UCD Housekeeping Mid-term Exam PDF

d HOUSEKEEPING 01 Mid-term exam • • • • • Oct 31 (Tue) in class, 9:45am – 10:45am Paper and pencil, closed-book exam Covers anything we covered in classes including lecture slides and cases, excluding textbook, from Week 1 (What is brand) to Week 7 (Quantitative methods) Total points: 50 points...

d HOUSEKEEPING 01 Mid-term exam • • • • • Oct 31 (Tue) in class, 9:45am – 10:45am Paper and pencil, closed-book exam Covers anything we covered in classes including lecture slides and cases, excluding textbook, from Week 1 (What is brand) to Week 7 (Quantitative methods) Total points: 50 points • Multiple-choice questions: 25 points (18 questions) • Short essay questions: 25 points (4 or 5 questions) In case of any issues, contact me by Oct 24 ([email protected]) d HOUSEKEEPING 01 Mid-term exam (Continued) • Brands provide consumers with all of the following except • A) Brands provide a shortcut to consumers and simplify decision making • B) Brands reduce the risk associated with purchase • C) Brands help consumers identify the maker of a product • D) Branded products always provide superior quality d HOUSEKEEPING 02 Draft I of the Brand Audit Report • • 03 Not graded * Will share my thoughts on them ASAP Draft II of the Brand Audit Report • • • Not graded * Due November 14 (week 10) Parts I-IV of the report CLASS 6-1 QUALITATIVE RESEARCH METHODS f Learning objectives # Explain how qualitative research methods can help us understand consumer perception about a brand # Describe how automated text analyses can provide valuable consumer insights BRAND AUDIT PROJECT Example: lululemon PRODUCT § § § § Product: everything for athletic activities (A) Superior quality & technology (e.g., Mirror, patented fabric) (A) Fashionable design (A) Community-based (e.g., plus-size, environment) (C) PRICE § § Premium price range (A) Minimum discounting with environment-related excuse (A), (C) CHANNEL § § Primarily direct channels: company-operated stores (offline), direct to consumers (online) (A), (B), (C) Indirect channels (e.g., wholesale, studios), (B), (C) PROMOTION § § § Grassroots: Community-based marketing (e.g., in-store yoga class, local events, Mirror, sustainability, inclusiveness, global and store ambassadors), (C) Experiential marketing (e.g., experiential store, Mirror) (B) Provider of Canadian Olympics team’s uniform (A) POPs & POPs A. Athleisure Brand (athletic = performance, quality / leisure = fashionable) B. Experiential C. Community-based POSITIONING Lululemon is an athleisure brand that builds and maintains meaningful connections with others (i.e., community) by facilitating sweat and growth (i.e., experience) BRAND AUDIT PROJECT § Part I: Historical overview and background § Part II: How is the brand currently positioned (refer to Lululemon case) § Associations established through marketing programs (i.e., product, price, communication, & channel strategies), target consumer segment (e.g., location, income), (in)direct competitors § Part III: How do consumers perceive the brand (refer to Land Rover America) § § Is brand awareness high enough? Is perceived brand image consistent with positioning? Part IV: What types of marketing activities are (in)consistent § Identify reasons why there are inconsistencies between the current consumer perceptions and an ideal brand positioning: Lack of awareness? Failure in communication with consumers? § Part V: Make recommendations on what can be done (refer to The Hunger Games, Proctor & Gamble Always Russia) § How to increase awareness? How can marketing programs be revised to shape desired image? RESEARCH HELPS US FIGURE OUT § Awareness: Do customers recall or recognize our brand? * § Image: How is our brand perceived by customers? § Is consumer perception of our brand consistent with our positioning? * § Who (i.e., what types of consumers) likes our brand? § Demographic factors (e.g., gender, age, income level) § Psychographic factors (e.g., values, beliefs, lifestyle) OVERVIEW OF RESEARCH METHODS Qualitative methods Quantitative methods § Free association § Projective techniques § Ethnography, Netnography § § § § § § § Survey, Experiment Social media monitoring Text analysis I: Topic modeling Text analysis II: Sentiment analysis Neural research Eye tracking Facial electromyography CAN WE HAVE A MEANINGFUL