![[Big]Data for Marketers:
Targetingthe Right Customers
Preparedby:
PanjiWinata
Jakarta,8th November2017](https://image.slidesharecdn.com/20171108bigdataformarketerstargetingtherightmarket-171109021834/85/Big-Data-For-Marketers-Targeting-the-Right-Market-1-320.jpg)
![Why We Need [Big] Data Analytics
7
Some of subjects who succeeded in getting more benefits by implementing [big]
data analytics on their business strategies / actions :](https://image.slidesharecdn.com/20171108bigdataformarketerstargetingtherightmarket-171109021834/85/Big-Data-For-Marketers-Targeting-the-Right-Market-7-320.jpg)
![Why We Need [Big] Data Analytics
8
Source:
MapR (June 2013); *Teradata has since launched a new Extreme Data Appliance (1700) as $2,000/TB in October
2013
From the storage cost perspective, Hadoop has much lower cost compared
to RDBMS data warehouse.
Beside that, Hadoop provides highly scalable capacity compared to any
RDBMS.](https://image.slidesharecdn.com/20171108bigdataformarketerstargetingtherightmarket-171109021834/85/Big-Data-For-Marketers-Targeting-the-Right-Market-8-320.jpg)
![CustomerSegmentationUsingRFM
14
“Customer Segmentation is the subdivision of a market into discrete
customer groups that share similar characteristics.”[3]
RFM is a method used for analyzing customer value.[4]
RFM stands for the three dimensions:
Recency – How recently did the customer purchase? (days since last purchase)
Frequency – How often do they purchase?
Monetary Value – How much do they spend?
The combination for each score of those dimensions will create different groups,
and then the groups can be clustered into several segments.
Pareto principle is the core of RFM.
Hence, the treatment for each customer segment should be different for gaining
the best benefits.](https://image.slidesharecdn.com/20171108bigdataformarketerstargetingtherightmarket-171109021834/85/Big-Data-For-Marketers-Targeting-the-Right-Market-14-320.jpg)
![38
[1] https://www.forbes.com/sites/gartnergroup/2013/03/27/gartners-big-data-
definition-consists-of-three-parts-not-to-be-confused-with-three-
vs/#6595be1942f6. Accessed on 7th November 2017.
[2] Jacobs, A. 2009. The pathologies of big data. Communications of the ACM 52(8),
36-44.
[3] http://www.bain.com/publications/articles/management-tools-customer-
segmentation.aspx. Accessed on 7th November 2017.
[4] Fader, P. S., Hardie, B. G., & Lee, K. L. (2005). RFM and CLV: Using iso-value
curves for customer base analysis. Journal of Marketing Research, 42(4), 415-430.
Reference](https://image.slidesharecdn.com/20171108bigdataformarketerstargetingtherightmarket-171109021834/85/Big-Data-For-Marketers-Targeting-the-Right-Market-38-320.jpg)
[Big] Data For Marketers: Targeting the Right Market
- 1. [Big]Data for Marketers: Targetingthe Right Customers Preparedby: PanjiWinata Jakarta,8th November2017
- 2. 2 About Me Lead data science in OVO (Lippo Group Digital) Big Data team which consist of 5 data scientists in developing analytics & generating actionable business insights from multiple industries data in: Financial Technology, E-Commerce, Retail, Hospital, etc. Master Degree of Information Technology, Universitas Indonesia. Final project: Architectural Design of Big Data with TOGAF Framework: Case Study at a Telecommunication Company https://www.linkedin.com/in/panjiwinata/ panji.winata@nusantaraanalytics.com Current Job
- 3. 3 What Is In This Slide o What is big data and what is the most technology used in big data What Is Big Data Why We Need Big Data Analytics o The reason why we need big data analytics Business Use Case 1: o 360 Degree of Customer Profile Business Use Case 2: o Customer Segmentation Using RFM Business Use Case 3: o Smartphone Adoption Prediction for Telecommunication Service Business Use Case 4: o Event Trigger for Preventing Customer Churn in Telecommunication Service o Big data and analytics illustration in short video from Harvard Business Review o The progression of analytics Big Data and Analytics Question & Answer
- 4. What Is Big Data 4 "Big data" is high-volume, - velocity and -variety information assets that demand cost-effective, innovative forms of information processing for enhanced insight and decision making. (Gartner, 2013) The Most used Big Data Technology basis: Big data is too large to be placed in a relational database and analyzed with the help of a desktop statistics/visualization package—data, perhaps, whose analysis requires massively parallel software running on tens, hundreds, or even thousands of servers. (Jacobs, A, 2009)
- 5. Big Data and Analytics 5
- 6. 6
- 7. Why We Need [Big] Data Analytics 7 Some of subjects who succeeded in getting more benefits by implementing [big] data analytics on their business strategies / actions :
- 8. Why We Need [Big] Data Analytics 8 Source: MapR (June 2013); *Teradata has since launched a new Extreme Data Appliance (1700) as $2,000/TB in October 2013 From the storage cost perspective, Hadoop has much lower cost compared to RDBMS data warehouse. Beside that, Hadoop provides highly scalable capacity compared to any RDBMS.
