24761 Chap.9 Primary Data Methods for Validating Insights
Primary Data Methods for Validating Insights
Why primary data enters the story
Week 6 turns from evidence that already exists to evidence a firm gathers first-hand for one question. Primary data is purpose-built and actionable, and in this subject it complements secondary findings rather than replacing them.
You will not collect any; the second report asks you to recommend the approach that would test your earlier conclusions.
The method families
Customer interviews and focus groups uncover reasons and reactions, ethnography watches what people actually do, surveys measure how widespread a view is, and experiments, including A/B and blind tests, check whether a change causes an effect.
Choosing well
The chapter compares the methods side by side, routes exploratory, descriptive and causal questions to the right family, reads a worked A/B test with computed conversion rates, and sets out what a recommendation must include: the method, the variables, why they matter and how they reduce uncertainty.
A short note on blind testing covers when hiding the brand helps and when it can mislead a product team.
What this chapter covers
- 01
What makes data primary, and why it complements secondary evidence
- 02
Interviews, focus groups and ethnography for exploring the why
- 03
Survey question types for measuring how many
- 04
Experiments: independent and dependent variables, control groups, random assignment
- 05
A/B testing and blind testing in marketing
- 06
Matching research objectives to methods
- 07
Writing the primary data recommendation for the second report
Worked example · free
Match research questions to methods for a credit union
- 1Question (a) asks why, so it is exploratory: recommend one-on-one interviews with members who abandoned a goal, since personal money habits are easier to discuss privately.
- 1Question (b) asks how many, so it is descriptive: recommend a survey of a large member sample with a closed question on intended use, compared across segments.
- 1Question (c) asks whether one change causes another, so it is causal: recommend an A/B test that randomly assigns members to the current or earlier reminder and compares late-repayment rates.
Key terms
- Focus Group
- A moderated discussion with a small set of target customers, typically six to ten, used to explore reactions to a product, message or concept.
- Ethnography
- Observing consumers in their everyday settings, physically or online, to learn what they actually do rather than what they report.
- Likert Scale
- Survey response format asking how strongly a person agrees or disagrees with a statement, used to measure attitudes and opinions.
- Control Group
- Participants in an experiment who do not receive the change being tested, giving a baseline for comparison with the treated group.
- Blind Test
- Evaluation in which brand names, packaging and other identifying cues are hidden so a product is judged on its own merits.
Primary Data Methods for Validating Insights FAQ
Do I have to collect primary data for this subject?
No. The Week 6 lecture page says students are not required to collect it. The second report instead asks you to recommend a primary data approach in no more than 200 words, explaining the method, the variables you would collect, why they matter and how they would validate the issue your secondary analysis found.
When should I choose interviews over a focus group?
Choose interviews when you need detailed personal journeys, sensitive stories or feedback a person may not share in front of others. Focus groups suit testing concepts, messages and shared experiences, where watching people react to one another adds insight. Both are qualitative and explore reasons rather than measuring how common a view is.
What makes a study a true experiment?
It needs an independent variable the researcher deliberately changes, a dependent variable that is measured, an experimental group compared with a control group, and random assignment of participants to those groups. Random assignment is what lets you credit a difference in outcomes to the change rather than to who happened to be in each group.
How is an A/B test different from a survey?
A survey asks people what they think, prefer or intend, and is good at describing opinions across many respondents. An A/B test shows two versions of one marketing element to randomly split groups and records what they actually do, such as clicking or buying, so it can support a claim about cause and effect.
What are the limits of blind testing?
Hiding the brand removes reputation and packaging from the judgement, which is its strength, but it also means the test does not reflect a real shopping situation. Preferences can change once people know the brand, and the sessions take time to organise, so treat the result as evidence about the product itself.
Assessment move
Take the issue you identified from secondary data and write down the single assumption it depends on most. Decide whether testing that assumption needs reasons, scale or cause, then use the decision tree to pick a method family. Draft four to six variables you would collect and, beside each, one line on why it matters and what result would change your view.
Practise reading an A/B result by computing both conversion rates and the difference in percentage points. Last, cut your recommendation to 200 words, keeping the method, the variables and the link back to the evidence, and read it aloud to check that every sentence earns its place.
Working through Primary Data Methods for Validating Insights in 24761? Sia is AskSia’s AI Marketing tutor — ask any 24761 Primary Data Methods for Validating Insights question and get a clear, step-by-step explanation grounded in how 24761 is taught and assessed. Read this chapter free, then take your hardest questions to Sia.