Highlights from DMA andcreating communicationwith left & right side brainBo Sannung, Nordic director of Centre of Excellen...
Digital Dashboard    Competitor Earned                                 Bought   Owned    You    Comp 1    Comp 2    Comp 3...
9 NOVEMBER 2012   2012 COPYRIGHT SAS INSTITUTE3
“The Emergence Of Customer Experience Management Solutions”Delivering Cross-Touchpoint Customer Experiences Drives Need Fo...
Channel / Media Trustworthiness                                Epsilon Targeting                 5     Step 3: Multichanne...
Time to Get it Right       Treat customers the way we want to be treated… …and generate double-digit increases in response...
On any given day, the                           Message Overload   customer will beexposed to nearly 3,000   media message...
“Per the DMA,          93% of                                             Let’s Define “Relevance”marketers using multiple...
5 Principles of Multichannel Marketing5     5: Customer Lifecycle Marketing: 1) Communications must be     deployed at app...
Don’t   re-engineer                        This economy and social                        media have profoundly       your...
VoC Learnings        Question                           Answer Which has more impact on           Engagement/Relationship ...
VoC Learnings       Question                                AnswerWhich is a more significant        Engagement/Relationsh...
VoC Learnings4. The Importanceof Community    Per McKinsey research, as cited in the Wall Street Journal, people who    pa...
 Community-driven,    online marketer      specializing in    T-shirts designed     by members of       According to the ...
As a result, customers and       prospects view     personalization asthe next step in a company‟s   •   Personalization i...
Meaningful Personalization Customers are also savvy regarding the  type of personalization they want. They want it to be...
Customer EngagementWe at Academic PCS would like to see Flash in 64-bit version as soon as possible.This is very important...
Customer Engagement“…customers with highest feedback scoresalso had the greatest lifetime values.Differences in lifetime v...
VoC Learnings1. Providing Value                                        “Self-serve makes it                               ...
VoC Learnings2. Relationships                                         “The fastest way to                                 ...
VoC Learnings3. The Web                                     “When you tell me to go     “I don’t just want to             ...
Global CMO Survey:                                       For 42% of CMOs:                           “…representing the voi...
Using Voice of Customerto Increase Engagement &               Drive Sales
Who We Are•   Launched in 2007•   Flash-sales category founder in    US and leader with over $500MM    revenue•   Curate b...
How We Use VoC•   Measure VoC: Utilize various sources including purchase,    browsing, waitlist, email click, as well as ...
How We Use VoC•   Personalization•   Merchandising•   Segmentation•   Policies•   Loyalty•   Social Engagement•   Customer...
How We Use VoC - Personalization•   We produce 2,500+ versions of personalized emails•   VoC drivers include: purchase,   ...
How We Use VoC - Personalization•   Favorite brands are appended to    one’s profile and help drive e-mail    personalizat...
How We Use VoC - Personalization“I don’t buy men’s goods on gilt.com becausethey sizing information isn’t good enough, you...
How We Use VoC -                  Merchandising •      Use Facebook’s Face Off Application to        empower members to cu...
How We Use VoC - Merchandising •      Announce winning selection on Wall and drive        to sale featuring the “Facebook ...
How We Use VoC -           Merchandising•   Use Facebook’s Face Off Application to    empower members to curate sales     ...
How We Use VoC - Merchandising  •      Crowd source ideas involving fans to create new         productsFans vote          ...
How We Use VoC - Merchandising•   Crowd source ideas involving fans to create new products          KPI                   ...
How We Use VoC - Segmentation    Brand Seekers                 Self-Expressionists “I am always shopping to     “My style ...
How We Use VoC - Policies•   Online Panels, Customer Service Feedback and    Research told us that Shipping Fees were bigg...
How We Use VoC - Policies  •       Measure the impact of new shipping fee policies on Consumer Awareness          and Sati...
How We Use VoC – Loyalty•   Quarterly member dinners provide “multi-    channel” insights and ensure that strategies    an...
How We Use VoC – Loyalty•   Quarterly member dinners…       KPI                  RESULTSSpending by Best   10-15X Higher t...
How We Use VoC – Social      Engagement•   Senior Officers engage with members                                          41
How We Use VoC – Social Engagement•   Senior Officers engage with members                                             42
How We Use VoC – New Businesses•       Gilt Taste idea originated        from Gilt Employee•       Business launched      ...
Bringing High Quality Customer Service Into The        Social Arena•   Authenticity: Team is encouraged to be    themselve...
Bringing High Quality Customer Service Into       The Social Arena•   Follow Up: Social feeds are tagged for follow    up,...
Bringing High Quality Customer        Service Into The Social Arena•   Transparency: All postings are valid               ...
Bringing High Quality Customer Service Into The            Social Arena• Surprise & Delight:   •   Per CSR Feedback, women...
5 Tactics to Leverage VoC1.   Listen and Invite Feedback2.   Respond, always, and make responses personal3.   Drive Awaren...
5 Tactics to Leverage VoCBONUS:•   Do Not Wait! to hear from your customers•   Recently launched an outreach program to   ...
5 Tactics to Leverage VoC•    Best Customer Outreach Call Program           KPI                         RESULTSReactivatio...
Max=(r2+k3)*(TIME-(n+g))                                                                     51         Copyright © 2011, ...
DRIVING VALUE FROM CUSTOMER RELATIONSHIPS IS     INCREASINGLY COMPLEXCustomers &ProspectsOffers,Services &PricingChannels ...
ACCENTURE RESEARCH 2011                                                            COMPANIES THAT INVEST IN ADVANCED77% OF...
54Copyright © 2011, SAS Institute Inc. All rights reserved.
Entry Points to ITV Player                                                                       Multiple                 ...
Next Best Product - ExamplesCase Study: Erste Bank Group                                                                  ...
The NBO 4 main components      Customer Behavior                                                             Importance   ...
Personalized “Next Best Product” offer executed across…Branch/Advisor                                nbp                  ...
How to get started ?Value                                                            Phase 3                              ...
LønsomhedsopgørelseOmsætning:Forbrugs DB                          100Abonnement DB                                        ...
Lønsomhed på kunde                                                                         61             Copyright © 2011...
62Copyright © 2011, SAS Institute Inc. All rights reserved.