- 9. 9 BUSINESS USE CASE 1: 360 Degree of Customer Profile
- 11. 11 360 Degreeof CustomerProfile bycombiningmultipledatasources(customertransactiondata,customermembershipdata, andothersdata)to transformthemintomeaningfulinformationformorepreciseactionableinsights Age Gender Fav. Merchant Name Fav. Merchant on Working Hour Segment Spending Segment Spending Lifestage Generation Relationship Status Fav. Merchant on Home Hour Location Segment Lifestyle Payment Preference Religion Product Tendency Interest Demographic Geographic Behaviours Interest Loyalty Transactions Payment External Data
- 12. 12 360 Degreeof CustomerProfileSample amultidimensionscustomerattributesofcustomers Demographic: • Mobile Phone: 081321494266 • Age 31 • Male • Older Millenial • Married • Islam Geographic: • Bogor Resident • Fav. Merchant: Maxx Coffee • Fav. Merchant on Working Hour: Maxx Coffee • Fav. Merchant on Home Hour: Cinemax Behaviour: • Length of Stay: 150 days • Product Tendency: Beverages • Segment Spending: Top Spender • Segment Spending Lifestage: Stable • Segment Lifestyle: Promo Seekers • Payment Preference: Credit Card Interests: • Online Payment • Cinema • Gadget • Culinary • E-Commerce Transactions: • 20 Times / Month • Revenue IDR 2.5 Mio • Fav. Product Name: Cappucino Loyalty: • Loyalty Points: 75,000 • Loyalty Segment: Medium • Ever Used Loyalty Points: No
- 13. 13 BUSINESS USE CASE 2: Customer Segmentation Using RFM
- 14. CustomerSegmentationUsingRFM 14 “Customer Segmentation is the subdivision of a market into discrete customer groups that share similar characteristics.”[3] RFM is a method used for analyzing customer value.[4] RFM stands for the three dimensions: Recency – How recently did the customer purchase? (days since last purchase) Frequency – How often do they purchase? Monetary Value – How much do they spend? The combination for each score of those dimensions will create different groups, and then the groups can be clustered into several segments. Pareto principle is the core of RFM. Hence, the treatment for each customer segment should be different for gaining the best benefits.
