PRIZM and Segmentation for Any Size Budget Mike Mancini Michelle FauntLeRoy Data Product Management Nielsen Claritas, The Nielsen Company Confidential & Proprietary • Copyright © 2009 The Nielsen Company Agenda • Segmentation and Precision Marketing • Getting the Most Out of Your Segmentation Budget – Case Study Example • • • • Leverage Segmentation Throughout Your Organization Maximizing Benefits of List Selects Using Segmentation in Modeling Plan to Grow with Nielsen Segment 43 33 24 25 March 26, 2009 Page 2 Confidential & Proprietary Copyright © 2009 The Nielsen Company Segmentation and Precision Marketing • Every industry wants to attract and retain as many profitable customers as possible • Segmentation is the process of grouping people or businesses based on shared traits/activities/behaviors • Customer knowledge helps steer improved operations Segment 43 33 24 25 March 26, 2009 Page 3 Confidential & Proprietary Copyright © 2009 The Nielsen Company #24 Segment Product/s with You Customer Knowledge 33 24 25 43 perspectives Full Wallet of Products Financial, Retail, Restaurant, Automobiles Channels Location, Phone, Online, Mail Lifestyles and Demographics Media Usage HH Traits, Travel, TV, Radio, Print, Hobbies, Sports, Outdoor, Internet, Music Handheld March 26, 2009 Neighborhood Composition Consumer Files Penetration, Vicinity, Local specific Finding Like-Type Consumers Page 4 Confidential & Proprietary Copyright © 2009 The Nielsen Company Nielsen Claritas Segmentation Products • PRIZM® is optimized by urbanization and affluence for a multitude of consumer behaviors including media usage, automobile ownership, communications, financial and insurance, and retail behavior • P$YCLE® is optimized for Income Producing Assets, financial and insurance behavior and consumption; identifies wealth • ConneXions® is optimized for technology adoption and communications behavior March 26, 2009 Page 5 Confidential & Proprietary Copyright © 2009 The Nielsen Company Segmentation Assignments • Segment Assignment – – – – – Household: one segment for one household based on name and address ZIP+6: typically single household ZIP+4: single street face, 3-5 households Census block group: market base, maps ZIP Code: ultimate fill-in, maps • Segment Distributions – Each geography has a unique composition based on its underlying Z6/Z4 assignments – Market ranking, planning, targeting • Behavioral Profiles – Propensity by segment for thousands of behaviors from Nielsen, MRI, NPD, Polk, Simmons and more – Segmentation Link Partner Network provides segmentation on wide variety of hundreds of market research surveys and service bureaus March 26, 2009 Page 6 Confidential & Proprietary Copyright © 2009 The Nielsen Company Segmentation Market Share 20 of the top 25 U.S. Banking and Holding Companies 6 of the top 10 Investment Companies 4 of the top 5 Property and Casualty Insurance Companies The top 4 Life Insurance Companies 13 of the top 15 Retailers 6 of the top 10 Specialty Retailers 7 of the top 10 Restaurant Chains 8 of the top 10 Newspaper Groups All National Long Distance Providers 6 of the top 8 Wireless Providers 5 of the top 10 Cable Network Groups 5 of the top 10 Energy and Utility Providers March 26, 2009 Page 7 Confidential & Proprietary Copyright © 2009 The Nielsen Company Fully Utilize Your Current Resources • Enhance what you know about your customers • Get more value by linking to internal metrics • Customer transaction data provides a wealth of insight • Other secondary segmentation tools are effective for internal communication, helpful when putting together business cases to demonstrate how to get the return on marketing campaigns • It takes effort to get the full benefit of your segmentation resources • Don’t expect simple solutions to complex questions • Don’t overly complicate solutions • Spend the time to do the analysis March 26, 2009 Page 8 Confidential & Proprietary Copyright © 2009 The Nielsen Company Getting the Most Out of Your Segmentation Budget March 26, 2009 Page 9 Confidential & Proprietary Copyright © 2009 The Nielsen Company Case Study: Selling Cupcakes to Customers Coffee Shop Up-sell Strategy • High end coffee shop wants to sell gourmet cupcakes to existing client base to increase revenue • Identify target groups within customer base to add new product • Appeal to different prospects within target groups Analysis and Measurement • Customer profile • Customer + revenue profile • Measuring lift • Creating target groups • Add more complex metrics March 26, 2009 Page 10 Confidential & Proprietary Copyright © 2009 The Nielsen Company PRIZM Customer Profile Buy from Coffee Shop, 1 month • PRIZM provides insight into households that are most likely, average, and least likely to patronize my high end coffee shop – Code customers (using directory or online or have Nielsen Claritas or partner code) – Use secondary research data (MRI, Simmons, Scarborough, etc.) March 26, 2009 Page 11 Confidential & Proprietary Copyright © 2009 The Nielsen Company PRIZM Customer Profile + Revenue Customers vs. U.S. Households Customer Revenue vs. U.S. Households • Customers by PRIZM plus Customer Revenue by PRIZM • Revenue by PRIZM enhances insight into value March 26, 2009 Page 12 Confidential & Proprietary Copyright © 2009 The Nielsen Company PRIZM Customer Data Adds Lift 100 90 22 segments 34% HH 67% customers 2.00 is max lift 80 % total customers 70 30 segments 45% HH 75% customers 1.69 is max lift 60 50 40 30 PRIZM profile of customers 20 PRIZM customers + PRIZM revenue Random Households 10 0 0 10 20 30 40 50 60 70 80 90 100 % total households March 26, 2009 Page 13 Confidential & Proprietary Copyright © 2009 The Nielsen Company Quintiles (5 groups) by PRIZM 50 45 45 Random Households 42 40 PRIZM profile of customers 35 PRIZM customers + PRIZM revenue % total 30 26 25 20 25 22 20 22 21 18 16 14 15 9 10 9 4 5 5 0 Quintile: 1 Quintile: 2 Quintile: 3 March 26, 2009 Quintile: 4 Page 14 Quintile: 5 Confidential & Proprietary Copyright © 2009 The Nielsen Company Target Group 1 Target Group Code Households % Comp Customers % Comp 1 02 Blue Blood Estates 1.0 2.5 251 2.8 284 1 04 Young Digerati 1.2 3.0 248 3.4 279 1 05 Country Squires 1.9 4.5 232 4.6 240 1 06 Winner's Circle 1.1 2.9 265 3.1 278 1 07 Money & Brains 2.0 4.8 238 5.9 295 1 13 Upward Bound 1.6 3.7 233 4.6 288 1 16 Bohemian Mix 1.8 2.9 163 3.0 172 1 17 Beltway Boomers 1.0 1.8 183 1.8 191 1 18 Kids & Cul-de-sacs 1.6 3.1 189 3.1 192 1 29 American Dreams 2.2 3.7 170 3.2 149 1 34 White Picket Fences 1.2 1.5 124 1.4 111 16.6 34.3 207 37.0 223 PRIZM Segment 1 Coffee Cupcakes and Kids Index Revenue % Share Index • Coffee, Cupcakes & Kids represents 16.6% of households, 34% of customers, and 37% of revenue • High value target group includes customers with children March 26, 2009 Page 15 Confidential & Proprietary Copyright © 2009 The Nielsen Company Target Group 2 Target Group Code Households % Comp Customers % Comp 2 01 Upper Crust 1.5 3.2 212 3.4 220 2 03 Movers & Shakers 1.6 4.3 265 4.3 268 2 08 Executive Suites 0.9 1.7 188 1.6 180 2 09 Big Fish, Small Pond 2.3 2.7 120 2.1 92 2 10 Second City Elite 1.2 1.7 142 1.9 157 2 11 God's Country 1.6 3.1 197 2.7 173 2 12 Brite Lites, Li'l City 1.5 3.4 229 4.4 292 2 15 Pools & Patios 1.3 1.9 141 2.1 163 2 19 Home Sweet Home 1.8 2.5 133 2.9 156 2 23 Greenbelt Sports 1.5 2.2 151 1.8 123 2 24 Up-and-Comers 1.2 1.6 129 1.1 94 16.4 28.2 172 28.3 173 2 Cupcake Couples PRIZM Segment Index Revenue % Share Index • Cupcake Couples represents 16% of households, 28% of customers, and 28% of revenue • High value target group includes couples without children at home March 26, 2009 Page 16 Confidential & Proprietary Copyright © 2009 The Nielsen Company On and Off-Target P$YCLE Target Group Name Households % Comp Customers % Comp Index Revenue % Share Index 1 Coffee Cupcakes and Kids 16.6 34.3 207 37.0 223 2 Cupcake Couples 16.4 28.2 172 28.3 173 3 Special Treat 21.1 20.4 96 19.8 94 4 Off Target 45.9 17.1 37 15.0 33 100.0 100.0 100 100.0 100 • Coffee, Cupcakes & Kids bring in over twice the revenue as