4 slides on analytics – that’s it !                 Copyright © 2011 SAS Institute Inc. All rights reserved.
Why Predictive Modeling?          100                        90                        80 Cuml Gains( Caputre)            ...
Predictive Modeling TechniquesDecision TreeAttempts to split a population into subgroups that tend to be morehomogeneous t...
Predictive Modeling TechniquesNeural Networks Data can be processed in parallel and complex relationships can befound quic...
Develop Treatment Strategy -- Example                                                   Three Tier Approach: 1) Predictive...
Applying Predictive Models to Marketing StrategyMarketing Objective                Question                               ...
Customer insightin action at Tesco             Copyright © 2011 SAS Institute Inc. All rights reserved.
Tesco Overview Formed in 1924 The UK’s largest food retailer Operating stores in all formats – convenience,     high st...
Tesco and SAS Currently use SAS across the business to help Select locations Plan investment in refurbishments Margin ...
Tesco- a truly customer focussed business                                                          “Our mission is to earn...
Translate data in to a clear picture of acustomer        Data                                                             ...
Shopping behaviour can explain a lot…                Data                                                                 ...
Step 1: develop a meaningful customersegmentation                                                                         ...
The Nectar Marketing CommunicationsSegmentation                               Engaged Enthusiasts                  Bonus S...
Step 3: Overlaying financial data allowsfor improving the allocation of customermarketing investment                      ...
Cross - upsale            Copyright © 2011 SAS Institute Inc. All rights reserved.
How to get started ?Value                                                            Phase 3                              ...
Kunde scoring - Produkt        Kunde    Prod A                                         Prod B      Prod C          1      ...
Kunde scoring - respons        Kunde   Kamp A                                         Kamp B       Kamp C          1      ...
Net lift                                                                             NET LIFT                           “W...
Aviva online experience            Copyright © 2011 SAS Institute Inc. All rights reserved.
How to get started ?Value                                                            Phase 3                              ...
Churn - Score        Kunde   Churn Q1                                     Churn Q2       Churn Q3          1            90...
Churn - messageRDM Generated offer ormessage                                                                              ...
Opsalg ved bookingRetrieve customer andflight informationCheck miles balance tosee if free flightavailableIf free flight n...
Analytics embedded in customer process          “Conversational Data”                                                     ...
Embedded analytics           Copyright © 2011 SAS Institute Inc. All rights reserved.
A Vast Universe of Data What are they requesting?                                                                      How...
Capture and Analyze a SmallPortion   What are they requesting?                                                            ...
We Miss the Elegant Patterns                                                                           92          Copyrig...
Customer experience across datasilos                                                                                      ...
AD-HOC ANALYTICAL                                                                             SEGMENTATION                ...
Pradigmeshift inMarketing processes           Copyright © 2011 SAS Institute Inc. All rights reserved.
Campaign Management Process                 Campaign                Management                 Exclusion                  ...
Is S3 the Optimal Solution?                      Setting the Max Offers                      constraint at 512K delivers  ...
Results - 3. Switch Focus to ProfitabilityEffect of using SAS MO                                                          ...
Campaign Management Process                 Campaign                Management                 Exclusion                  ...
offer optimization - a complex problem              choices and constraints:              •   many customers, offers, chan...
and actually what happens is….•   subjective planning and decision making•   poor ROI from marketing campaigns    •   low ...
Why optimization outperforms other decisioning techniques                                                                 ...
Simple Example – Prioritizing Campaigns•       Model scores define response probability for each campaign•       Probabili...
Level 1: Campaign Prioritization •   Constraints: 1 customer: 1 campaign / campaign - 3 customers •   Campaigns assigned p...
Level 2: Customer Offer Prioritization •   Constraints: 1 customer: 1 campaign / campaign - 3 customers •   Priority assig...
Level 3: Optimization Approach •   Constraints: 1 customer: 1 campaign / campaign - 3 customers •   Optimisation evaluates...
Campaign Management Process                 Campaign                Management                 Exclusion                  ...
Cultural Impact Post OptimizationMinimal impact on Campaign Management Process                           Campaign         ...
Tips to get started Get started Know your customers motivation to engage with you and  do a communication segmentation ...
AD-HOC ANALYTICAL                                                                              SEGMENTATION               ...
«Marketing Treatment Strategy» - veien til økt kundeverdi
Upcoming SlideShare
Loading in …5
×

«Marketing Treatment Strategy» - veien til økt kundeverdi

967 views
894 views

Published on

Foredragsholder er Bo Sannung, Head of Center of Excellence, Customer Intellegence i SAS Institute: Marketing Treatment Strategy handler om hvordan man gir hver enkelt kundesegment spesifikke tilbud for å optimalisere kundeverdien.

Hvordan individualisere håndteringen av kunder og samtidig optimalisere kundeverdien? Gjennom eksempler vil du få innsikt i Marketing Treatment Strategies - både den analytiske delen med kundeinnsikt, segmentering og prediksjon, men også hvordan omsette innsikten i praksis. Under innlegget vil du få praktiske råd og verktøy. I tillegg vil Sannung også dele trender og tendenser fra DMA 2012 konferansen.
http://www.dma12.org

Bo Sannung er Nordisk direktør for Customer Intelligence i SAS Institute. Bo Sannung har bakgrunn fra byråbransjen og store nordiske selskaper innenfor salg, marked, CRM og analyse. Sannung underviser også på Copenhagen Business School og CRM Akademiet.

