Real-Time Customer Intelligence: The New Heartbeat for Growth and Profitability

1,231 views

Published on

Consumerization is driving transformational change for every enterprise and therefore, customer intelligence is a core capability for every company to compete in the New Era.

Published in: Technology, Business
1 Comment
2 Likes
Statistics
Notes
  • Hi Balu, this is a very insightful presentation. Would it be possible for you to share this document with satish.kalpathy@gmail.com? thanks
       Reply 
    Are you sure you want to  Yes  No
    Your message goes here
No Downloads
Views
Total views
1,231
On SlideShare
0
From Embeds
0
Number of Embeds
1
Actions
Shares
0
Downloads
0
Comments
1
Likes
2
Embeds 0
No embeds

No notes for slide
  • With the Cetas Analytics Solution, there are 3 things that are important to note :1. Single interface to handle a variety of data feeds at different velocities including live data streams or otherwise. A key point here is that there is zero ETL, No copying of data and no schema is req'd.2. The Cetas real-time analytics engine can automatically correlate billion plus events per day across multiple dimensions of data coming from a variety of sources.3. The third point to note is that our solution can surface insights automatically or thru exploration for you to take immediate action.
  • Real-Time Customer Intelligence: The New Heartbeat for Growth and Profitability

    1. 1. Real-time Customer Intelligence:The New Heartbeat For Growth & Profitability INSTANT INTELLIGENCEBalu Rajagopalbrajagopal@vmware.comVMware and Cetas Confidential; Do NOT DistributeWeb: www.cetas.netTwitter: @CetasAnalyticsBlog: www.cetas.net/blogYouTube: www.youtube.com/CetasAnalytics © 2009 VMware Inc. All rights reserved
    2. 2. What is Big Data? DATAVOLUME 2.0 Zettabytes in 2011 Enterprise Data MachineZettabyte To Machine Exabyte Petabyte Interactions Terabyte Transactions Mainframe PC Internet Mobile Machine TimeChart based on IDC and UC Berkeley Data Growth Estimates, Source: IDC & CosmoBC.com:http://techblog.cosmobc.com/2011/08/26/data-storage-infographic/ 2
    3. 3. What is “Big Data Analytics”? Volume Velocity Variety Value $From Terabytes to 10’s of Billions Multi-Structured Business Petabytes of Daily Records Insights ADAPTABILITY SCALABILITY FLEXIBILITY ACTIONABILITY 3
    4. 4. The Perfect Storm for Retailers & Brands MARKET FORCES BUSINESS CHALLENGES① Consumerization ① “Show Rooming”② Mobility & Connectivity ② Margin Pressure③ Purchase Choices ③ Demand Predictability④ Social Influences ④ Customer Defection⑤ Price Sensitive Economy ⑤ Differentiation Difficulties 4
    5. 5. Path To Retail Growth & Profitability BUYER Individual Search Who, What, How Micro Segments Learn Macro Consult Real-Time Customer Intelligence What/When/Whom? Multi-dimensional Buy Stock Patterns/Trends Tell Offer Activity Tracking Influence Monitoring SELLER SHOPPER 5
    6. 6. Big Data Analytics Retail ContextShopper  Customer centric: Who is the customer, his/herBehavior Analysis patterns, trends, likes, dislikes, frequencies, ...Product Mix  Customer-Product relationship: CategoriesAnalysis interested? products/games/apps favored? Ad potential? ….Market Basket  Customer-Product relationship: CategoriesAnalysis interested? products/games/apps favored? Ad potential? ….Social & Other  Social: Who’s influencing whom? Network size?influences Type of interactions? …Instant Decision  Actionable insights: offerings to present? What toMaking stock? Ad placement? Purchase Likelihood? … 6
    7. 7. SOLUTION REAL-TIME CUSTOMER INTELLIGENCE USING BIG DATA ANALYTICS7
    8. 8. Customer Intelligence Frame ADAPTED FROM INSTANT INTELLIGENCE 8
    9. 9. About the Big Retailer Revenue • Top 20 Global Retailer • Generates over 10 billion in revenue worldwide • Average margin per transaction - ~ 10% • Average Market Basket Size - ~ $ 50 Product Categories and SKUs • Over 10,000 product categories • Millions of SKUs, 1000s of stores Shopper Base • Millions of registered online and mobile app shoppers 9
    10. 10. Questions That Needed Answers (in Real-time, with Drill Down) Behavioral • Do Women generate more revenue than men ? • Do Women buy products that are of higher margin than men ? • Does Demographic patterns for age, income and household size roughly match US population (e.g., % with income $100K+)? Online Shopping Trends • Is Average weekend revenue higher than weekday revenue? 10
    11. 11. Retailer Dataset Summary① Time series data of Click & Purchase activity by user ID for April 2012② Customer master, customer behavioral interests, behavioral interest taxonomy③ Mobile App & Website analytics (real-time streams)④ Loyalty Data⑤ Number of shopper related events: Over 10M 11