the average household, and they are my best cupcake targets • Cupcake Couples another valuable target • Special Treat are average customers; let’s sell them a cupcake treat! • Assess value of marketing to Off-Target group March 26, 2009 Page 17 Confidential & Proprietary Copyright © 2009 The Nielsen Company Use Internal Metrics P$YCLE Target Group Name Households % Comp Customers % Comp Index Revenue % Share Index Average Monthly Profit 1 Coffee Cupcakes and Kids 16.6 34.3 207 37.0 223 $4,904 2 Cupcake Couples 16.4 28.2 172 28.3 173 $3,751 3 Special Treat 21.1 20.4 96 19.8 94 $2,622 4 Off Target 45.9 17.1 37 15.0 33 $1,989 100.0 100.0 100 100.0 100 • Use internal metrics like monthly revenue or average monthly profit to plan marketing campaigns • Adding customer data is valuable and free! March 26, 2009 Page 18 Confidential & Proprietary Copyright © 2009 The Nielsen Company Leverage Segmentation Throughout Your Organization March 26, 2009 Page 19 Confidential & Proprietary Copyright © 2009 The Nielsen Company Leverage Segmentation Everywhere • Product Optimization – Offer the right products to the right target groups • Sales and Marketing Channel – Optimize channel mix by understanding customers’ channel preference • Site Selection – Identify sites with best segments concentrated in surrounding trade areas • Finance or Measurement Groups – Use segmentation to find hidden gems in the low lifetime value groups – Measure cost of acquisition versus the lifetime value • Customer Service – Offer high value service for most valuable target groups – Route inbound calls based on target groups • Media Strategy and Advertising – Share target groups with ad agency so targets are not over simplified into age and gender (female age 18-45) March 26, 2009 Page 20 Confidential & Proprietary Copyright © 2009 The Nielsen Company Maximizing Benefits of List Selects March 26, 2009 Page 21 Confidential & Proprietary Copyright © 2009 The Nielsen Company Using Lists for Cost Effective Lead Generation • Generating a list using PRIZM, P$YCLE or ConneXions will enable you to get a more precise target than demographic selects alone • Combined with demographics or a specific behavior, PRIZM will provide more performance lift • PRIZM is on a variety of compiled lists • Save money by knowing your targets before ordering lists – Use segmentation ground counts to size the opportunity – Use segmentation suppression • Test and control March 26, 2009 Page 22 Confidential & Proprietary Copyright © 2009 The Nielsen Company Data Choices for List Selects • Demographics – Demographic data about consumers collected from multiple sources • Behavior – Data collected via magazine inserts, surveys, web sites and product registration data such as: go to the theater, play tennis, take vitamins • Financial Behavior Data – Credit cards, personal loans, ACE, other consumer financial behavior • Segmentation – Household level for precision – ZIP+4 level for cost effective saturation • Other – Telephone number, address, and residency information, email addresses March 26, 2009 Page 23 Confidential & Proprietary Copyright © 2009 The Nielsen Company Economic Environment Impacts Email • • • • • Increasing postage rates Rising print and paper costs Decreasing consumer response rates Increasingly complex regulatory environment Marketers reporting: – Motivating customers to shop is becoming increasingly complex – New customer acquisition and retention is not as dependable as it has been in the past – Marketing across multiple channels is top priority – Targeting and segmentation is key for campaign success March 26, 2009 Page 24 Confidential & Proprietary Copyright © 2009 The Nielsen Company Nielsen Claritas Partners with Top Service Bureaus • Email Append (B2C and B2B) – Append email addresses to existing customer addresses • Reverse Email Append (B2C and B2B) – Append client’s file of email addresses and return name and postal address • Email Update (email change of address) – People change their email address regularly, therefore updates