Published in: Business
0 Comments
2 Likes
Statistics
Notes
  • Be the first to comment

No Downloads
Views
Total views
967
On SlideShare
0
From Embeds
0
Number of Embeds
0
Actions
Shares
0
Downloads
6
Comments
0
Likes
2
Embeds 0
No embeds

No notes for slide

«Marketing Treatment Strategy» - veien til økt kundeverdi

  1. 1. Highlights from DMA andcreating communicationwith left & right side brainBo Sannung, Nordic director of Centre of Excellence - IMM 9 NOVEMBER 2012 2012 COPYRIGHT SAS INSTITUTE1
  2. 2. Digital Dashboard Competitor Earned Bought Owned You Comp 1 Comp 2 Comp 3 Comp 4 9 NOVEMBER 2012 2012 COPYRIGHT SAS INSTITUTE2
  3. 3. 9 NOVEMBER 2012 2012 COPYRIGHT SAS INSTITUTE3
  4. 4. “The Emergence Of Customer Experience Management Solutions”Delivering Cross-Touchpoint Customer Experiences Drives Need For New Capability
  5. 5. Channel / Media Trustworthiness Epsilon Targeting 5 Step 3: Multichannel Marketing
  6. 6. Time to Get it Right Treat customers the way we want to be treated… …and generate double-digit increases in response and revenue Overview
  7. 7. On any given day, the Message Overload customer will beexposed to nearly 3,000 media messages. They will pay attention to 52. They will positively remember 4. The chance they will remember your ad is 0.013%!D. Mastervich, VP, Sales Strategy, U.S. Postal Service, VDP Conference Presentation 7 Step 3: Multichannel Marketing
  8. 8. “Per the DMA, 93% of Let’s Define “Relevance”marketers using multiplechannels have attempted to 1. Right message. integrate their messaging. Only 27.4% of these said their efforts 2. Right time. are „effective‟. . .” DMA Report, “Rowing 3. Right person. as One: Integrated Marketing Today,” 4/11 4. Delivered per that individual’s media preferences. Integrated, multichannel irritation! Without this, all we have achieved is. . . 8 Step 3: Multichannel Marketing
  9. 9. 5 Principles of Multichannel Marketing5 5: Customer Lifecycle Marketing: 1) Communications must be deployed at appropriate points in the buying cycle, and 2) Contacts should be driven by opt-in preferences.4 4: Re-conceive Inbound as a high value customer interface. By definition, Inbound callers are more 1) Qualified, and 2) Likely to spend.3 3: Synchronize your multichannel mix with and value. precision 2: Create processes for generating feedback from your social media2 channels and your sales and service reps. This will provide ongoing qualitative and quantitative VoC guidance.1 1: Start with the Customer (VoC). Step 3: Multichannel Marketing
  10. 10. Don’t re-engineer This economy and social media have profoundly your changed buyer’s priorities relationship and expectations. marketing If you have notstrategies from recalibrated strategies within the past 12 the isolation months, you are out of of your sync with your customers. conference VoC insights ensure you room. . . develop truly customer- focused strategies to drive relevance and revenue. 10 Step 1: VoC Research
  11. 11. VoC Learnings Question Answer Which has more impact on Engagement/Relationship retention and repeat strength has 12 times more purchases; influence on retention and Customer Satisfaction or repeat purchases than Customer Satisfaction. Engagement/Relationship? Satisfaction is a minimum expectation. 11 Step 1: VoC Research
  12. 12. VoC Learnings Question AnswerWhich is a more significant Engagement/Relationship strength has 18 times more influence on driver of word of mouth word of mouth recommendations recommendations; than Satisfaction. Customer Satisfaction or This has profound implications forEngagement/Relationship? re-allocating greater budget for Retention/ Relationship building. 12 Step 1: VoC Research
  13. 13. VoC Learnings4. The Importanceof Community Per McKinsey research, as cited in the Wall Street Journal, people who participate in an effective online community, return to a site: times as often times as long This represents a 45 time increase in loyalty! Step 1: VoC Research
  14. 14.  Community-driven, online marketer specializing in T-shirts designed by members of According to the Sloan Management Review: the community. 95% of those purchasing from Threadless.com have Community is made up voted and posted comments…before making a of 3 groups: purchase. 1. Purchasers 2. Designers 3. Reviewers Results: • Over 1 million users, • Over $30 million dollars in annual sales, • Approximately 30% margins. Step 1: VoC Research
  15. 15. As a result, customers and prospects view personalization asthe next step in a company‟s • Personalization is viewed as a service and benefit, not commitment to service just a sales tool. excellence. • Online shoppers view personalization as a requirement for their preferred shopping venues, rather than as simply a perk. • Many BtoB decision-makers use Amazon as their point of reference regarding expectations for BtoB personalization. • BtoB and BtoC marketers have to at least match Amazon! 15 Step 2: Opt-In Engagement
  16. 16. Meaningful Personalization Customers are also savvy regarding the type of personalization they want. They want it to be more than just transaction- based. “I expect more than just ‘we’ve looked at everything you’ve bought over the last X years and this is what we think you’ll like’. With today’s technology, I expect much more than that!” Step 2: Opt-In Engagement
  17. 17. Customer EngagementWe at Academic PCS would like to see Flash in 64-bit version as soon as possible.This is very important creating and taking advantage of current hardware technologies. Step 1: VoC Research
  18. 18. Customer Engagement“…customers with highest feedback scoresalso had the greatest lifetime values.Differences in lifetime value between customers withlowest and highest feedback scores ranged from:43% among retail customers to 288% among keybusiness accounts." Forrester Research, 12/8/11 Step 1: VoC Research
  19. 19. VoC Learnings1. Providing Value “Self-serve makes it easy for you, not the “Don’t just sell me the customer.” service. Provide ongoing value at key times.” “Email blasts do not “The quality of your equal ‘relationships’.” service is key to how we judge you.” Step 1: VoC Research
  20. 20. VoC Learnings2. Relationships “The fastest way to be forgotten is to buy from you.” “We buy. You disappear without a trace. Oh, except for the monthly bills.” “Relationship? You guys are about ‘buy and die’!” Step 1: VoC Research
  21. 21. VoC Learnings3. The Web “When you tell me to go “I don’t just want to “An easy navigation and to the web for service, transact. I want to commerce process is a especially when I am connect with your minimal competency. . . growing old waiting for company, your brand You better be at least as a phone rep, what I hear and your community.” good as Amazon.” is, ‘Go. . . help yourself.”In Step 4, we’ll analyze the site BtoB magazine ranked #1, and see how it compares, per VoC Research findings. Step 1: VoC Research