- 15. CustomerSegmentationUsingRFM 15 Let’s do customer segmentation using RFM analysis with a simple example below
- 16. Recency:Let’sclassifytherecencyinto3 categories 16 We will classify the 3 categories as below: High Recency (recency score 3): Top 33.33% from ascending order of the recency Medium Recency (recency score 2): not classified as high or low recency Low Recency (recency score 1): Bottom 33.33% from ascending order of the recency
- 17. Frequency:Let’sclassifythefrequencyinto3categories 17 We will classify the 3 categories as below: High Frequency (frequency score 3): Top 33.33% from descending order of the frequency Medium Frequency (frequency score 2): not classified as high or low frequency Low Frequency (frequency score 1): Bottom 33.33% from descending order of the frequency
- 18. Monetary:Let’sclassifythemonetaryinto3categories 18 We will classify the 3 categories as below: High Monetary (monetary score 3): Top 33.33% from descending order of the monetary Medium Monetary (monetary score 2): not classified as high or low monetary Low Monetary (monetary score 1): Bottom 33.33% from descending order of the monetary
- 19. Creating the Customer Segment 19 We create 8 segments by combining the recency, frequency, and monetary Scores as below:
- 20. Creating the Customer Segment 20 Apply those customer segment formula into the data:
- 21. KeyInsightsof TheSimpleSampleCustomer SegmentationUsingRFM 21 Key Insights 22% customers (11% from Champion & 11% from Can’t Loose Them segment) generated 43% total revenue (20% by champion & 23% by Can’t Loose Them) with 63% total transactions (29% by champion & 34% by Can’t Loose Them) Focusing your efforts on critical segments of customers is likely to give you much higher return on investment! Every segments has to be treated differently, as recommended below
- 22. 22 BUSINESS USE CASE 3: Smartphone Adoption Prediction for Telecommunication Service
- 23. 23 Due to the Decreasing PT.XTelco ARPU Trend in Last 5 Years, It is interesting to analyze the Strategy for PT. XTelco Business. Red Ocean Strategy vs Blue Ocean Strategy ? Which one ? Source : http://www.blueoceanstrategyaustralia.com.au/what-is-bos/red-vs-blue/ How about if we choose both of them : - Exploit existing demand (eg. Broadband Service, etc) - Create and capture new demand (eg. IoT M2M Mobile Health Care, etc) First Step is we need to know What is The Critical Enabler in customer Perspective For both of the Strategies.
- 24. 24 Harvard Business Review : “Smartphones will be the occasion for significant “reverse innovation” in the coming months and years. Telecommunication Industry must jump on this opportunity.” (https://hbr.org/2011/05/smart-phone-a-reverse-innovati.html) The Hypothesis : In Customer Perspective, Smartphones will be the critical enabler to exploit existing demand, and also create and capture new demand in PT.Xtelco Market. The Hypothesis in PT. XTelco is proven (Smartphone is the critical enabler) by these facts below:
- 25. Analytical Solution With Smartphone Adoption Predictive Model, we can predict the Propensity of The Smartphone Adoption and create segmented offer/device bundling to non Smartphone user on the PT. Xtelco Market. Based on CRISP DM (Cross Industry Standard Process For Data Mining), The Global Process is: 25
- 26. 26 Target Segments • Predictor Variable Binning : Quantiles • Predictor Variable Selection : Backward Stepwise • Logistic Regression Model • Out-Sample Set Validation • Evaluation Using CCR (Correct Classification Rate, Sensitivity, Specificity) All active Smartphone Users who were non Smartphone Users in the Previous month. The predictor variables : • Customer Interests • Device (Data Capable, Device Network, Device Type) • Demography (LOS, Age, Home Location, Office Location, Weekend Location) • Usage Details (ARPU, Broadband ARPU, Voice ARPU, SMS ARPU, Roaming ARPU, Digital Service ARPU) • Credit Card Log History (Interconnect Data) Technique Used DataUsed Non Smartphone Users Objective Identify and Predict Smartphone Adoption Based on Propensity Model Benefit • Accelerate Smartphone User Adoption on Customers • Better Device Bundling Strategy & Partnership • Increase Postpaid Revenue by Long Term Package Offering (via Bundling Contract)
