MAKING SENSE OF BIG DATA. Making Sense of Big Data 1

Transcription

1 MAKING SENSE OF BIG DATA Making Sense of Big Data 1

2 HELLO! NICE TO MEET YOU! Shingly Lee Workshop Coordinator Brand Lover Eclectic foodie on Amy Martin Workshop Coordinator Fascinated by the latest and greatest Walmart (starting 2015) Travel Enthusiast The brands we represent: Making Sense of Big Data 2

3 AND A SPECIAL GUEST! DR. CEREN KOLSARICI Assistant Professor, COMM 433 Marketing Analytics Ian R. Friendly Fellow of Marketing New Researcher Achievement Award Distinguished as the American Marketing Association- Sheth Consortium Fellow Ph.D. in Marketing, Mcgill University Making Sense of Big Data 3

4 1 The Big Data Movement 4 Case Studies ft. Dr. Ceren Kolsarici UNLOCKING BIG DATA 2 What s the Big Deal? 3 How Can Analytics Help? Making Sense of Big Data 4

5 Why NOW? Making Sense of Big Data 5

6 CONSUMER LANDSCAPE IS CHANGING 1. THE MULTITASKERS Making Sense of Big Data 6

7 CONSUMER LANDSCAPE IS CHANGING 2. MARKET FRAGMENTATION Making Sense of Big Data 7

8 CONSUMER LANDSCAPE IS Jeff Weiner, CEO of Phil Libin, CEO of Evernote 3. THE EMPOWERED CONSUMER Making Sense of Big Data 8

9 Making Sense of Big Data 9

10 DIGITAL IS THE NEW NORMAL Making Sense of Big Data 10

11 MEDIA FACEOFF TRADITIONAL 111.5M VIEWS Making Sense of Big Data 11

12 MEDIA FACEOFF NEW MEDIA 2.1B VIEWS Making Sense of Big Data 12

13 MOBILE RULING YOUR WORLD Making Sense of Big Data 13

14 DATA EXPLOSION 90% OF THE DATA IN THE WORLD TODAY HAS BEEN CREATED IN THE LAST TWO YEARS ALONE Making Sense of Big Data 14

15 Making Sense of Big Data 15

16 BIG(GER) DATA Making Sense of Big Data 16

17 THE FOUR V S OF BIG DATA Making Sense of Big Data 17

18 Making Sense of Big Data 18

19 Making Sense of Big Data 19

20 3 TYPES OF ANALYTICS Making Sense of Big Data 20

21 DESCRIPTIVE ANALYTICS Making Sense of Big Data 21

22 PREDICTIVE ANALYTICS Making Sense of Big Data 22

23 PRESCRIPTIVE/NORMATIVE ANALYTICS Making Sense of Big Data 23

24 SOPHISTICATION CURVE Making Sense of Big Data 24

25 BOTTOM LINE IMPACT Zone of Death Wish Marketing Zone of Exceptional Marketing (Well Below Average) (Below Average) (Average marketing program) (Above Average) (Well Above Average) Marketing Performance Critical Troubling Average Pleasing Amazing Marketing Share Growth Precipitous Significant Modest Increase Dramatic Decline Decline Decline Increase New Product Success Rate 0% 5% 10% 25% 40%+ Advertising ROI Negative 0% 1-4% 5-10% 20% Promotional Programs Disaster Un - profitable Marginally Unprofitable Profitable Very Profitable Customer Satisfaction 0-59% 60-69% 70-79% 80-89% 90-95% Customer Retention/Loyalty 0-44% 45-59% 60-74% 75-89% 90-94% Making Sense of Big Data 25

26 CASE #1 GENERATING THE PERFECT CONTENT: NETFLIX

27 SOME CONTEXT In Q3 of 2011, Netflix announced continuing its DVD service under the name Qwikster and a price increase for its streaming service.

28 ORIGINAL CONTENT STRATEGY To keep and grow its subscriber base, repair the damaged brand name Netflix turned to a brand-new strategy: creating its original content. Create the perfect TV-series. Netflix did not need a fortune teller to see how successful their new show would be. They knew! Even before anyone shouted "action." But how? GOAL

29 ACCESS TO DELUGE OF INFORMATION No one in the industry knows more about the audiences than Netflix 33 million subscribers worldwide 30 million plays a day: when you pause, rewind and fastforward, star ratings, searches, time and day, devices Tags for the movies and TV shows: Genre, cast, award nomination, length, production studio, etc. Traditionally Match available shows with audiences based on preferences (i.e. Recommendations) New Design original content (Why not?) DATA

30 HOW DID NETFLIX KNOW THE RECIPE FOR SUCCESS? No primary data collection: audience testing, market research, focus groups, etc. By being a great "data detective" Let the data predict what people would like based on their past viewing habits Mine extremely rich data to generate actionable insights But how? Looking for correlated patterns of behaviour across individuals MODEL Source: "Giving Viewers What They Want," by David Carr, New York Times, Feb. 24, 2013

31 OVERALL IMPACT House of Cards was the first original series by Netflix Political drama based on BBC mini-series of the same name Costs $100m for two seasons The show quickly became critics and audiences favorite First season: 13 episodes, February 1, 2013, 9 Emmy and 14 GG nominations Second season: 13 episodes, February 14, 2014, 13 Emmy nominations IMDB rating: 9.1 It would only make sense to invest if audience likes it and Netflix can get new subscribers (i.e. 500K new subscriptions in two years to break even) Did the strategy pay off? INSIGHTS

