Quantitative Market Research: Methods & Examples
Quantitative market research measures needs, preferences and behavior with numbers. Learn to choose methods, questions and denominators with a worked example.
Quantitative market research uses numerical data to measure needs, preferences or behavior in a defined audience. Surveys, experiments and recorded actions can answer different questions. Choose the measure for the decision, keep the question and answer counts beside each result and check whether the sample supports the conclusion you want to draw. If your team will run the questionnaire itself, our AI market research tools comparison covers survey workflows, participant sources and research support.
If a survey combines understanding and usefulness in one score, the team won't know whether to change the wording or investigate the service. The measure you choose determines which business question the numbers can answer.
Example: understanding a delivery deadline
The following six records are hypothetical. They show why understanding the deadline and finding the service useful need separate questions and calculations.
The proposition for recent grocery-delivery users reads: “Select a two-hour delivery window at checkout. Changes are allowed until the evening before delivery.”
Six records reveal more than one average
Record | Responses |
|---|---|
R1 | Deadline understood: Yes (1) Usefulness: 5 |
R2 | Deadline understood: Yes (1) Usefulness: 4 |
R3 | Deadline understood: No (0) Usefulness: 4 |
R4 | Deadline understood: Yes (1) Usefulness: 3 |
R5 | Deadline understood: No (0) Usefulness: 2 |
R6 | Deadline understood: Yes (1) Usefulness: Cannot judge |
Understanding: 4/6. Ratings 4–5: 3/5 substantive ratings, with one Cannot judge. R3 likes the offer but misread the deadline; R4 understood it but found it less useful.
The target audience is recent grocery-delivery users. The team must decide whether to revise the deadline wording. Ask people to explain the deadline before showing answer choices. Rate usefulness separately. Neither measure checks whether the service can deliver on time.
For a real study, define recent use, location, recruitment and exclusions. Say whether the result describes eligible respondents, a target human population or a simulated population. Those are different claims.
Define who has a chance to respond or act: a buyer shown an offer, an account due to renew or a customer able to order. Then define the answer or behavior to measure. Agree what a high or low result would change.
A brand study may measure awareness after a defined prompt. A concept study may check unaided understanding. An experiment may record actions under assigned conditions. All can produce numbers without answering the same question.
The questionnaire and stored values define one measure
- Have you received a grocery delivery in the past 30 days? Yes; No; Prefer not to answer. Continue eligible respondents only.
- Show the proposition exactly as written. When can you last change the delivery window? Open response.
- How useful would this arrangement be for your household? 1 Not at all; 2 Slightly; 3 Moderately; 4 Very; 5 Extremely; Cannot judge.
- What information is missing from this description? Open response.
Variable | Values and rule |
|---|---|
record_id | Type and allowed values: Unique text Rule: One eligible record per participant |
recruitment_source | Type and allowed values: Text Rule: Preserve actual source, not an inferred demographic |
deadline_text | Type and allowed values: Original text Rule: Store before coding |
comprehension | Type and allowed values: 1 correct; 0 incorrect; missing unanswerable Rule: Correct requires the evening before delivery |
usefulness | Type and allowed values: Integer 1–5; Cannot judge; missing Rule: Do not encode “Cannot judge” as 3 |
information_needed | Type and allowed values: Text Rule: Code separately from comprehension |
The data dictionary keeps software settings from changing the meaning. Coding Cannot judge as 3 turns uncertainty into moderate usefulness. Coding missing understanding as 0 turns an unanswerable record into a wrong answer. Keep the distinctions set in the questionnaire.
Plan buyer-group comparisons before collecting data. Define purchase history for new and repeat buyers, or the boundary for a service zone. Collect context that helps answer a planned question.
Different questions need different denominators
Use the six records above. Code correct deadline understanding as 1 and incorrect understanding as 0, while keeping unanswerable records separate. Read the usefulness scale as 1 Not at all through 5 Extremely, with Cannot judge outside that scale.
Correct understanding is 4/6. Ratings 4–5 are 3/5 of substantive usefulness answers. Report “Cannot judge” separately, outside that denominator. Show the rating shares. A mean could hide the misunderstanding.
If the example rule required all six records to understand the deadline, this wording would fail. That is a teaching rule, not an industry standard. Read the wrong answers and revise the deadline before testing again.
Read an incorrect answer before revising
Record R3 says: “I can change it during the two-hour slot.” Under the stated evening-before rule, code comprehension as 0 even though usefulness is 4. R1 says: “No later than the evening before it arrives,” which receives 1. The favorable rating doesn't repair the misunderstanding. Test this deadline-first candidate: “Change your delivery window until the evening before delivery. At checkout, choose a two-hour delivery window.” Keep the offer and exposure conditions unchanged.
R3 rates usefulness 4 but misunderstands the deadline. R4 understands it but rates usefulness 3. R3 suggests a wording problem; R4 raises a question about whether the service fits the household. Read both before requesting another copy edit.
Use the answer shares and original explanations needed for the decision. A count may reveal the information gap. Add a more complex statistic when it answers a question the simpler result cannot.
An audience comparison needs the same complete offer
Compare groups when the offer, buying criteria and answer counts are comparable. If new users see different delivery terms from repeat users, both audience and offer change. Describe the result as that combined comparison.
Understanding uses six records. Usefulness uses five substantive ratings, with one Cannot judge reported separately. A percentage of all six answers a different question from a percentage of those five ratings. Show the denominator beside each result.
Check who the human sample can reach and who responds. A large overall total may still leave too few people in a critical group. For simulated responses, record the audience data and task validation. Generated profile counts do not create a human-sample margin of error.
Use the report to choose the next investigation
Ask for each question’s answer shares, valid counts and the records explaining mistakes. Keep routing, missing values and recruitment coverage in the analysis. Marketing needs the question and audience to know what a number means.
Sampling uncertainty depends on the human design and estimator. For simulated comparisons, explain the audience data and validation for the task; generating more profiles does not create more independent human observations. The AAPOR disclosure standards provide a reference for reporting the design.
Use our analysis walkthrough for the data and our market research survey guide for fieldwork.
Sapien's study workflow takes a simulated audience built from data through research to an interactive report and interpretation. Audience comparisons, recommendations and follow-up questions can help explain the simulated responses. For this delivery offer, use that work to distinguish a misunderstood deadline from service terms that don't fit the household.
Book a demo to discuss a simulated comparison of the deadline-first wording and the original, with understanding and usefulness reported separately.
Test the deadline-first wording with unchanged service terms and ask people to explain it unaided. If understanding improves but usefulness stays weak, investigate household needs or compare another complete offer. To learn whether ordering changes, observe it in a behavioral test.
