Sample size

Sample size is the number of observations, respondents, interviews, transactions or units included in a research study. In market research, it determines how precisely findings can describe a target population and how much confidence decision-makers can place in quantitative results.

What is sample size?

Sample size is the count of units selected for analysis from a broader population. In market research, these units are usually consumers, business buyers, households, users, stores, customer records or digital interactions. The concept is central to survey research, concept testing, segmentation, customer experience studies, brand tracking and many other forms of quantitative analysis.

The role of sample size is to make inference possible. Researchers rarely study an entire market because full coverage is costly, slow or technically impossible. Instead, they select a sample and use its results to estimate characteriztics of the population, such as awareness, purchase intent, satisfaction, market share perception or preference for a product feature. A well-defined sample size helps control sampling error, which is the random difference between the result obtained from a sample and the result that would be obtained from the full population under the same measurement conditions.

Sample size in market research is not a measure of research quality by itself. A large sample drawn from the wrong population, recruited through a biased channel or measured with a weak questionnaire may still produce misleading results. Conversely, a smaller but properly designed sample can be more useful when it reflects the target group, uses appropriate quotas or stratification and answers a clearly defined business question. For this reason, sample size should always be assessed together with sample design, recruitment method, response quality, weighting and analytical objectives.

Application of sample size in practice

Sample size is used whenever market researchers need to determine how many cases are required to support a decision. It is specified during research design, before fieldwork begins, because it affects budget, timing, feasibility, data quality procedures and the level of detail available for analysis.

In practice, sample size in market research is applied in several common situations:

• Consumer surveys: to estimate attitudes, needs, usage patterns, purchase drivers or brand perceptions within a defined consumer population.
• B2B research: to study decision-makers, procurement teams, channel partners or professional users, often in smaller and harder-to-reach populations.
• Concept and product testing: to compare reactions to alternative propositions, packaging designs, pricing options or communication messages.
• Customer experience measurement: to assess satisfaction, loyalty, effort or perceived service quality across customer segments and touchpoints.
• Brand tracking: to monitor changes in awareness, consideration, usage or preference over time, where stable sample design is important for comparability.
• Segmentation studies: to identify meaningful groups in the market and profile them with sufficient analytical reliability.

Managers and analysts use sample size decisions to balance precision against cost. A larger sample usually reduces random sampling error, but the improvement becomes less efficient as the sample grows. The optimal sample size depends on the business risk of the decision, the expected variability of responses, the need to compare subgroups and the required level of statistical precision.

For example, a study designed to estimate a single overall preference may require a different sample size than a study designed to compare preferences across multiple age groups, regions, customer tiers or purchase categories. In B2B markets, feasibility may also be a constraint, because the total population of qualified respondents can be limited. In such cases, sample size planning should reflect both statistical requirements and realistic access to respondents.

Sample size and related methods

Sample size is part of a broader methodological ecosystem that includes sampling frame, sampling method, questionnaire design, fieldwork control and statistical analysis. It should not be confused with the population size or with the number of invitations sent to potential respondents. The sample size refers to the number of completed and usable observations included in the analysis.

Several related concepts are important when interpreting sample size in market research:

• Population: the full group about which the research aims to draw conclusions, such as all current customers, category buyers or enterprise software decision-makers.
• Sampling frame: the operational list, panel, database or source from which respondents can be selected.
• Sampling method: the procedure used to select units, such as probability sampling, quota sampling, stratified sampling or purposive recruitment.
• Margin of error: an expression of expected sampling variability for an estimate, under specified assumptions.
• Confidence level: the probability framework used to describe how often a sampling procedure would capture the true population value if repeated many times.
• Weighting: an adjustment applied when the achieved sample differs from known population structure or required analytical proportions.
• Effective sample size: the analytical sample size after considering weighting, clustering or other design features that influence precision.

Sample size differs across quantitative, qualitative and mixed-methods research. In quantitative research, sample size is primarily linked to estimation accuracy, statistical testing and subgroup analysis. In qualitative research, the term may refer to the number of interviews, focus groups or observed cases, but the logic is different: the goal is depth, variation and saturation rather than numerical representativeness. In mixed-methods studies, sample size planning often combines both logics. A quantitative survey may define the scale of market patterns, while qualitative interviews explain why those patterns occur.

Hume’s Institute treats sample size as one element of research design rather than as an isolated target. In applied projects, it is typically considered together with business objectives, available population data, planned comparisons, analytical models and the practical conditions of recruitment.

How to calculate sample size for a quantitative survey?

The question of how to calculate sample size for a quantitative survey starts with defining the decision that the data must support. A statistically elegant calculation is not useful if the population is unclear, if the required subgroups are not specified or if the planned analysis demands more cases than the design can deliver.

A practical approach to determining sample size in market research includes the following steps:

1. Define the population: specify who is in scope and who is excluded, for example recent category buyers, active customers or senior procurement decision-makers.
2. Clarify the key measure: identify whether the study estimates a proportion, mean score, incidence rate, preference share or difference between groups.
3. Set the desired precision: decide how much uncertainty is acceptable for the main estimate or comparison.
4. Estimate variability: use prior data, pilot results or conservative assumptions when the expected distribution of responses is unknown.
5. Account for subgroup analysis: ensure that important segments have enough completed cases for interpretation.
6. Adjust for design effects: consider the impact of weighting, clustering, stratification or panel structure on effective precision.
7. Plan for non-response and data cleaning: recruit enough respondents to allow for incomplete interviews, quality exclusions and eligibility screening.

For simple probability-based estimates of a proportion, researchers often use a standard formula in which the required sample size is driven by confidence level, expected proportion and acceptable error. In simplified notation, the initial estimate can be expressed as:

n = (Z² x p x (1 – p)) / e²

In this formula, n is the required sample size, Z is the critical value associated with the confidence level, p is the expected proportion and e is the acceptable margin of error. If the total population is limited and known, a finite population correction may be applied. If the design uses weighting or clustering, the calculated sample size may need to be increased to maintain effective precision.

This calculation should be interpreted as a planning tool, not as a mechanical rule. In commercial research, sample size in market research must also reflect questionnaire length, respondent availability, panel quality, incidence rate, fieldwork mode and the consequences of a wrong decision. The strongest sample size decisions are therefore both statistical and managerial: they define the level of evidence needed to act responsibly.

Limitations of sample size

Sample size improves the potential precision of a study, but it does not eliminate all sources of error. A survey with a large sample may still be affected by coverage error, non-response bias, poorly worded questions, order effects, interviewer effects or low respondent engagement. These problems are not solved by adding more cases if the underlying measurement process remains flawed.

Key limitations of sample size in market research include:

• Representativeness depends on selection, not only volume: the sample must reflect the target population in ways relevant to the research question.
• Subgroups reduce available precision: a total sample may look sufficient, while individual segments remain too small for reliable comparison.
• Weighting can reduce effective sample size: strong weights may correct structure but increase variance in estimates.
• Rare audiences require special planning: low-incidence groups may demand screening, targeted recruitment or alternative designs.
• Business decisions differ in risk: exploratory studies, tactical optimizations and high-investment decisions may require different evidence thresholds.

For this reason, sample size should be planned as part of the full research architecture. It is most valuable when connected to a clear population definition, transparent sampling method, disciplined fieldwork, valid measurement and an analysis plan aligned with the business question.