CONVERSATION? Customers do not have an incentive nor capability to express true feeling § Consumers do not have an incentive to share their perception § Imagine you are a marketer for Durex (i.e., a condom brand). You want to understand customer perception of Durex. You recruit participants and ask their experience of using Durex. Do you expect them to talk to you honestly? § Consumers are not clear about how they feel about a brand § Imagine you are a marketer for Colgate (i.e., a toothpaste brand). You asked research participants why they started purchasing Colgate repeatedly. Do you expect them to know why? QUALITATIVE RESEARCH TECHNIQUES § Identify possible brand associations in non-intrusive ways § Diagnostic tools that can be used to uncover true opinions and feelings when consumers are unwilling or unable to express themselves § First step in exploring consumer perceptions § Unstructured measurement approaches FREE ASSOCIATION § Q: “What comes to your mind when you think of a brand?” § Simplest way to clarify and profile brand associations § Useful Questions § What do you like the best or least about the brand? § What do you find unique about the brand? § Concern: What types of probes to give to participants? PROJECTIVE TECHNIQUES Completion task (e.g., bubble task) § To fill in bubble based on what they believe is happening or being said in the scene PROJECTIVE TECHNIQUES Comparison task § To convey impressions by comparing brands to people, countries, animals, activities, occupations, cars, etc. Samsung LG Hyundai Lotte PROJECTIVE TECHNIQUES Third-person projection (1/5) § In 1949, Nescafé launched a new product: instant coffee § Product advantage: coffee in 3 mins § Despite successful pre-tests, sales did not take off. So they did a survey. § Why don’t you buy instant coffee? § Typical answer: “I don’t like the taste” PROJECTIVE TECHNIQUES Third-person projection (2/5) § Instruction (Haire 1950, Journal of Marketing) § Read the shopping list below. Try to project yourself into the situation as far as possible until you can more or less characterize the woman who does the groceries. Then write a brief description of her personality and character. Wherever possible indicate what factors influenced your judgment. PROJECTIVE TECHNIQUES Third-person projection (3/5) § Stimulus: Randomly presented one of the two below § 1½ pounds of hamburger § 1½ pounds of hamburger § 2 loaves of Wonder bread § 2 loaves of Wonder bread § 1 can Rumford’s Baking Powder § 1 can Rumford’s Baking Powder § Nescafe instant coffee § Maxwell coffee (Drip Ground) § 2 cans Del Monte peaches § 2 cans Del Monte peaches § 5 lbs. potatoes § 5 lbs. potatoes PROJECTIVE TECHNIQUES Third-person projection (4/5) § Results (content-coded): I think this woman is… Nescafe instant coffee Maxwell ground coffee Lazy 48% 4% Failing to plan household chores 48% 12% Spending too much 12% 0% Thrifty 4% 16% Not a good wife 16% 0% A good wife 4% 16% PROJECTIVE TECHNIQUES Third-person projection (5/5) ETHNOGRAPHY § Ethnographic research aims to extract and interpret deep cultural meaning of events and activities through research techniques such as consumer immersion, site visits, shopalongs, etc. § Con: subject to interpretation NETNOGRAPHY § Study actual communications on digital environments https://www.youtube.com/watch?v=SLC_sw4a1mM CLASS 6-2 AUTOMATED TEXT ANALYSES (QUANTITATIVE METHODS) SOCIAL MEDIA MONITORING Social media analytics service SOCIAL MEDIA MONITORING Social media analytics service § Provide a summary of key statistics associated with a brand § Topics related to a brand from millions of reviews in text format § Sentiment (positive or negative and to what extent) associated with social media messages or reviews in text format § Number of engagements (e.g., Facebook Likes or Twitter Re-tweets) § How are these insights generated? TOPIC MODELING Purpose § Topic modeling identifies (a) market trends and consumer opinions about a brand by extracting latent themes and topics from text data and (b) key words associated with each topic § What do consumers care about our brand? § What images are strongly associated with our brand? TOPIC MODELING Demonstration § Data: TripAdvisor reviews on hotels (N=20,491) § You can also collect your own text data by asking: § When it comes to X brand, what comes to your mind? Please