1. Segmentation plays a key role in loyalty marketing by dividing customers into groups based on common attributes and behaviors. This allows companies to better understand their customers and maximize relationships.
2. There are various levels of segmentation from basic demographics and purchases to more advanced psychographics and transaction data. Companies can use both supervised and unsupervised segmentation.
3. Effective segmentation identifies strategic business focuses, provides insights into customer needs, and helps companies focus communications and campaigns. It is a process that aims to create meaningful customer groups.
8. Inbound Right-Time MarketingList PullSCVSegmentAnalyticsEvent DetectionCampaign MgmtReal TimeMaturity of Direct Marketing Marketing Effectiveness: ROI Courtesy of SAS
11. Levels of Segmentation Information Required Courtesy of SASProducts OwnedNo Segmentation
12. Levels of Segmentation Information Required Courtesy of SASChannel UtilizationProducts OwnedNo Segmentation
13. Levels of Segmentation Information Required Courtesy of SASDemographicsChannel UtilizationProducts OwnedNo Segmentation
14. Levels of Segmentation Information Required Courtesy of SASTransaction InformationDemographicsChannel UtilizationProducts OwnedNo Segmentation
15. Levels of Segmentation Information Required Courtesy of SASPsycho- graphicsTransaction InformationDemographicsChannel UtilizationProducts OwnedNo Segmentation
16. Levels of Segmentation Information Required Courtesy of SAS1Psycho- graphicsTransaction InformationDemographicsChannel UtilizationProducts OwnedNo SegmentationSegment of One
22. Customer Segmentation•Identifies strategic business focus and direction•Analysis of customer behavior to gain insight into customer needs and preferencesKey Benefits & Capabilities
23. What makes a segment? Measurable identifying elements that distinguish from othersSegments desirably have these characteristics:
24. What makes a segment? Defined contact points or channels through which communication is possibleSegments desirably have these characteristics:
25. What makes a segment? Quantifiable size so that cost computations may be done for targeting themSegments desirably have these characteristics:
26. What makes a segment? Have generally unique stated or implied needsregarding the product or serviceSegments desirably have these characteristics:
27. What makes a segment? Stability and robustness to random shocks(applies to some applications) Segments desirably have these characteristics:
28. What is Segmentation? “a process of creating groups of customers whohave SIMILAR behavior and characteristics”
29. Segmentation TypesUnsupervised data-driven segmentation; segments determined after data gathering and processing using statistical analysesSupervised segmentation based on pre-defined factors
30. Supervised Segmentation
•Usually uses less variables with pre-defined “cuts”.
•Ad-hoc, user-driven
•Other variables are used as mere profilers and not active segmenters
•Applicable when user has a distinct focus and variables of interest are readily available.
30
31. Some Prototype Segmentations
Customer Value versus Tenure
Customer Value versus Transaction Type & Frequency
Customer Value versus Risk
Profit Margin or Profit Rate against Tenure, Transaction Frequency or Risk
Purchase Behavior
Other possible information:
31
Variety of Products Availed
Life Stage
Family Life Cycle
The Remittance Market
32. Segmentation Variables
•Measures the amount of business brought in by the customer
•Also measures the capacity of a customer for cross-sell/upsell
•There is difficulty in measuring “high”, “medium” and “low” value.
•There are varying indicators of value
•ADB (CA/SA) , Investments
•Loan amount/ Outstanding Balance
•Total purchase per transactionCustomer Value
33. Segmentation Variables
•Measures the loyalty of customer with respect to time
•Usually a “net time value”, i.e. lulls between product availment are not counted
•Skewness in data is an issueTenure
34. Segmentation Variables
•Identifies the “sleepers” from “transactors”
•Number of Transactions per Month is a usual metric.
•Time-between-transactionsis a good substitute segmentation variableTransaction Frequency
35. Segmentation Variables
•Tag customers given certain warning signals
•common indicators are:
•Low ADB
•Defaults
•Lapses and claimsRisk Indicators
36. Segmentation Variables
•Metric for each customer’s contribution to total profit
•Used to level the number of products with the value of products availedProfitability
37. Segmentation Variables
Common in Market Research but also evident in transactional information
•Utility/benefit from product
•Usage rate
•Loyalty vis-à-vis switching, hopping, ambivalence
•Propensity/Proclivity to buy/avail/take-up
•Temporal stimuli (payday, holidays, special events) Behavior
38. Segmentation Variables
Some segmentation variables are also profiling variables
•Age, number of dependents, marital status
•Ownerships (home, car, business, etc.)