  22. 22. Global CMO Survey: For 42% of CMOs: “…representing the voice of the customer is one of the most critical factors in ensuring personal success as a marketer”. “CMOs and their peers understand that the real challenge is …to become the experts of the customers…They must understand what customers represent for the whole organization to help shape the strategy for the overall business.” -- Luca Paderni, VP and Principal Analyst, Forrester. Heidrick & Struggles and Forrester Research, 1/23/12 Step 1: VoC Research
  23. 23. Using Voice of Customerto Increase Engagement & Drive Sales
  24. 24. Who We Are• Launched in 2007• Flash-sales category founder in US and leader with over $500MM revenue• Curate broad range of daily sales• Evolved beyond women’s fashion to Men, Home, Kids, Travel, Food & local offerings 25
  25. 25. How We Use VoC• Measure VoC: Utilize various sources including purchase, browsing, waitlist, email click, as well as an advisory panel to get member feedback• Share VoC Insights Internally: Weekly presentation by the Customer Service team to share VoC insights to senior management• Disposition Reporting: To keep middle-management up to date• Customer First Experience: mandatory experience for all Gilt Employees focused on connecting employees with actual VoC 26
  26. 26. How We Use VoC• Personalization• Merchandising• Segmentation• Policies• Loyalty• Social Engagement• Customer Service• Launch of New Businesses 27
  27. 27. How We Use VoC - Personalization• We produce 2,500+ versions of personalized emails• VoC drivers include: purchase, browsing, email click, and brand preference 28
  28. 28. How We Use VoC - Personalization• Favorite brands are appended to one’s profile and help drive e-mail personalization 29
  29. 29. How We Use VoC - Personalization“I don’t buy men’s goods on gilt.com becausethey sizing information isn’t good enough, youhave only general size information, you need tohave brand specific size info.” Clothing by A.P.C is cut on the slimmer side of the sportswear spectrum, making for a modern European fit … 30
  30. 30. How We Use VoC - Merchandising • Use Facebook’s Face Off Application to empower members to curate salesWhich handbag do you like best? Vote for your favorite andwe‟ll feature it as our Facebook Fan Pick in the Kooba salestarting Thurs. 3/22 at noon ET. Click below the play button below to vote right from your newsfeed. 31
  31. 31. How We Use VoC - Merchandising • Announce winning selection on Wall and drive to sale featuring the “Facebook Fan Pick” labelAnd the Kooba Face-Off winner is…the Maci, with 206votes! Find the Maci in today‟s Kooba sale along withother styles we love from the line: http://gi.lt/GKofqP 32
  32. 32. How We Use VoC - Merchandising• Use Facebook’s Face Off Application to empower members to curate sales KPI RESULTSEngagement 7X Higher than average postLikes & CommentsUnique 5X Higher than average postImpressionsUnique fans who haveseen the post 33
  33. 33. How We Use VoC - Merchandising • Crowd source ideas involving fans to create new productsFans vote Fans voteon favorite on favorite design color We‟re thrilled to announce that Rebecca Minkoff will be We‟re exited to reveal that the winning sketch producing a handbag exclusively for Gilt Members. Even be produced by Rebecca Minkoff exclusively to better – we want you to be a part of the process. Vote on for Gilt is “Luscious Hobo with Spine Studs”! your favorite design by liking one of the two sketches, and Now‟s your chance to select the handbags the sketch with the most votes will be produced. Be sure color. Vote on one of the swatches below by to tell your friends to vote… liking the picture… And the winning Rebecca Minkoff handbag combination is…”Luscious Hobo” with spine studs in soft leather metallic rose gold. Keep your eyes peeled for this creation, available only to Gilt members. Big thank you to everyone who voted. Winner is shown 34
  34. 34. How We Use VoC - Merchandising• Crowd source ideas involving fans to create new products KPI RESULTSEngagement 27X Higher than average postLikes & CommentsUnique 4X Higher than average postImpressionsUnique fans who haveseen the post 35
  35. 35. How We Use VoC - Segmentation Brand Seekers Self-Expressionists “I am always shopping to “My style is an expression of my keep up with the latest personality. I am always lookingfashions. I own the hottest for inspiration …” brands” 36
  36. 36. How We Use VoC - Policies• Online Panels, Customer Service Feedback and Research told us that Shipping Fees were biggest customer pain point• Verified with quantitative research and testing, then reduced fees 37
  37. 37. How We Use VoC - Policies • Measure the impact of new shipping fee policies on Consumer Awareness and Satisfaction: “In general, the cost of shipping on Gilt is:” Just rightOld Policy 16% 52%New Policy Just right Much too high Somewhat too high Just right Lower than you would expect I dont know enough about the 38 current shipping policies to answer
  38. 38. How We Use VoC – Loyalty• Quarterly member dinners provide “multi- channel” insights and ensure that strategies and policies are on track 39
  39. 39. How We Use VoC – Loyalty• Quarterly member dinners… KPI RESULTSSpending by Best 10-15X Higher than averageCustomers customerChurn Rate 50%+ Lower than average customer 40
  40. 40. How We Use VoC – Social Engagement• Senior Officers engage with members 41
  41. 41. How We Use VoC – Social Engagement• Senior Officers engage with members 42
  42. 42. How We Use VoC – New Businesses• Gilt Taste idea originated from Gilt Employee• Business launched within 5 months• With that speed VoC was crucial to getting it right: • Customer Surveys • Advisory Board • Usability 43
  43. 43. Bringing High Quality Customer Service Into The Social Arena• Authenticity: Team is encouraged to be themselves 44
  44. 44. Bringing High Quality Customer Service Into The Social Arena• Follow Up: Social feeds are tagged for follow up, even if it takes months 45
  45. 45. Bringing High Quality Customer Service Into The Social Arena• Transparency: All postings are valid 46
  46. 46. Bringing High Quality Customer Service Into The Social Arena• Surprise & Delight: • Per CSR Feedback, women often volunteer that they are pregnant. • Team is trained to actively engage with members and empowered to surprise and delight. 47
  47. 47. 5 Tactics to Leverage VoC1. Listen and Invite Feedback2. Respond, always, and make responses personal3. Drive Awareness of VoC in organization, make it core to the Culture4. Start somewhere, you don’t need a lot of resources to begin listening to your Customers5. Follow up, we are sometimes wrong and so are customers 48
  48. 48. 5 Tactics to Leverage VoCBONUS:• Do Not Wait! to hear from your customers• Recently launched an outreach program to proactively re-activate lapsed (best) members “Thank you for reaching out and I look forward to working with you. What fun! " 49
  49. 49. 5 Tactics to Leverage VoC• Best Customer Outreach Call Program KPI RESULTSReactivation +45% vs. control groupIncremental Sales +40% vs. control groupFrom reactivated customers 50
  50. 50. Max=(r2+k3)*(TIME-(n+g)) 51 Copyright © 2011, SAS Institute Inc. All rights reserved.