- 27. 27 - MSISDN - AGE_SEGMENT - GENDER_CATEGORY - LOS_SEGMENT - CUSTOMER_TYPE - CREDIT_CARD_OWNER_FLAG - DEVICE_DATA_CAPABLE_M1 - DEVICE_DATA_CAPABLE_M2 - DEVICE_NETWORK_M1 - DEVICE_NETWORK_M2 - DEVICE_TYPE_M1 - DEVICE_TYPE_M2 - DATA_USER_M1_FLAG - DATA_USER_M2_FLAG - BROADBAND_CHARGE_PKG_M1 - BROADBAND_CHARGE_PKG_M2 - DAY_OF_USAGE_M1 - DAY_OF_USAGE_M2 - REGION_COVERAGE_M1 - REGION_COVERAGE_M2 - ARPU_SEGMENT_M1 - ARPU_SEGMENT_M2 - BROADBAND_ARPU_M1 - BROADBAND_ARPU_M2 - VOICE_ARPU_M1 - VOICE_ARPU_M2 - SMS_ARPU_M1 - SMS_ARPU_M2 - ROAMING_ARPU_M1 - ROAMING_ARPU_M2 - DIGITAL_ARPU_M1 - DIGITAL_ARPU_M2 - SLI_ARPU_M1 - SLI_ARPU_M2 - PAYU_ARPU_M1 - PAYU_ARPU_M2 - FLASH_ARPU_M1 - FLASH_ARPU_M2 - BB_ARPU_M1 - BB_ARPU_M2 - TECHNO_UPDATE_L2M_FLAG - SPORT_L2M_FLAG - NEWS_L2M_FLAG - SHOPPING_L2M_FLAG - SOCMED_L2M_FLAG - CHAT_L2M_FLAG - TRAVELER_L2M_FLAG - GAMES_L2M_FLAG - BUSINESS_L2M_FLAG - MUSIC_L2M_FLAG - OL_PAYMENT_L2M_FLAG ThePredictorVariables&TheVariableImportance only12outof50predictorvariablethatconsideredimportant
- 29. The Likelihoodof the SmartphoneAdopters 29 NB. The Generation Category based on: “Consumer Decision Journey, McKinsey Quarterly, by David Court, Dave Elzinga, Susan Mulder, and Ole Jørgen Vetvik, July 2009” AVERAGE LOS THAT SWITCH TO SMARTPHONE: 6.2 MONTHS, WITH AVERAGE AGE 48 YEARS OLD, AVERAGE USAGE 16 DAYS PER MONTH.
- 30. Tabular Output of The Model smartphonerecommendation=Yifadoptionprobability>0.5 30
- 31. 31 Business Use-Case & Action Non Smartphone Users Smartphone Adoption Predictive Model Business Use-Case Device Bundling Partnership Strategy Smartphone Adoption on Customers Long Term Package Offering (Bundling Contract Based)
- 32. 32 Revenue Uplift Projection NB : The revenue uplift projection can be higher with the broadband package (bundling contract based program). The other intangible benefit is enabler of the blue ocean business (IoT, M2M, Mobile Health Care, etc)
- 33. 33 BUSINESS USE CASE 4: Event Trigger Campaign for Preventing Customer Churn in Telecommunication Service
- 34. 34 Output Event Trigger Campaign Sample Churn Prevention Business Case Flow in TelecommunicationIndustry Ever Champion segment who has decreasing usage trend in last 3 consecutive months Free Voice Free SMS Free Internet Churn Potential BusinessCaseIllustration Action / Offer • Aggressive churn prevention program with interesting offer • Recharge bonus offering Target Customer • Ever Champion Segment customer, and • Downgrade usage in last 3 consecutive months Trigger • Customer browsing competitor website in specific page ( promo / price ), where the Freq. access > 3x / week
- 35. 35 EventTriggerCampaign–ChurnPrevention Campaign Information Campaign Trigger Channel : SMS Criteria Segments : ex.Champions who has decreasing usage trend in last 3 consecutive months Offers : Minimum recharge accumulation >= IDR 50 k get bonus free internet 2 GB Broadcast date : 7-30 Nov 2017 Wording : Nikmati GRATIS INTERNET 2 GB dengan isi ulang mulai Rp 50.000. Promo s.d 30/11/17. Cek bonus di *345#. S&K berlaku Objective : To prevent the ex.champion segment churns Mechanic : Offer recharge bonus 1 Ever Champion segment who has decreasing usage trend in last 3 consecutive months Customer browsing competitor website in specific page ( promo / price ), where the Freq. access > 3x / week
- 36. 36 Thank you……
- 37. 37 Questions ?
- 38. 38 [1] https://www.forbes.com/sites/gartnergroup/2013/03/27/gartners-big-data- definition-consists-of-three-parts-not-to-be-confused-with-three- vs/#6595be1942f6. Accessed on 7th November 2017. [2] Jacobs, A. 2009. The pathologies of big data. Communications of the ACM 52(8), 36-44. [3] http://www.bain.com/publications/articles/management-tools-customer- segmentation.aspx. Accessed on 7th November 2017. [4] Fader, P. S., Hardie, B. G., & Lee, K. L. (2005). RFM and CLV: Using iso-value curves for customer base analysis. Journal of Marketing Research, 42(4), 415-430. Reference