32 GENERAL FRAMEWORK ADAPTED TO NETFLIX What is the business/marketing question you want to answer? E.g. What type of show should Netflix invest in developing that will appeal to subscribers and attract new ones? What data do you have available (or can obtain) to help you answer this question? E.g. Netflix has data on viewing habits of subscribers and what their portfolios of shows viewed look like. What do you do with the data to help you answer your question? E.g. Netflix looks for patterns in viewing habits, correlation analysis. What does the analysis reveal? E.g. A sizeable segment of subscribers who watch political thrillers also watch Kevin Spacey movies and David Fincher movies. They also watch an old British miniseries called House of Cards. What is the marketing decision driven by the data analytics? Create a political thriller and involve Kevin Spacey and David Fincher U.S. series version of the old British miniseries House of Cards. Source: Adapted from Andrew Stephen, Social Media Analytics Course, Katz Graduate School of Business, University of Pittsburg

33 CASE #2 THE KEY DRIVERS FOR SUCCESS FOR PRODUCTS WITH SEQUENTIAL DISTRIBUTION: ADVERTISING AND WOM SYNERGIES "Dynamic Effectiveness of Advertising and Word-of-Mouth in the Sequential Distribution of Short Life Cycle Products," Norris I. Bruce, Natasha Zhang, Ceren Kolsarici, Journal of Marketing Research, 2012, 49(4),

34 SEQUENTIAL DISTRIBUTION Windowing or sequential distribution is most prevalent for new products with short life cycles. Motion pictures, book publishing, fashion, music and art Revenues from sequential distribution are crucial for firms: Hollywood studios, on average, spend $71M to produce and $36M to market a film A movie on average only makes $47M theatrical revenues

35

36 WHEN AND HOW MUCH TO ADVERTISE FOR A MOVIE? The two key drivers for movie revenues are advertising and third-party reviews (e.g. critics reviews and WOM) How do ad effectiveness and WOM effectiveness fluctuate between box office to rental stages of a movie? How do they differ and interact? How do they vary across different movies? Is there a better way to allocate advertising resources? GOAL

37 HOW IS THE DATA STRUCTURED? For both theatrical and video stages Revenues ($) Advertising Spending ($) IMDB ratings (Volume & Valence) Critics reviews (Valence) Movie specific variables Genre, Runtime, Big Studio, Oscar Nominations, Sequel, Budget etc. DATA

38 DESCRIPTIVE ANALYTICS Firms can use advertising and WOM strategically to support and elevate each other's effectiveness at different stages of the PLC. Theater Stage Video Stage DATA

39 DESCRIPTIVE ANALYTICS MODEL

40 PREDICTIVE ANALYTICS: AD AND WOM SYNERGY IN ACTION Diminishing ad effectiveness over time Advertising wear-in possible, particularly for new products Advertising and WOM exert independent yet interdependent influences on demand Higher ad elasticities early on in PLC replaced by higher WOM elasticities later INSIGHTS

41 NORMATIVE ANALYTICS More efficient media planning would generate greater profits 30%(70%) of films could designate lowerthan-observed (higher-than-observed) theatrical ad budget 17% (73%) of films could designate lowerthan-observed (higher-than-observed) video ad budget Recommended pre- versus post release budget split. Critics' favourites (allot more to pre-) Action (allot less to pre-) Up to 15% increase in log revenues with the new allocation pattern. INSIGHTS

42 THANKS FROM QMAA! Making Sense of Big Data 42

43 QMAA GIVES YOU COOL THINGS THAT WILL HELP YOU IN LIFE 1 Resume tip sheet 2 Cover letter tip sheet 3 LinkedIn tip sheet 4 What is due? When? 5 The T-Chart That Will Change Your Life Making Sense of Big Data Recruiting s Dirty Little Secrets 43 43

44 WHAT IS A CASE ANYWAYS? SO YOU WANT TO BE A MEMBER? Making Sense of Big Data Recruiting s Dirty Little Secrets 44 44

45 WHAT IS A CASE ANYWAYS? WOAH, HOLD UP WHAT AM I SIGNING UP FOR? QMAA MEMBERS: ALUMNI MEMBERSHIP (Coffee chats, mock interviews) MARKETING CAREER NIGHT AGENCY NIGHT FIRM VISITS MONTHLY NEWSLETTER TO KEEP YOU IN THE KNOW INCLUSION IN EXCLUSIVE QMAA RESUME BOOK Making Sense of Big Data Recruiting s Dirty Little Secrets 45 45

46 WHAT IS A CASE ANYWAYS? ALRIGHT TIME FOR THE NITTY GRITTY 1 Hit up 2 Fill out our Google form. It s fast and easy 3 Send your resume to membership@qmaa.com in PDF form Making Sense of Big Data Recruiting s Dirty Little Secrets 46 46

47 OF COURSE, Talk to ANYONE on the QMAA EXEC to find out more. SERIOUSLY. Making Sense of Big Data Recruiting s Dirty Little Secrets 47 47

48 Because this is US when we help YOU Making Sense of Big Data Recruiting s Dirty Little Secrets 48 48