4 parts, about 20 minutes. Your work saves itself in this browser. See the guide again
Market research is for deciding with data instead of hunches. People who do it prepare questions that don't influence the answers, choose the right people to ask, read the results carefully and respect participants' privacy.
A good question is neutral (it doesn't suggest the answer), asks one thing and collects only the data really needed: under the GDPR personal data must be kept to a minimum.
For exampleAn ice cream shop asks "How often do you eat ice cream in winter?", not "Don't you think ice cream in winter is fantastic too?".
It's the group of people you ask. It must resemble the customers you want to reach and be large enough. Asking only friends or only fans creates a bias: the results come out more positive than reality.
For exampleFor a new ice cream, 300 people who eat ice cream at least once a month, half customers and half not.
Answers are turned into percentages (answers ÷ total × 100). Careful: "I'd buy it" is an intention, not a purchase. Real sales are usually lower than what people say.
For example90 people out of 300 say they'd try the new flavour: 90 ÷ 300 = 30%. It doesn't mean 30% will really buy it.
Before investing a lot you run a small test and listen to the objections. The data collected is used only for the stated purpose, and results aren't touched up to look better.
For exampleThe ice cream shop tries the new flavour for two weeks in a single shop before putting it everywhere.
For each question check three things: is it neutral or does it push towards an answer? Does it ask one thing or two at once? Is the data it asks for really needed for the research? If any of the three checks fails, the question needs redoing.
For example"Do you like the taste and price of our ice cream?" asks two things at once: if someone says no, you don't know which they mean.
Here's the draft questionnaire. Which questions are fine and which need redoing?
Why it matters. Questions change the answers. A good question is neutral, asks one thing and collects only the data really needed: under the GDPR personal data must be kept to a minimum.
The right sample is made of people similar to the customers you want to reach, numerous enough, who agreed to take part. Too small, too friendly or too different from the audience gives wrong results.
For exampleAsking 10 friends whether the ice cream is good almost always gets a yes: it's a bias.
Choose the sample, meaning the people who'll answer the questionnaire.
Why it matters. The sample must resemble potential customers, be large enough and not be made up only of people who are already fond of us. Otherwise the results look good but don't tell the truth (bias).
For percentages divide by the total and multiply by 100. For the price, add up the ranges from the top: a person who'd pay €2.50 is also fine with €2.00. Look for the highest price that works for at least 75 people in 100.
For exampleIce cream, maximum price: €4 10%, €5 40%, €6 35%, €7 15%. €5 works for 40 + 35 + 15 = 90%; €6 only for 50%. The price you're looking for is €5.
Using the results on the card, answer Marta.
Why it matters. Questionnaire numbers are read with caution: they're intentions, not sales. They help you understand whether it's worth trying, and at what price, not promise a result.
There can be more than one right answer. Keep the careful actions (trying on a small scale), the ones that respond to the objections raised and the ones that respect privacy. Discard inflated claims, using data for other purposes and touched-up results.
For exampleIf many people say "too sweet", the ice cream shop tries a less sugary version instead of ignoring the data.
What do you advise Marta? Choose all the right things, and only those.
Why it matters. Research is for deciding better, not for confirming what we wanted to hear. The data collected is used only for the stated purpose (GDPR, purpose limitation), and results don't become exaggerated slogans.
Marta has written to you: Shall we make bottled iced tea?. You'll solve it in 4 parts, in about 20 minutes.
Here you'll find the request and a card with the numbers. Use them to answer: that's where the right answers are.
The parts one below the other. You don't need to write anything: choose in the tables, among the proposals and in the menus. Watch out for traps!
Checks turn green when something is right, and I tell you what to fix. Each hint you open costs 5 points.
Your work stays saved in this browser. At the end click Submit and compare to see your points and the solution.