BBM016 Consumer Insights PDF

BBM016 Consumer Insights Janina Steinmetz Today ETHICS SECONDARY DATA SURVEYS Primary versus secondary data recap What are the differences between the two? What are the advantages and disadvantages of each? Doing research ethically: Principles Respect Make sure participation...

BBM016 Consumer Insights Janina Steinmetz Today ETHICS SECONDARY DATA SURVEYS Primary versus secondary data recap What are the differences between the two? What are the advantages and disadvantages of each? Doing research ethically: Principles Respect Make sure participation is voluntary Beneficence Do no harm Justice Be fair Doing research ethically Informant autonomy Explain what you’ll ask them about, how long it will take, etc. Confidentiality Informants must be anonymous (don’t ask for personal data, e.g., addresses, birthdates, etc.) Only us the data for your research, don’t show them to anyone else Voluntary participation Informants can stop the survey or interview anytime Doing research ethically Don’t try to squeeze the participant! Treat them like you want to be treated Respect, honesty Ask only the information than you need Don’t make their task more tedious than necessary Be professional (e.g., on time, have your recording device ready, etc.) Don’t: Ask about race, religion, sexual orientation! Ask about political beliefs or about illegal activities (e.g., drugs) Ask about health information (e.g., illnesses) Ask persons under the age of 18 Use deception Use questions with the potential for ‘labelling’ by the researcher or participant (e.g. ‘I am ugly’) Create the potential for psychological stress, anxiety, humiliation or pain Research Ethics: Data octopus Protect your participants’ data! Don’t share them, don’t send via email Research Ethics documents On Moodle, Market Research, week 3, you fill find 3 templates Participant information (Mandatory) Consent (Mandatory) Flyer (Optional) Every respondent MUST read the participant information sheet Online: include a sentence, for example: “By proceeding with this survey, you consent to these conditions.” Interview: record informants’ explicit consent Diversity in marketing Industry Example: Dove’s problematic campaign How could they get it so wrong? What about the ads you get? Look at your social media feed and see what ads you get Do you feel they are diverse and inclusive? Discuss with your neighbor! Diversity in your research Think about the diversity of your informants (e.g., in your target segment) Sampling Question asking … Secondary data Answer certain research questions directly Complement primary research Diagnose & provide background information to the research problem Help to refine objectives Help to design primary research (e.g., selecting a research approach, methods, sampling) Help understand/interpret results from primary research Confirm/validate results of primary research External secondary data sources External: Business Intelligence (BI) converges different data sources Syndicated report (e.g., Nielsen) Academic (e.g., on Google Scholar) Industry (trade organizations) Government (e.g., census) Blogs Web analytics Google search terms Where to start Check the library for syndicated reports, academic sources etc. https://libraryservices.city.ac.uk/ Check what internal data the startup has (e.g., pitch deck) Think about what information you need and check internal and external sources Decide which information is missing and what you can collect with primary research Secondary data Strengths: ▪ Weaknesses: Cheap ▪ Might be outdated Quick ▪ Might not be objective Easy ▪ Might not be what you need for your question Can be of great quality ▪ Can be of bad quality Limit to relevant sources Be critical with secondary data Source Fact versus opinion Date Academic versus industry Try to determine the quantity and quality of the collected data (large sample, valid method, etc.) Why when who what where how Useful additional resources https://ourworldindata.org/diet-compositions A fun source of secondary data Go to https://trends.google.com/trends/ What’s your favorite product or brand? Search in google trends: When in the year do people search for it the most/the least? Where in the world is the search volume for it greatest? How does that compare to a competitor product? A fun source of secondary data Go back to your questions for Lizzie Which of these could be addressed by data available from https://trends.google.com/trends/ ChatGPT You CAN use it to get started on a new topic, to get a summary, etc. BUT: it sometimes makes up things, so you need to verify everything you take from it Example: You: ChatGPT, what’s the percentage of vegans in London? ChatGPT: about 5% We’ll need to verify: Where does your information come from? You can’t just cite ChatGPT, but the source it uses, to make sure the information is accurate Survey research Survey research Structured Data Collection Use of a