    12. 12. Click and purchase activity DataColumn Field Name Description Value 1 Event Timestamp Timestamp of click or purchase event. m/d/yyyy h:mm 2 User ID Unique user ID 6 digit numeric 17 and under 18-24 25-44 45-64 3 Age Group Quintile of age range 65 and over $0-$25,000 $25,000- $50,000 $50,000- $75,000 $75,000- $100,000 4 Income Group Quintile of household income range $100,000 and over Male Female 5 Gender Shoppers gender Unknown East Central 6 Region Shoppers georaphic location in United States West 7 Household Size Quintile of household size 1-5 1 if click event 8 Click Event Event is a click 0 if not click event 1 if purchase event 9 Purchase Event Event is a purchase 0 if not purchase event 10 Revenue Total shopping cart revenue if purchase event USD 11 Margin Gross margin after COGS if purchase event USD12
    13. 13. Customer Master DataColumn Field Name Description Value 1 User ID Unique user ID 6 digit numeric 2 Member Since Date of initial membership m/d/yyyy h:mm Silver Gold Platinum 3 Loyalty Level Membership class Unknown Bank Transfer Paypal Credt Card 4 Preferred Payment Method How customer typically pays Unknown Poor Fair Good 5 Payment History Credit history with vendor Unknown Low Medium High 6 Promotion Receptivity Demonstrated openness to offers Unknown No Yes 7 Mobile App Download Downloaded vendor app Unknown13
    14. 14. Customer Behavioral Data Interest Category Arts & Entertainment  Taxonomy: Autos & Vehicles Beauty & Fitness Books & Literature Business & Industrial • 25 top level categories Computers & Electronics Finance • 249 sub categories Food & Drink Games Hobbies & Leisure Home & Garden  One shopper can have multiple Internet & Telecom interests Jobs & Education Law & Government News Online Communities People & Society Pets & Animals Real Estate Reference Science Shopping Sports Travel World Localities14
    15. 15. Real-time Web Analytics DataColumn Field Name Description Value 1 Hour Summary Hourly bucket for site analytics aggregations m/d/yyyy h:mm 2 Average Revenue Per Visit Average transaction revenue per site visit USD 3 Average Item Value Average value of items checked out USD 4 Average Num Orders Per Visit Average number of orders per site visit Numeric 5 Average Num Items Per Order Average items in cart at checkout Numeric 6 Shopping Cart Abandonment Rate Percent of users who add to cart but dont check out % 7 Shopping Cart Session Percent Percent of users who add at least one item to cart % 8 Average Time On Site Average time spend shopping Minutes 9 Average PVs Per Visit Average page views Numeric 10 Percent Single Page Visits Bounce rate % 11 Percent New Users Percent first time users %15
    16. 16. Customer Intelligence THE ANALYSIS16
    17. 17. Customer Intelligence Analysis Elements① Multi-dimensional Views of Shopper② Shopper Activity Analysis③ Website Analysis④ Loyalty Analysis (Using Batch Query)⑤ Real-Time Dashboard Views17
    18. 18. MULTI-DIMENSIONAL VIEW OF SHOPPER18
    19. 19. Multi-Dimensional View Analysis Steps Charts & Tables • Measures of Interest: Revenue, margin • Break down by dimension of interest • Age, gender, HHsize, income group, or region Time trends • Measure of Interest: sum of revenue (could also do avg margin) • Started with daily the use time pivot to drill down • Break down by dimension of interest • Age, Gender, HHsize, Income group, or region Custom measures • Select a custom measure and operate on it like regular measure19
    20. 20. Multi-dimensional analysis Total revenue by gender and region20
    21. 21. Multi-dimensional analysis with filteringTotal revenue by income group and region, with filters:Age group = 18-24Gender = Female 21Household Size = 4
    22. 22. SHOPPER ACTIVITY ANALYSIS22
    23. 23. Shopper Activity Analysis Steps Summary aggregates • Look at revenue broken down age, gender, other dims of interest Uniques aggregates • Look at unique counts of shoppers by day • Remove time trend and break down by age, gender, other dims of interest Dashboards • Leverage pre-defined dashboards and review • Demo and geo theme • Revenue theme • Shopper theme 23
    24. 24. Activity analytics – simple time trend Weekend Total revenue (USD) by day. Weekend revenue is approx 20% lower than weekdays.24
    25. 25. Custom measures - create Compute margin % from margin and revenue measures25
    26. 26. Custom measures - chart Custom measure margin % broken down by gender. Males generate lower margins.26
    27. 27. Aggregates summary chart (1) Avg. monthly transaction revenue (USD) by age and gender. Male revenue is significantly lower.27
    28. 28. Aggregates summary chart (2) Avg. margin (USD) by gender for three days in April. Male revenue is significantly lower on each day.28
    29. 29. REAL-TIME WEBSITE ANALYTICS29
    30. 30. Real-Time Website Analytics Steps Hourly aggregations from web site analytics tool that shows shopping metrics of interest Use the date field called “Hour Summary” Correlate values from Cyber Monday industry reports (e.g., shopping cart abandonment rate)30
    31. 31. Unique Daily Visitors by Gender Unique daily visitors to site by gender.31