  51. 51. DRIVING VALUE FROM CUSTOMER RELATIONSHIPS IS INCREASINGLY COMPLEXCustomers &ProspectsOffers,Services &PricingChannels &Business Web E-mail Mail Mobile Print Social Phone Branch ATM Advisor TV Radio Service Finance Collections RiskFunctions $ Checking Credit Cards Loans Mortgages InsuranceProducts Savings Lines Investments 52 Copyright © 2011, SAS Institute Inc. All rights reserved.
  52. 52. ACCENTURE RESEARCH 2011 COMPANIES THAT INVEST IN ADVANCED77% OF HIGH-PERFORMING ANALYTICAL CAPABILITIES OUTPERFORMCOMPANIES HAVE ANALYTICAL THE S&P 500 ON AVERAGE BY 64%CAPABILITIES ABOVE AVERAGE INADEQUATE INFORMATION ACCESS REDUCES KNOWLEDGE WORKERS’ PRODUCTIVITY BY 54%65% OF HIGH-PERFORMING COMPANIESHAVE SIGNIFICANT DECISION-SUPPORTAND ANALYTICAL CAPABILITIES 53 Copyright © 2011, SAS Institute Inc. All rights reserved.
  53. 53. 54Copyright © 2011, SAS Institute Inc. All rights reserved.
  54. 54. Entry Points to ITV Player Multiple entry points to ITV Player 55Copyright © 2011, SAS Institute Inc. All rights reserved.
  55. 55. Next Best Product - ExamplesCase Study: Erste Bank Group 56 Copyright © 2011, SAS Institute Inc. All rights reserved.
  56. 56. The NBO 4 main components Customer Behavior Importance “The probability of the “The profitability customer to aquire the generated if the product” customer aquire the product” Restrictions Previous Contacts “The product can be “The product was sold to the customer” already offer to this customer” 57 Copyright © 2011, SAS Institute Inc. All rights reserved.
  57. 57. Personalized “Next Best Product” offer executed across…Branch/Advisor nbp 58 Copyright © 2011, SAS Institute Inc. All rights reserved.
  58. 58. How to get started ?Value Phase 3 Cross- and up-sale Phase 2 Phase 1 Churn & Credit Risk Profitability 59 Time Copyright © 2011, SAS Institute Inc. All rights reserved.
  59. 59. LønsomhedsopgørelseOmsætning:Forbrugs DB 100Abonnement DB 100= Samlet DB1 200Direkte kapacitetsomkostninger:Salg 10Marketing 20Kampagne 30= Samlet direkte kapacitetsomkostninger 60= Samlet DB2 140Indirekte kapacitetsomkostninger:Kundecenter 15Billing / Produktskifte 25Debitorer 35= Samlet direkte kapacitetsomkostninger 75= Samlet omkostninger 135= Samlet DB2,5 65Øvrige omkostninger 50= EBIT 15 60 Copyright © 2011, SAS Institute Inc. All rights reserved.
  60. 60. Lønsomhed på kunde 61 Copyright © 2011, SAS Institute Inc. All rights reserved.
  61. 61. 62Copyright © 2011, SAS Institute Inc. All rights reserved.
  62. 62. 4 slides on analytics – that’s it ! Copyright © 2011 SAS Institute Inc. All rights reserved.
  63. 63. Why Predictive Modeling? 100 90 80 Cuml Gains( Caputre) 72% 70 62% 60 48% 50 40 30% 30 20 Targeting the top 10% of customer base capture 30% churners 10 0 0 1 2 3 4 5 6 7 8 9 10 DecileBenefits of Modeling vs. Random Targeting •Increased response rate by contacting the right customers •Reduced campaign cost by selecting the most-likely to act customers •Conveying the right message by understanding target population 64 Copyright © 2011, SAS Institute Inc. All rights reserved. 16
  64. 64. Predictive Modeling TechniquesDecision TreeAttempts to split a population into subgroups that tend to be morehomogeneous than the original sample. Each of the subgroups continue tobe split into even smaller subgroups until the model cannot be improved.Pros: Allows for non-linear relationships, very intuitive Cons: Clumping ofprobabilities and less distribution 8.00%Clustering 7.00% Cluster 3Identify groups of individuals based on their proximity to each other. 6.00% 3 22.5% of the upgraders 5.90% churn Proportion ChurnThe cluster procedure and discriminate* analysis utilizes an effective 5.00% $36.67 avg. ARPU Cluster 2 4.19% of the upgraders 3.17% churnmethod for finding initial clusters with a standard iterative algorithm for 4.00% $173.40 avg. ARPU 3.00% 2minimizing the sum of squared distances from the cluster means . Cluster 1 73.4% of the upgraders 2.00% 1 1.80% churn $87.14 avg. ARPU 1.00%Logistic Regression 0.00% $0 $25 $50 $75 $100 $125 $150 $175 $200A generalized linear model for predicting probabilities. Logistic Regression ARPUcalculates the probability of a particular record being a member of a targetgroup, based on the values of the predictor fields.Yi = B0 + B1Xi1 + B2Xi2 + … + BkXik + EPredicted Churn = B0 + B1(Cell Minutes) + B2(Customer value) + E 65 Copyright © 2011, SAS Institute Inc. All rights reserved.
  65. 65. Predictive Modeling TechniquesNeural Networks Data can be processed in parallel and complex relationships can befound quickly. Nodes in Neural Networks sums information from othernodes connected to it and passes information to the other nodes.Pros: Allows for more complex, non-linear, relationships Cons:Interpretation very difficult - Called a “black box”Survival ModelMethod of statistical analysis used for determining time-to-event for one-time Survival Curves by Credit Class 2004 100%Events. Includes both the actual probability of event and effects of A_survcovariates. Enables to: 90% B_surv C_H_surv •Study survival trends by demographic area, channel, credit 80% D_surv class, rate plan, type of churn etc 70% E_surv Remaining (%) •Estimate remaining lifetimes for present customers N_surv 60% Other_surv 50% total_surv 40% 30% 20% 10% 0 365 730 1095 1460 1825 2190 2555 2920 3285 3650 Tenure (days) 66 Copyright © 2011, SAS Institute Inc. All rights reserved.
