of items.for market basket data has the following keycomponents: A word of caution regarding association rules: Retailers need clarity and utility of the results,• The order is the fundamental data structure for which are in the form of rules about groups of market basket data. products. There is an intuitive appeal to an asso- ciation rule because it expresses how tangible• Individual items in the order are represented products and services group together. separately as line items.• Product reference tables provide more descrip- tive information about each product. This A Co-occurrence Three- should include the product hierarchy and other dimensional Matrix information that might prove valuable for analysis.• The customer table is an optional table and should be available when a customer can be identified. Detergent 1 0 0 1 1Association rules were originally derived from Soda 2 0 0 2 1point-of-sale data that describes what productsare purchased together. The sheer bulk of this Milk 1 1 1 0 0transactional data (see Figure 2) — recording, atthe item level, every purchase through stores, Cleaner 1 1 1 0 0online storefronts and other channels – makes it Detergent Sodavery hard to understand the repeated patterns of OJ 4 1 1 2 1 Milk Cleanerpurchasing that provide insights into customer OJ OJ Cleaner Milk Soda Detergentbehavior and preferences. Orange juice, milk and window cleanerThe basic process for finding association rules is appear together in exactly one transaction.illustrated in Figure 3. There are three importantsteps for creating such rules: Figure 2 cognizant 20-20 insights 3
  4. 4. Three Basic Steps for Building Association Rules 1 First, determine the right set of items and the right level. For instance, is pizza an item or are the toppings items? Topping Probability 2 Next, calculate the probabilities and joint probabilities of items and combinations of interest, perhaps limiting the search by using thresholds on support or value. If mushroom, then pepperoni! 3 Finally, analyze the probabilities to determine the right rules.Figure 3We created a market basket analysis (MBA) • Increasing the size and value of the marketsolution framework to help clients demystify this basket.data and provide actionable insights into customerpurchasing behavior and their responsiveness • Testing and learning by using the marketplace as a laboratory.to certain channels/promotions. Although theirroots are in analyzing point-of-sale transactions, • Determining the “magic” price points for this store.association rules can be applied outside the retailindustry to find relationships among other types • Matching inventory to need by customizingof “baskets.” Examples of potential cross-industry store and assortment to trade area demo-applications include: graphics.• Items purchased on a credit card, such as rental • Optimizing store layout. cars and hotel rooms. Thus, pattern analysis can be used to drive• Information on value-added services pur- decisions on how to differentiate store assortment chased by telecom customers (call waiting, call and merchandise as well as to effectively combine forwarding, DSL, speed call, etc.) can help oper- offers of multiple products, within and across ators determine how to improve their bundling categories, to drive higher sales and profits. of service packages. These decisions can be implemented across an• Unusual combinations of insurance claims can entire retail chain, by channel or, if the data is be a sign of fraud. analyzed at the store level, customized offers can be formulated and deployed at a local level.Retailers’ Use of MBA Analysis It is imperative to understand the critical missingRetailers use MBA analysis to explore transaction link; traditional basket analysis fails to providedata to determine the affinities of what people actionable insights. However, when done correctly,buy to detect changes in basket composition, MBA can uncover how effective promotion ofsize and value, and to discover new insights into a given product can increase sales and profitcustomer buying behavior. This includes: influenced by related products and drive larger• Identifying more profitable advertising and baskets or greater transaction volumes. Hence, promotions. our approach is aimed at revealing transaction• Targeting offers more precisely to improve ROI. patterns, illuminating key cause-and-effect rela- tionships, to help retailers isolate the incremen-• Generating better loyalty card promotions with tal impact of any given promotion (i.e., transac- longitudinal analysis. tions that wouldn’t have occurred but for the• Attracting more traffic into the store. promotion). Affinity analysis3 on its own cannot cognizant 20-20 insights 4
  5. 5. accomplish this; however, conducting a test vs. milk at an aggressive every-day low price (EDLP);control measurement along with market basket sell-through is high and sales look good. And theanalysis helps to unravel the cause-and-effect company is confident that the sacrificed margin isrelationships related to the incremental impacts justified as it must be driving traffic to its storesof promotions. and generating incremen- tal sales of other items. Predictive models helpsThe analysis framework links actions to outcomes However, upon looking more retailers to direct theand empowers the retailers to truly understand closely at the baskets thatthe mind of the consumer. The solution contain milk, the retailer right offer to the rightframework is tool agnostic and can be deployed realizes that those baskets customer segments/as an analytical layer to a vast majority of COTS tend to be single-SKU, or profiles, as well astools available in the market. Personalization and otherwise small baskets. Intailor-made campaigns are all the rage in retail reality, the pricing strategy gain understanding oncampaigns and promotions. In fact, gone are the was not at all efficient, what is valid for whichdays when one-size-fits-all large promotions are but this would have been customer, predict theapplied to increase basket size. With an increase impossible to determinein margin pressures, marketers are trying to focus without gaining visibility probability score ofon getting that extra mile. into the market basket. customers respondingMBA analysis will help retailers refine their Using this insight, the to that offer and retailer decides to raise itsapproach to drive an effective loyalty card scheme EDLP on milk. The retailer understand theor online shopping registration by bringing apersonal touch. Combining POS data with other expects sales of milk to customer value gaingeographical level demographic information, drop, and it may even lose from offer acceptance.customer interaction information across channels some customers. But thosesuch as e-commerce, loyalty club Web sites or customers were not profitable; the