    32. 32. Unique Daily Visitors by Age Range Unique daily visitors to site by age range.32
    33. 33. Unique Monthly Visitors by Income Group Unique monthly visitors to site by income group. 33
    34. 34. Unique Monthly Visitors by Purchase Made Unique visitors to site who made a purchase on any day. 34
    35. 35. Out-of-Box Vertical Insights Dashboard35
    36. 36. Loyalty Analysis (Batch Query)36
    37. 37. Loyalty Analysis Using Batch Query Query “Revenue by Loyalty Level” - a simple join of event level data stream with customer master (profile) data Query “Join with Customer Master” and customer profile dimensions Query “Join with Customer Interests” - a more complex multi-join with customer master and interest categories.37
    38. 38. Batch query definitionData sources Drag and drop canvas Expression builder38 HiveQL expression
    39. 39. Batch query result Joining event level data with customer master (profile) data Most revenue comes from “Silver” members39
    40. 40. Batch query result (2) Combining dimensionsfrom event level data andcustomer master (profile) data Most revenue comes from $25k-$50k group 40
    41. 41. Batch query result (3) Female top interests by revenue Male top interests by revenue41
    42. 42. Batch query result (4) Good credit top interests by revenue Poor credit top interests by revenue Joining event data with in-house profile data helps you understand your customers more.42
    43. 43. REAL-TIME DASHBOARD VIEWS43
    44. 44. Dashboard (1) Demographic and Geo Focus44
    45. 45. Dashboard (2) Revenue Focus45
    46. 46. Dashboard (3) Shopper Focus46
    47. 47. Customer Intelligence APPROACH, ANALYTIC FINDINGS, TAKEAWAYS47
    48. 48. Questions That Needed Answers (in Real-time, with Drill Down) Behavioral • Do Women generate more revenue than men ? • Do Women buy products that are of higher margin than men ? • Does Demographic patterns for age, income and household size roughly match US population (e.g., % with income $100K+)? Online Shopping Trends • Is Average weekend revenue higher than weekday revenue? 48
    49. 49. The Answers (From Analytics) Behavioral • Women generate more revenue than men • Women buy products that are of higher margin than men • Demographic patterns for age, income and household size roughly match US population (e.g., % with income $100K+) Online Shopping trends • Average weekend revenue is about 20% lower than weekday revenue (Correlates with third-party data showing people shop at work)49
    50. 50. Interesting AssociationsBehavioral preferences by gender Females prefer: 302 Shopping Apparel 310 Shopping Luxury Goods 315 Shopping Toys 135 Beauty & Fitness Face & Body Care 136 Beauty & Fitness Fashion & Style 138 Beauty & Fitness Hair Care 213 Home & Garden Bed & Bath 222 Home & Garden Kitchen & Dining 102 Arts & Entertainment Entertainment Industry 100 Arts & Entertainment Celebrities & Entertainment News 141 Books & Literature Childrens Literature 197 Games Online Games Male prefer: 317 Sports College Sports 322 Sports Motor Sports 323 Sports Sport Scores & Statistics 324 Sports Sporting Goods 173 Computers & Electronics Consumer Electronics 207 Hobbies & Leisure Outdoors 185 Finance Investing 227 Home & Garden Yard & Patio 219 Home & Garden Home Improvement 121 Autos & Vehicles Motorcycles 125 Autos & Vehicles Trucks & SUVs 114 Autos & Vehicles Boats & Watercraft 50
    51. 51. Interesting Associations Behavioral preferences in dataset by payment history:Payment History = Poorprefer: 181Finance Credit & Lending 242Law & Government Military 318Sports Combat Sports 322Sports Motor Sports 100Arts & Entertainment Celebrities & Entertainment NewsPayment History = Goodprefer: 281Real Estate Real Estate Listings 282Real Estate Timeshares & Vacation Properties 330Travel Air Travel 334Travel Cruises & Charters 337Travel Specialty Travel51
    52. 52. The Analytics Solution Semi- Structured structured Products Transactions Logs E-mails In-Apps Sensors …. …. Unstructured Social Audio Photo & Video Inventory Ad Impr., Clicks, Conv. • Zero ETL  Volume  • No Copy Velocity Cetas Real-time  Variety • No DW Correlate Analytics  Variance • No Schema Billion+ Events per day Customer Intelligence 52
    53. 53. 5 Takeaways – How To Extract Customer Intelligence CONNECT THE DOTS (DECISION CONTEXT) Profiles, Sessions, Activities, Items in Baskets/Carts, Purchases, …① Big Data – Source, Know, Manage, & Govern your big data② Customer – Reconcile different versions of the same customer③ Collaboration & Sharing – Enrich with third-party and partner data④ Data-Driven Decision Making (DDM) – Start w/small project successes⑤ Collective Wisdom – Machine plus Human Intelligence Still Required ! 53
    54. 54. Cetas User Interface 54
    55. 55. Sign up today at www.cetas.net!55
    56. 56. DROP BUSINESS CARD AT OUR BOOTH WE ARE GIVING AWAY AN IPAD BE PRESENT TO WIN DRAWING ON FRIDAY AT 3:00pm56 WWW.CETAS.NET

    ×