  66. 66. Develop Treatment Strategy -- Example Three Tier Approach: 1) Predictive Modeling 2) Segmentation 3)Value 100 8.00% Model can capture 62% of 90 Churners by targeting 30% of 7.00% Cluster 3 the entire base 80 22.5% of the upgraders Cuml Gains( Caputre) 72% 6.00% 5.90% churn 3 Proportion Churn 70 $36.67 avg. ARPU Cluster 2 62% 5.00% 4.19% of the upgraders 60 3.17% churn $173.40 avg. ARPU 48% 4.00% 50 3.00% 2 40 Cluster 1 30% 73.4% of the upgraders 30 2.00% 1 1.80% churn $87.14 avg. ARPU 20 1.00% 10 0.00% 0 $0 $25 $50 $75 $100 $125 $150 $175 $200 ARPU 0 1 2 3 4 5 6 7 8 9 10 Decile1. Model –Churn model to select at-risk customers2. Segmentation – Multivariate segmentation to understand theprofile and usage patterns of specific target populations3 . Value – Derived from revenues, costs, and expected customerlifetime based on survival analysis to optimize the right offer to theright customer 67 23 Copyright © 2011, SAS Institute Inc. All rights reserved.
  67. 67. Applying Predictive Models to Marketing StrategyMarketing Objective Question Modeling Approach Treatment Strategy Why Will Customer Churn? Propensity to Churn Reconciliation When Will Customer Survival Model (Time until Churn & churn) Churn? Value Segment Who is Savable? Propensity to Stay Decile H M L 1 Propensity Score 2 3 Who Will Buy? What? 4 Propensity to Buy 5 Maximize Revenue Which product will Product Basket Customer Buy Next? When Will Customer Survival Model (Time until Buy? Purchase) 68 Copyright © 2011, SAS Institute Inc. All rights reserved.
  68. 68. Customer insightin action at Tesco Copyright © 2011 SAS Institute Inc. All rights reserved.
  69. 69. Tesco Overview Formed in 1924 The UK’s largest food retailer Operating stores in all formats – convenience, high street, super markets and hyper markets. Operating in 13 countries around the world The world’s leading internet grocery retailer Substantial Finance and telecoms businesses 70 Copyright © 2011, SAS Institute Inc. All rights reserved.
  70. 70. Tesco and SAS Currently use SAS across the business to help Select locations Plan investment in refurbishments Margin and revenue reporting Analysis of operational performance in Tesco.com Through SAS with Dunnhumby 71 Copyright © 2011, SAS Institute Inc. All rights reserved.
  71. 71. Tesco- a truly customer focussed business “Our mission is to earn and grow the lifetime loyalty of our customers”“Contiually increasing Tesco has a core aim “tovalue for customers understand customersto earn their lifetime better than anyone”loyalty.Tesco PLC, Annual review and Sir Terry Leahy, Chief Executivesummary financial statement 72 Copyright © 2011, SAS Institute Inc. All rights reserved.
  72. 72. Translate data in to a clear picture of acustomer Data Miss virtanen 73 Copyright © 2011, SAS Institute Inc. All rights reserved.
  73. 73. Shopping behaviour can explain a lot… Data Miss virtanen is a busy young lady looks after her health and loves fresh produce drives to the supermarket on a Saturday morning reads lifestyle magazines has a cat doesn’t pay attention to the price of products does look out for promotionsTesco know 12 million customers as well as wenow know Miss Virtanen 74 Copyright © 2011, SAS Institute Inc. All rights reserved.
  74. 74. Step 1: develop a meaningful customersegmentation Research Data • 2800 survey respondents Segmentation Requirements • Shopping behaviour • Simple and intuitive • Loyalty programme participation • Categorises appropriate share of the customer database • Satisfaction • Segments of significant sizes • Sufficient differentiation • Actionable KnowledgeTransactional Data• Points accrual transactions Participation Engagement• Points redemption transactions• Shopping behaviour across 17 retail and service brands Value Satisfaction• Card usage vs automatic points collection Opportunity Segmentation• Response to promotions 75 Copyright © 2011, SAS Institute Inc. All rights reserved.
  75. 75. The Nectar Marketing CommunicationsSegmentation Engaged Enthusiasts Bonus Seekers Savvy Supermarket Shoppers Contented X- Swipeless Savers Routine Grocery Nectar Indifferents Shoppers Shoppers 76 Copyright © 2011, SAS Institute Inc. All rights reserved.
  76. 76. Step 3: Overlaying financial data allowsfor improving the allocation of customermarketing investment Average Customer Profitability and Ability to Promote by SegmentHigh Low profitability, some opportunity to improvePromotability Index via incentives Low profitability, little Highly profitable opportunity to improve segmentsLow via incentives Low Profitability Index High 77 Copyright © 2011, SAS Institute Inc. All rights reserved.
  77. 77. Cross - upsale Copyright © 2011 SAS Institute Inc. All rights reserved.
  78. 78. How to get started ?Value Phase 3 Cross- and up-sale Phase 2 Phase 1 Churn & Credit Risk Profitability 79 Time Copyright © 2011, SAS Institute Inc. All rights reserved.
  79. 79. Kunde scoring - Produkt Kunde Prod A Prod B Prod C 1 90 20 90 2 80 8 4 3 60 9 65 4 55 3 21 5 75 16 50 6 75 65 60 7 75 15 5 8 65 14 33 9 80 47 36 80 Copyright © 2011, SAS Institute Inc. All rights reserved.
  80. 80. Kunde scoring - respons Kunde Kamp A Kamp B Kamp C 1 90 20 90 2 80 70 75 3 60 75 65 4 55 80 75 5 75 60 50 6 75 65 60 7 75 90 65 8 65 60 60 9 80 30 75 81 Copyright © 2011, SAS Institute Inc. All rights reserved.