•Employment (nature of business, position, job tenure)
•Geographic information
•Delinquencies/ Fraud history, if any
•ChannelsProfiling Variables
40. Company A
•Launched a loyalty card
•Has big data on transactions
•Known as an innovator
•Challenge is to avert the impact of patent expiry and generic erosion
41. Company B
•Has different/diverse businesses in different industries
•Has product ownership, transactional data
•Challenge is to maximize customer relationship through cross-sell and upsell
42. Step 1: List Pull
•Involves definition of target population
•By featured product/s
•By time period of observation and analysis
•By geographic coverage
•Brainstorm on Key Metrics and required raw data
•Demographics
•Transactional behavior
•Profitability Drivers
List of Customers
43. List of CustomersStep 2: Single Customer View
•Consolidation of customer level information throughout the entire collection of data to be used for analytics
•Through the SCV, the analyst can tract a specific customer’s profile, behavior & profit contribution.
•The SCV is the recipient of scores
derived from analytics exercises.
44. Step 2: Single Customer ViewSCV lends itself to queriesStatistical MatchingRemoved inactive accountsRemoved cancelled accountsCorporateRetail
45. Step 3: Segmentation
•Identify and understand best and worst performing customers
•Input for programs that focus on the following:
•Increasing profitability
•Motivating positive behavioral changes:
•Activate sleepers
•Increase usage of active customers
•Leads to best targets for cross-selling and up-selling
•Protect our most valued customers
•It’s more expensive to acquire a new customer than retain a good one.
46. Sample SegmentationSource: Data Mining Techniques in CRM: Inside Customer Segmentation, K Tsiptsis & A Korianopoulos
Use of card for entertainment (bars, resto)
Use of card for gym, fitness centers.
Highest internet usage
47. Sample SegmentationSource: Data Mining Techniques in CRM: Inside Customer Segmentation, K Tsiptsis & A Korianopoulos
Increased purchases at apparel stores and accessory stores
High balances
48. Sample SegmentationSource: Data Mining Techniques in CRM: Inside Customer Segmentation, K Tsiptsis & A Korianopoulos
Use of card for travel & airfare
Highest international usage
Highest internet usage
49. Sample SegmentationSource: Data Mining Techniques in CRM: Inside Customer Segmentation, K Tsiptsis & A Korianopoulos
Daily needs
Use of the card mainly for supermarkets and gas.
50. Sample SegmentationSource: Data Mining Techniques in CRM: Inside Customer Segmentation, K Tsiptsis & A Korianopoulos
Lowest purchase frequency
Infrequent but high value transactions
Main spend is electronic / appliance
51. Sample SegmentationSource: Data Mining Techniques in CRM: Inside Customer Segmentation, K Tsiptsis & A Korianopoulos
>=50% spend on Installment
Low retail spend
Revolver
52. Sample SegmentationSource: Data Mining Techniques in CRM: Inside Customer Segmentation, K Tsiptsis & A Korianopoulos
Use of card for heath purposes and DIY shops
Lowest internet usage
Infrequent but high value purchases
53. Sample SegmentationSource: Data Mining Techniques in CRM: Inside Customer Segmentation, K Tsiptsis & A Korianopoulos
Diverse Card Usage.
Purchase at different merchants
Moderate balance amount
High purchase frequency
54. Sample SegmentationOne segmentation led to another segmentation that targets loyalty. Patient SegmentationDoctor Segments (Example) High Growth Potential
Highest %Highly-compliant low dosage users
Also some highly-compliant high dosage users.
Lowest %Low-value patients
ProfileNot recruiting actively.
Most are interns.
55. Step 4: Analytics
•Wide array of statistical analysis aimed at understanding the customer base and the derived segments.
•Typical techniques are product association (market basket analysis), portfolio analysis (reports).
57. Step 5: Event Detection
•Attempt to answer the question, “Who among my customers are likely to leave me?”
•This is usually addressed by Churn Modeling.
Example:
Actual
Churned
Stayed
Total
Model
Says
“Churn”
3,151
1,335
4,486
“Stay”
529
2,985
3,514
Total
3,680
4,320
8,000
Using logistic regression analysis, themodelwas able to capture 87% of the true state of nature (true churners and true stayers). Further drill-down is done within the four outcome states.
61. There are solutions which optimize Customer Management Process that reflects the voice of the customer, promotes retention and relationship building, supports business goals, leverages events / triggers, and is cross- channel and cross Business Unit.
62. Step 7: Inbound Right-Time Marketing
•“Right message at the right place and at the right time”
•Objective is to make heralds out of the customers
63. Step 8 : Optimization
•Cutting edge innovation
•Tailor-fit customer relationship
•Affinity and pride is established
•Must beware of oversolicitation.
64. Please Remember
•The goal of the segmentation analysis is to create manageable and meaningful customer groups among customers.
65. Please Remember
•Segmentation is instrumental in increasing shareholder value by identifying:
•Most high-value segment(s)
•Segments with high potential for cross selling and/or up-selling
•By focusing communications on a targeted segment, a causal effect would be a reduction in campaign costs
66. Please Remember
•Segment definition
•Supports retention, service prioritization and cross selling / up-selling efforts
•Serves as input in developing new products
•Segmentation is both a science and an art!
67. Thank you for your attention! Prof. Francisco N de los ReyesSchool of StatisticsUniversity of the Philippines, Diliman