formal questionnaire that presents questions in a prearranged order. Fixed response alternative questions: Questions that require respondents to choose from a set of predetermined answers Advantages of Survey Methods Deliver precise E.g., 70% of respondents like coffee numerical estimates E.g., satisfaction with a product in increasing/decreasing Reduces bias caused by individuals (individual Objectivity managers, interviewers, interviewees, etc.) Helps overcome individual limitations Questionnaire is simple to administer, analyze, Statistics allow complex analyses interpret Disadvantages of Survey Methods Explains the “What” but not the “Why” Survey cannot reveal what you don’t ask about Rely on self-report data E.g., respondents might lie or exaggerate Structured questions and fixed-response alternatives may result in loss of validity Wording questions is not easy Disadvantages of Survey Methods Rely on large samples of voluntary participants, and many people are unwilling to take surveys Preventing error: Keep it short Industry example: Survey limits The world’s most expensive burger cost more than 2000 Euros in a Dutch restaurant It’s off menu now, but it was covered with a gold leaf The burger creates buzz, but that’s something people hardly admit in surveys Survey example Luxury cosmetics brands wants to find out whether there is a market for a skin care line for men The brand expects results such as: X% of men between 18-30 are interested in luxury skin care On average, men use skin care products twice a week On average, men are 8.7 satisfied (on a scale from 1-10) with their current products) X% of students are interested in the brand, versus X% of young professionals Survey error: Example Launches survey on social media Problem 1: only men who already like the brand click on the survey to participate (self-selection bias) Problem 2: participants don’t want to admit that they can’t afford the luxury brand (social desirability bias) Survey error Respondent Errors: ▪ Response Error: Non-response error (refusals) ▪ Acquiescence bias Self-selection bias (people who volunteer) ▪ Extremity bias ▪ Interviewer bias ▪ Auspices bias (e.g., survey from Greenpeace or Shell) ▪ Social desirability bias Measurement error Questionnaire errors: Respondent fatigue E.g., long confusing survey; respondents get frustrated Questionnaire bias How much do you spend on shampoo per month in pound? Write down the number and whether you think that’s little/average/a lot < 50 pound per month 50-100 pound per month > 100 pound per month Survey error: Example The luxury cosmetics brand has designed a survey and launched it on social media, and hundreds of men have participated Problem 1: The survey is very long, and many quit halfway through, without having completed demographic questions (respondent fatigue) Problem 2: Some questions push respondents in a certain direction (questionnaire bias) Sample Design Errors Population specification error E.g., target all from 18-60 when it should be working population Sample frame error E.g., target all customers from 9-17:00 when the store is open till 20:00 Sample selection error E.g., leave out people who are busy looking E.g., only ask existing customers Survey error: Example The luxury skin care brand found that 10% of the men in the survey would be interested in trying the product Problem 1: Men who are not active on social media were not surveyed (sample selection error) Problem 2: The survey said that men from 18-30 were invited to participate, but men over 30 might also belong to the target segment (population specification error) Preventing error: Keep your survey invitation neutral to make sure you don’t deter/attract certain people Make sure you sample people from your target segment The better your segmentation, the more targeted your survey can be Preventing social desirability Preventing social desirability Example: do you care about buying eco-friendly products? Yes/No Better: To what extent are the following qualities important to you? Price Quality Eco-friendliness Taste … Preventing social desirability Example: do you like our product? Yes/No Better: what did you think about the following aspects of our product? Price Quality Eco-friendliness Looks … Preventing social desirability Think about your Lizzie questions What issues might be affected by social desirability? Create a question where you think almost everyone will say yes to Then make it more neutral What did you change? Survey: Takeaway A large sample alone doesn’t prevent bias (e.g., self-selection) Make sure to complement with secondary research and qualitative data For every survey question, ask What is the point of this question? yourself: Do I need this information? Questionnaire design Bad example Questionnaire design Bad example Questionnaire design Good example End of Session 3 Prepare for next week Take the Session 3 Quiz on Moodle! Get started on your research