describe your thoughts and feelings about X brand in as much detail as possible. § You need a “big” N § Use R, Python with codes from online sources TOPIC MODELING How it works § Step 0: Load text data (document #1 out of 20,491 documents) § nice hotel expensive parking got good deal stay hotel anniversary, arrived late evening took advice previous reviews did valet parking, check quick easy, little disappointed non-existent view room room clean nice size, bed comfortable woke stiff neck high pillows, not soundproof like heard music room night morning loud bangs doors opening closing hear people talking hallway, maybe just noisy neighbors, aveda bath products nice, did not goldfish stay nice touch taken advantage staying longer, location great walking distance shopping, overall nice experience having pay 40 parking night. TOPIC MODELING How it works § Step 1: Remove non-words (numbers, emojis, punctuation) § nice hotel expensive parking got good deal stay hotel anniversary, arrived late evening took advice previous reviews did valet parking, check quick easy, little disappointed non-existent view room room clean nice size, bed comfortable woke stiff neck high pillows, not soundproof like heard music room night morning loud bangs doors opening closing hear people talking hallway, maybe just noisy neighbors, aveda bath products nice, did not goldfish stay nice touch taken advantage staying longer, location great walking distance shopping, overall nice experience having pay 40 parking night. TOPIC MODELING How it works § Step 2: Remove stopwords (prepositions, articles, etc.) § nice hotel expensive parking got good deal stay hotel anniversary arrived late evening took advice previous reviews did valet parking check quick easy little disappointed non existent view room room clean nice size bed comfortable woke stiff neck high pillows not soundproof like heard music room night morning loud bangs doors opening closing hear people talking hallway maybe just noisy neighbors aveda bath products nice did not goldfish stay nice touch taken advantage staying longer location great walking distance shopping overall nice experience having pay parking night TOPIC MODELING How it works § Step 3: Remove meaningless words (e.g., yay, tho) § nice hotel expensive parking good deal stay hotel anniversary arrived late evening advice previous reviews valet parking check quick easy disappointed existent view room room clean nice size bed comfortable woke stiff neck high pillows soundproof heard music room night morning loud bangs doors opening closing hear people talking hallway noisy neighbors aveda bath products nice goldfish stay nice touch advantage staying longer location great walking distance shopping overall nice experience pay parking night TOPIC MODELING How it works § Step 4: Remove certain key words (e.g., hotel, hilton) § nice hotel expensive parking good deal stay hotel anniversary arrived late evening advice previous reviews valet parking check quick easy disappointed existent view room room clean nice size bed comfortable woke stiff neck high pillows soundproof heard music room night morning loud bangs doors opening closing hear people talking hallway noisy neighbors bath products nice goldfish stay nice touch advantage staying longer location great walking distance shopping overall nice experience pay parking night § Step 5: Parse text into discrete words TOPIC MODELING How it works § Step 6: Find recurring patterns in co-occurrence of words TOPIC MODELING How it works § Step 7: Determine the number of topics with the best fit § Minimum density within topic and minimum within-topic divergence § Maximum cross-topic divergence TOPIC MODELING How it works § Step 8: Topic modeling with the optimum number of topics food staffs location room condition facilities reviews about cost efficiency check-in and out SENTIMENT ANALYSIS Purpose § Assess how positively or negatively consumers feel about a brand, product, or service through sentiment reflected in text, as compared to that toward competing brands § Effectiveness § Non-intrusive way to assess consumer opinion, hence more likely to provide a “true” consumer opinion than a survey § Sentiment reflected on online reviews predict movie sales, stock market returns, etc. in the future SENTIMENT ANALYSIS Demonstration § Data: TripAdvisor