are recommended • Prospect Match and Deploy – Select a prospect list. The service bureau deploys: – P$YCLE and PRIZM selections available – Results are shared after campaign deployment March 26, 2009 Page 25 Confidential & Proprietary Copyright © 2009 The Nielsen Company Using Segmentation in Modeling March 26, 2009 Page 26 Confidential & Proprietary Copyright © 2009 The Nielsen Company Using Segmentation in Regression Modeling Produces Results • The winner of the Direct Marketing Association Analytic Challenge used Nielsen Claritas segmentation products • • • Use segmentation products appropriately A transformation of results needed in most instances Use of appropriate segmentation product can result in more accurate predictions How should I model with segmentation? March 26, 2009 Page 27 Confidential & Proprietary Copyright © 2009 The Nielsen Company Transform The Variable • Segment codes are not continuous variables - Segment codes correlate with affluence but should not be treated as a continuous variable - Do not put the segment code itself into the model • Segment codes should be transformed into useful statistics like - Average balance - Market penetration - Rank of balance - Rank of penetration As tempting as it is – DO NOT use the segment code itself in a regression model Better options exist! March 26, 2009 Page 28 Confidential & Proprietary Copyright © 2009 The Nielsen Company Choose From Available Products • Choose the most appropriate segmentation product for best results Standard Deviation = $518,406 – – Standard Deviation = $323,402 – – Understand how products are created to avoid pitfalls Correlation/linear trends are not important Separation is important Match the product’s purpose with goals – (i.e. cell phones with technology segmentation product, 401k accounts with a financial product) Choose P$YCLE because of distinctiveness and product specifications March 26, 2009 Page 29 Confidential & Proprietary Copyright © 2009 The Nielsen Company Segmentation Modeling Tips 1. Break up the problem – keep penetration separate from profit per purchase 2. Choose the segmentation product appropriately 3. Transform the segmentation product – don’t use the segment code itself 4. Incorporate in-house information on the client (i.e. number of purchases) 5. Test for over-fitting the model 6. Understand potential correlations 7. Consider interaction terms 8. Apply segmentation parameters to missing observations 9. Adjust/remove the unknown segments/clusters March 26, 2009 Page 30 Confidential & Proprietary Copyright © 2009 The Nielsen Company Plan to Grow with Nielsen March 26, 2009 Page 31 Confidential & Proprietary Copyright © 2009 The Nielsen Company Nielsen Segmentation Benefits • PRIZM, P$YCLE and ConneXions are linked to Nielsen currencies, third-party databases, and client data • Nielsen surveys available in various formats with segmentation : – – – – – – – – Nielsen Media Research Local / National / Hispanic People Meter Nielsen Online NetView / MegaPanel Nielsen Mobile Insights / Total Comm Survey Nielsen IAG Advertising Effectiveness Nielsen Consumer Home and Retail CPG Scanner Data Nielsen Claritas Market Audit Survey Nielsen Claritas Insurance Audit Survey Nielsen Claritas Convergence Audit Survey • Speed to market = cost and time efficiency March 26, 2009 Page 32 Confidential & Proprietary Copyright © 2009 The Nielsen Company Breadth of Consumer Insights – Link Partner Network Media NPD Group Apparel & Restaurant Retail Nielsen Claritas Convergence Audit Telecom Auto General Lifestyle/ Consumer Research Direct Marketing Services March 26, 2009 Page 33 Confidential & Proprietary Copyright © 2009 The Nielsen Company Thank you – Products Mentioned – Nielsen Claritas PRIZM – Nielsen Claritas P$YCLE – Contact Us – Nielsen Claritas ConneXions – info.cds@nielsen.com – Nielsen Claritas Segmentation – 800-737-7429 – www.claritas.com March 26, 2009 Page 34 Confidential & Proprietary Copyright © 2009 The NielsenCompany Company Confidential & Proprietary • Copyright © 2007 The Nielsen