OperationsFramework / modelModelAccessible

Quantitative surveys

When and how should quantitative surveys be applied?

AccessibleTacticalIndividual3 min read

Contents

A quantitative survey seeks to ‘quantify’ something from a numerical or statistical point of view.

A quantitative survey collects standardised responses that can be counted and analysed statistically. It can estimate opinions or behaviours in a defined population, compare groups, monitor change or test relationships—provided the sample, questions and analysis support the intended inference.

When to use it

Use a quantitative survey when the concepts and response categories can be defined in advance and the decision requires numerical evidence.

Typical uses include:

Estimating whether demand exists for a product or service.

Measuring awareness of an offer or brand.

Estimating purchase interest and potential market size.

Describing the characteristics of high-value or frequent customers.

Measuring buying habits and channel use.

Tracking changes in needs, satisfaction or attitudes over time.

Surveys are flexible, but that flexibility does not guarantee validity. A biased sampling frame, leading question, low response rate or changed wording can produce a precise-looking answer to the wrong question. Incentives may improve participation, but their effect on who responds should be considered.

Origins

Quantitative surveys developed from censuses, administrative counting, probability theory, social statistics and sampling research. No single person invented the method. Modern practice combines questionnaire design with sampling, measurement and statistical inference so that claims about a population can be tied to a transparent design.

What it is

A quantitative survey usually presents a consistent set of questions and predefined response options. Standardisation makes responses easier to compare and aggregate, but respondents may not see their experience reflected in the available choices. Carefully placed open fields or prior qualitative research can reduce that problem.

Quantitative data may also come directly from operations rather than a survey. Use a questionnaire when the required construct—such as awareness, intention or satisfaction—must be reported by people and cannot be observed reliably from behaviour alone.

Why it matters

A well-designed sample can provide an efficient estimate of how a defined population thinks or behaves. It can measure incidence, compare subgroups and quantify uncertainty around estimates.

Repeated surveys can track change, but comparability requires stable wording, response options, mode and sampling—or an explicit method for adjusting changes. Recontacting the same people creates a panel, which supports individual-level change analysis but introduces attrition and conditioning risks.

How to use it

Begin with the decision and target population. Translate the research question into measurable constructs, then choose a sampling frame and recruitment method. Decide in advance how estimates will be weighted, which subgroup comparisons matter and how much uncertainty is acceptable.

When writing the instrument:

Explain the purpose, sponsor, expected effort, data use and privacy protections in plain language.

Use concise wording that the intended population understands.

Ask one thing at a time; “How attractive and easy to use is product X?” combines two judgements.

Provide balanced, exhaustive and mutually understandable options, including “not applicable” where needed.

Keep only questions connected to the research objective.

Pilot the survey for comprehension, accessibility, ordering effects and technical problems.

Online tools can reduce cost and accelerate collection, but phone, post and face-to-face administration may better reach some groups. Mode affects coverage and responses, so select it based on the target population and preserve comparability across waves.

Possible data sources

The data come from responses to a purpose-built quantitative survey sent to a defined group such as customers, employees, suppliers or investors.

Collection may be online, mobile, postal, telephone or interviewer-administered. Administrative or behavioural data can be linked where lawful and appropriate, but linkage does not correct a poor questionnaire or unrepresentative sample.

How difficult or costly is it to collect?

Cost varies with population accessibility, sample design, mode, language, accessibility, incentives and the precision required. Interviewer and postal modes usually cost more and may require manual data entry or additional follow-up.

Online collection is often fast and economical, but a cheap convenience sample can be more misleading than a smaller, deliberately recruited one. Budget for questionnaire testing, nonresponse analysis, cleaning, weighting and quality checks as well as distribution.

Practical example

A travel agent wants to time and target its marketing. It asks customers, “How many times have you flown overseas in the last 6 months?” and offers: Never; 1–2; 2–5; 6–10; More than 10.

The standardised categories support aggregation, but they overlap at the boundaries. The agent should revise them so each possible answer belongs to only one range, define whether trips or flight legs count, recruit a sample relevant to the target market and report uncertainty before extrapolating.

Top practical tip

Define the target population, sampling frame and intended inference before choosing the channel. Make participation accessible to that population and report coverage and nonresponse, because sample size alone does not make findings representative.

Top pitfall

Do not optimise only for a short completion time. A concise survey can still fail through leading wording, overlapping options, poor coverage or selective nonresponse. Pilot the full journey and validate that each item measures the intended construct.