  81. 81. Net lift NET LIFT “WOULD-BUY-ANYWAY” CLIENTS “SWING” CLIENTS  Will buy anyway  Haven’t made up their  A communication may mind disturb their buying  Can be positively process influenced by communication PREDEICTIVE MODELLING “NO-IMPACT” CLIENTS “DON’T-POKE” CLIENTS  Won’t accept offer  Not likely to accept offer  No impact of  But likely to end relation communication if communicated to 82 Copyright © 2011, SAS Institute Inc. All rights reserved.
  82. 82. Aviva online experience Copyright © 2011 SAS Institute Inc. All rights reserved.
  83. 83. How to get started ?Value Phase 3 Cross- and up-sale Phase 2 Phase 1 Churn & Credit Risk Profitability 84 Time Copyright © 2011, SAS Institute Inc. All rights reserved.
  84. 84. Churn - Score Kunde Churn Q1 Churn Q2 Churn Q3 1 90 20 90 2 80 70 75 3 60 75 65 4 55 80 75 5 75 60 50 6 75 65 60 7 75 90 65 8 65 60 60 9 80 30 75 85 Copyright © 2011, SAS Institute Inc. All rights reserved.
  85. 85. Churn - messageRDM Generated offer ormessage 86 Copyright © 2011, SAS Institute Inc. All rights reserved.
  86. 86. Opsalg ved bookingRetrieve customer andflight informationCheck miles balance tosee if free flightavailableIf free flight notavailable, make valuebased loyalty offer 87 Copyright © 2011, SAS Institute Inc. All rights reserved.
  87. 87. Analytics embedded in customer process “Conversational Data” Prioritized offers and about customers’ consistent treatment financial objectives & for each customer existing relationshipsInternet Banking Contact Center Personal Tellers U.S. Bank ATM(New!) Bankers (New!) Behavioral Insights Predictive Analytics Relationship Strategies Mining up to 15 million Evaluating customer value, Converting insights into transactions each day to creditworthiness, purchase decisions and guidance that identify out-of-pattern propensity and future is passed to legacy systems behaviors that may signal potential for over 13 million and customer facing need needs consumers employees 88 Copyright © 2011, SAS Institute Inc. All rights reserved.
  88. 88. Embedded analytics Copyright © 2011 SAS Institute Inc. All rights reserved.
  89. 89. A Vast Universe of Data What are they requesting? How many did they buy? What are my competitors doing? Who is buying what?What else did they consider? Where did they buy it? When did they buy? What prices where they quoted? How much did they pay? Did they buy multiple products? 90 Copyright © 2011, SAS Institute Inc. All rights reserved. 90
  90. 90. Capture and Analyze a SmallPortion What are they requesting? How many did they buy? What are my competitors doing? Who is buying what? What else did they consider? Where did they buy it? When did they buy? What prices where they quoted? How much did they pay? Did they buy multiple products? 91 Copyright © 2011, SAS Institute Inc. All rights reserved. 91
  91. 91. We Miss the Elegant Patterns 92 Copyright © 2011, SAS Institute Inc. All rights reserved. 92
  92. 92. Customer experience across datasilos Research/Metric– Non-guest centric Visitation data that helps understanding of Guest Dimensions of data Offer History - Survey Attendance Mindset Patterns RM Commun Resort ication GSM History Segmen Marketing tation mix ScoresOperations Data – created from WDPRmarketing efforts Geo- Model Demogra Scores The phic Guest Resort Social Individual – What we know Shop Media about the guest from their past behavior Internet Prior Inbound Guest activity – Response to registrat Stay marketing efforts ion VPK Pass Request holder Behavioral Data Resort and MDV Theme Operations Data Park Spending Inbound Guest Data Research/Metric Data 93 Copyright © 2011, SAS Institute Inc. All rights reserved.
  93. 93. AD-HOC ANALYTICAL SEGMENTATION AD-HOC CHURN PREDICTION ANALYTICS AD-HOC PRODUCT ASSOCIATION ANALYSIS AUTOMATED CHURN PREDICTION CUSTOMER AND PRODUCT PROFITABLITY ANALYSIS AUTOMATED CROSS-/UP SELL CHURN PREDICTION Generic roadmap COMMUNICATION ANALYTICS (TIMING AND CHANNEL) PREDICTIVE ANALYTICS CLV ANALYTICS BEHAVIORAL PROFILING CAMPAIGN LIFT ANALYTICS PRICE AND PROMOTION ANALYTICS Customer Analytics DESIGN OF EXPERIMENTS ANALYZING RELATIONS BETWEEN CUSTOMERSCopyright © 2011, SAS Institute Inc. All rights reserved. CROSS-CHANNEL DIRECT MARKETING OPTIMZATION ANALYZING CUSTOMER DIALOG CUSTOMER LINK ANALYTICS ANALYZING CHANNEL SPILLOVER EFFECTS TEXT ANALYTICS ANALYZING DIMINISHING RETURN ON INVESTMENTS DIRECT MARKETING OPTIMIZATION MARKETING MIX OPTIMIZATION 94
  94. 94. Pradigmeshift inMarketing processes Copyright © 2011 SAS Institute Inc. All rights reserved.
  95. 95. Campaign Management Process Campaign Management Exclusion Criteria Campaign Campaign Campaign Management Management Management Creating Campaign Campaign Target Planning Execution Groups Analysis Response Modelling etc 96 Copyright © 2011, SAS Institute Inc. All rights reserved.
  96. 96. Is S3 the Optimal Solution? Setting the Max Offers constraint at 512K delivers profit of Kr 13.5M Sensitivity analysis tells us that we could find more expected value, but the value Maybe the Max Offers value is too is limited – Increasing the high – do we really want to make number of offers to 590K most of the offers with only a would only deliver additional marginal expected value? expected value of Kr 0.09M – surely not worth the additional expenditure 97 Copyright © 2011, SAS Institute Inc. All rights reserved.
  97. 97. Results - 3. Switch Focus to ProfitabilityEffect of using SAS MO Scenarios10.000.000 S1 - Solution • 512k offers 8.000.000 • Expected Profit: Kr -4.93M 6.000.000 S7 – MO Maximize Profit • Max offers 512k constraint 4.000.000 • Maximize profit objective • Only makes 178k offers 2.000.000 • Expected Profit: Kr 7.986M 0 Offers Expected Profit-2.000.000-4.000.000-6.000.000 S1 Solution S7 Max Profit 98 Copyright © 2011, SAS Institute Inc. All rights reserved.