improvedorder or service hotlines, as well as attitudinal margin on the future milk sales will result indata captured through surveys at points of inter- profits being net-positive.action, is sifted and analyzed to provide valuable MBA empowers merchants to buy smarterinsights to further refine targeting strategy. and strengthen their negotiating position withFurther, predictive models are built on historical vendors by providing the merchants with betterpurchase data, as well as other attribute data, to information about customer buying behavior.add predictability to customer responsiveness to Integrated Solution Framework for Effectivepromotions. This empowers retailers to embrace Decision Makingfocused targeting. Predictive models help retailersto direct the right offer to the right customer Our MBA solution offering is coupled with ansegments/profiles, as well as gain understanding integrated analytics and BI platform calledon what is valid for which customer, predict the iTrackTM that integrates multiple stakeholderprobability score of customers responding to that metrics and views with a robust data model tooffer and understand the customer value gain generate business insights for effective decisionfrom offer acceptance. making. Together they provide:MBA Helps Merchandisers • An integrated view of business metrics across dimensions, with embedded roles andMerchandisers need to see long-term trends to privileges.decide how much to buy and how the assortmentfits into the business model. Here are some • A single version of truth.ways that leading retailers can leverage MBA to • An end-to-end platform and process from dataempower their merchandisers: acquisition to reporting.• Can the retailer sell fast enough to cover car- • Multipledata sources integrated to create a rying costs? comprehensive data mart for mining.• Will the initial markup provide sufficient margins • Visibility to a comprehensive set of business to promote sell-through? metrics.These high-value, high-risk decisions can be sig- • Informed decision making through insights and report comparison.nificantly improved with the customer insightsprovided by MBA. For example, a retailer sells • Report customization. cognizant 20-20 insights 5
  6. 6. This platform allows marketing and merchan- areas. This solution provides agility and flex-dising executives to understand sales patterns, ibility to merchants, marketers, operations andcustomer preferences and buying patterns so others managers within the retail organizationthey can take appropriate actions to improve to analyze relevant information and assimilateproduct sales and margins, and create targeted it with the past as well as what should be theand profitable promotions. It also enables retailers road ahead, by leveraging a common, integratedto extract and slice information in meaningful solution framework/platform.ways to identify challenges or opportunityFootnotes1 “Best Positioned, Grocery Store Kroger,” Adage News, Feb. 23, 2009.2 Undirected data mining seeks patterns or similarities among groups of records without the use of a particular target field or collection of predefined classes.3 Affinity analysis, the heart of market basket analysis, determines which potential purchases go together in a single shopping cart. Retail chains use affinity analysis to plan the arrangement of items on store shelves or in a catalog so that related items often purchased together will be seen together. Affinity analysis can also be used to identify cross-selling opportunities and to design attractive packages or groupings of products and services. Affinity analysis is one simple approach to generating rules from data.About the AuthorNilanjana Singh Chandra supports Business Development initiatives within Cognizant Analytics. Shehas over 14 years of cross-sectoral experience in providing solutions and thought leadership in salesand marketing analytics, brand strategy consulting, campaign management, multichannel closed loopmarketing and advanced promotion response analytics to global companies. Nilanjana can be reached atNilanjana.Chandra@cognizant.com.About Cognizant AnalyticsCognizant Analytics (CA) combines business consulting, in-depth domain expertise, predictive analyticsand technology services to help clients gain actionable and measurable insights and make smarterdecisions that future-proof their businesses. The practice offers comprehensive solutions and servicesin the areas of sales operations and management, product management and market research. CA’sexpertise spans sales force and marketing effectiveness, incentives management, forecasting, segmenta-tion, multichannel marketing and promotion, alignment, managed markets and digital analytics. With itshighly experienced group of consultants, statisticians and industry specialists, CA prepares companiesfor the future of analytics through its innovative “Plan, Build and Operate” model and a mature “GlobalPartnership” model. The result: solutions that are delivered in a flexible, responsive and cost-effectivemanner. http://www.cognizant.com/enterpriseanalytics.About CognizantCognizant (NASDAQ: CTSH) is a leading provider of information technology, consulting, and business process out-sourcing services, dedicated to helping the world’s leading companies build stronger businesses. Headquartered inTeaneck, New Jersey (U.S.), Cognizant combines a passion for client satisfaction, technology innovation, deep industryand business process expertise, and a global, collaborative workforce that embodies the future of work. With over 50delivery centers worldwide and approximately 150,400 employees as of September 30, 2012, Cognizant is a member ofthe NASDAQ-100, the S&P 500, the Forbes Global 2000, and the Fortune 500 and is ranked among the top performingand fastest growing companies in the world. Visit us online at  www.cognizant.com or follow us on Twitter: Cognizant. World Headquarters European Headquarters India Operations Headquarters 500 Frank W. Burr Blvd. 1 Kingdom Street #5/535, Old Mahabalipuram Road Teaneck, NJ 07666 USA Paddington Central Okkiyam Pettai, Thoraipakkam Phone: +1 201 801 0233 London W2 6BD Chennai, 600 096 India Fax: +1 201 801 0243 Phone: +44 (0) 20 7297 7600 Phone: +91 (0) 44 4209 6000 Toll Free: +1 888 937 3277 Fax: +44 (0) 20 7121 0102 Fax: +91 (0) 44 4209 6060 Email: inquiry@cognizant.com Email: infouk@cognizant.com Email: inquiryindia@cognizant.com©­­ Copyright 2012, Cognizant. All rights reserved. No part of this document may be reproduced, stored in a retrieval system, transmitted in any form or by anymeans, electronic, mechanical, photocopying, recording, or otherwise, without the express written permission from Cognizant. The information contained herein issubject to change without notice. All other trademarks mentioned herein are the property of their respective owners.

×