reviews on hotels (N=20,491) § LIWC (Linguistic Inquiry and Word Count) Software § Free resource: http://textanalyzer.org (awesome website!) § Simply upload a .csv or .xlsx file and you’ll get sentiment scores for each text document which are calculated automatically § This website was developed and is provided by world-renowned marketing researchers on word-of-mouth SENTIMENT ANALYSIS How it works § Step 0: Load text data § nice hotel expensive parking got good deal stay hotel anniversary, arrived late evening took advice previous reviews did valet parking, check quick easy, little disappointed non-existent view room room clean nice size, bed comfortable woke stiff neck high pillows, not soundproof like heard music room night morning loud bangs doors opening closing hear people talking hallway, maybe just noisy neighbors, aveda bath products nice, did not goldfish stay nice touch taken advantage staying longer, location great walking distance shopping, overall nice experience having pay 40 parking night. SENTIMENT ANALYSIS How it works § Step 1: Parse text into discrete words § nice hotel expensive parking got good deal stay hotel anniversary, arrived late evening took advice previous reviews did valet parking, check quick easy, little disappointed non-existent view room room clean nice size, bed comfortable woke stiff neck high pillows, not soundproof like heard music room night morning loud bangs doors opening closing hear people talking hallway, maybe just noisy neighbors, aveda bath products nice, did not goldfish stay nice touch taken advantage staying longer, location great walking distance shopping, overall nice experience having pay 40 parking night. SENTIMENT ANALYSIS How it works § Step 2: Identify sentiment-bearing words based on dictionaries § nice hotel expensive parking got good deal stay hotel anniversary, arrived late evening took advice previous reviews did valet parking, check quick easy, little disappointed non-existent view room room clean nice size, bed comfortable woke stiff neck high pillows, not soundproof like heard music room night morning loud bangs doors opening closing hear people talking hallway, maybe just noisy neighbors, aveda bath products nice, did not goldfish stay nice touch taken advantage staying longer, location great walking distance shopping, overall nice experience having pay 40 parking night. SENTIMENT ANALYSIS How it works § Step 3: Assign a sentiment score to each word § 30 -40 20 nice hotel expensive parking got good deal stay hotel anniversary, arrived late evening took advice previous reviews did valet parking, check quick easy, little 30 20 disappointed non-existent view room room clean nice size, bed comfortable woke 10 stiff neck high pillows, not soundproof like heard music room night morning loud -20 bangs doors opening closing hear people talking hallway, maybe just noisy 30 30 neighbors, aveda bath products nice, did not goldfish stay nice touch taken 30 40 advantage staying longer, location great walking distance shopping, overall nice experience having pay 40 parking night. SENTIMENT ANALYSIS How it works § Step 4: Compute overall sentiment score for each document § Overall number of words for the document: 87 § Number of words related to positive sentiment in the document: 9 § Number of words related to negative sentiment in the document: 2 § Overall positivity score: (30+20+30+20+10+30+30+30)/87 = 2.29 § Overall negativity score: (-40-20)/87 = -0.68 § Overall sentiment of the document: 2.29-0.68 = 1.61 SENTIMENT ANALYSIS How it works § Step 5: Compute overall sentiment score from entire data § Overall sentiment score of document #1 on Lululemon: 1.61 § Overall sentiment score of document #2 on Lululemon: 2.34 § Overall sentiment score of document #3 on Lululemon: -1.65 § Overall sentiment score of document #1 on UnderArmour: 0.89 § Overall sentiment score of document #2 on UnderArmour: -0.22 0.62 § Overall sentiment score of document #1 on UnderArmour: 1.20 0.76 AUTOMATED TEXT ANALYSIS § Limitation § The programs cannot identify sarcasm, context, irony, or idioms § Example: Cynical attitude toward a brand (e.g., “Yeah, I bet I will come back to this restaurant”) is not properly reflected in outcome § Benefit § Process massive amount of text data very quickly through automation d NEXT CLASS 01 Quantitative research methods