  98. 98. Campaign Management Process Campaign Management Exclusion Criteria Campaign Campaign Campaign Management Management Management Creating Campaign Campaign Target Planning Execution Groups Analysis Response Modelling etc 99 Copyright © 2011, SAS Institute Inc. All rights reserved.
  99. 99. offer optimization - a complex problem choices and constraints: • many customers, offers, channels • necessary strategic actions (must push offer A) • customer contact policies • many operational constraints: • budgets, resources, capacities challenges: • which offers maximise profit / ROI, but stay within constraint boundaries? • how do you find the optimal investment strategy? • how do you reconcile competing goals (product vs. customer)? • how do you plan effectively for change? 100 Copyright © 2011, SAS Institute Inc. All rights reserved.
  100. 100. and actually what happens is….• subjective planning and decision making• poor ROI from marketing campaigns • low response rates • ineffective use of channels • unnecessarily high costs• ineffective contact policy • over-contacting customers • lack of enforcement• no ability to understand tradeoffs between key elements e.g. volume vs. profit 101 Copyright © 2011, SAS Institute Inc. All rights reserved.
  101. 101. Why optimization outperforms other decisioning techniques 102 Copyright © 2011, SAS Institute Inc. All rights reserved.
  102. 102. Simple Example – Prioritizing Campaigns• Model scores define response probability for each campaign• Probability * Expected Revenue = Expected Value• Expected Value drives campaign allocation• Constraints: 1 customer 1 campaign & 1 campaign 3 customersClient Camp A Camp B Camp C 1 100 120 90 2 80 70 75 Campaign C 3 60 75 65 4 55 80 75 5 75 60 50 6 75 65 60 Campaign Campaign 7 75 90 65 B A 8 65 60 60 9 80 140 75 103 Copyright © 2011, SAS Institute Inc. All rights reserved.
  103. 103. Level 1: Campaign Prioritization • Constraints: 1 customer: 1 campaign / campaign - 3 customers • Campaigns assigned priority • Top customers selected for each campaign based on their expected value Client Camp A Camp B Camp C 1 100 120 90 ? 2 80 70 75 3 60 75 65 4 55 80 75 5 75 60 50 ? 6 75 65 60 ? 7 75 90 65 8 65 60 60 ? 9 80 140 75 ? Expected Return: 675 104 Copyright © 2011, SAS Institute Inc. All rights reserved.
  104. 104. Level 2: Customer Offer Prioritization • Constraints: 1 customer: 1 campaign / campaign - 3 customers • Priority assigned based on the customer • Top campaign selected for each customer based on their expected value Client Camp A Camp B Camp C 1 100 120 90 2 80 70 75 3 60 75 65 4 55 80 75 5 75 60 50 6 75 65 60 7 75 90 65 ? 8 65 60 60 ? 9 80 140 75 ? Expected Return: 705 (+30) 105 Copyright © 2011, SAS Institute Inc. All rights reserved.
  105. 105. Level 3: Optimization Approach • Constraints: 1 customer: 1 campaign / campaign - 3 customers • Optimisation evaluates ALL possible solutions to find the best • While also respecting constraints Client Camp A Camp B Camp C 1 100 120 90 2 80 70 75 3 60 75 65 4 55 80 75 5 75 60 50 6 75 65 60 7 75 90 65 8 65 60 60 9 80 140 75 Expected Return: 780 (+75) 106 Copyright © 2011, SAS Institute Inc. All rights reserved.
  106. 106. Campaign Management Process Campaign Management Exclusion Criteria Campaign Campaign Campaign Management Management Management Creating Campaign Campaign Target Planning Execution Groups Analysis Response Modelling etc 107 Copyright © 2011, SAS Institute Inc. All rights reserved.
  107. 107. Cultural Impact Post OptimizationMinimal impact on Campaign Management Process Campaign Analysis Management Exclusion Response Criteria Modelling etc Campaign Campaign Campaign Marketing Campaign Management Management Management Optimization Management Campaign Creating Target Groups Campaign Campaign Planning Eligible Campaign Optimization Execution Groups Execution Scenario 1 Scenario 2 Scenario 2 Scenario 3 • Ideally marketers now create eligible groups, rather than target groups • Propensity models drive campaigns.. • ….along with what the business is trying to achieve (goals and constraints) • An extra step, yes, but it can smooth the process – campaigns become simpler 108 Copyright © 2011, SAS Institute Inc. All rights reserved.
  108. 108. Tips to get started Get started Know your customers motivation to engage with you and do a communication segmentation Start to collect data on: Transactions, response behaviour, social behaviour, demographics Do simple datamining Explore your datamining segments in order to apply communication segment to each customer/prospect Execute, Execute 109 Copyright © 2011, SAS Institute Inc. All rights reserved.
  109. 109. AD-HOC ANALYTICAL SEGMENTATION AD-HOC CHURN PREDICTION ANALYTICS AD-HOC PRODUCT ASSOCIATION ANALYSIS AUTOMATED CHURN PREDICTION CUSTOMER AND PRODUCT PROFITABLITY ANALYSIS AUTOMATED CROSS-/UP SELL CHURN PREDICTION Generic roadmap COMMUNICATION ANALYTICS (TIMING AND CHANNEL) PREDICTIVE ANALYTICS CLV ANALYTICS BEHAVIORAL PROFILING CAMPAIGN LIFT ANALYTICS PRICE AND PROMOTION ANALYTICS Customer Analytics DESIGN OF EXPERIMENTS ANALYZING RELATIONS BETWEEN CUSTOMERSCopyright © 2011, SAS Institute Inc. All rights reserved. CROSS-CHANNEL DIRECT MARKETING OPTIMZATION ANALYZING CUSTOMER DIALOG CUSTOMER LINK ANALYTICS ANALYZING CHANNEL SPILLOVER EFFECTS TEXT ANALYTICS ANALYZING DIMINISHING RETURN ON INVESTMENTS DIRECT MARKETING OPTIMIZATION MARKETING MIX OPTIMIZATION 110

×