Descriptive statistics organise, summarise and present data so that patterns in a research sample can be understood before decisions are made. In market research, descriptive statistics turn survey responses, transaction records or behavioural data into clear indicators of customer attitudes, market structure and performance.
What are descriptive statistics?
Descriptive statistics are statistical methods used to describe the observed characteristics of a dataset. Their purpose is not to prove that a pattern exists in a wider population, but to show what the collected data look like: what answers occur most often, what a typical result is, how widely respondents differ and whether the distribution contains unusual values.
In market research, descriptive statistics are usually the first analytical layer after data cleaning and coding. They provide a factual view of results from quantitative surveys, customer satisfaction studies, segmentation projects, concept tests, employee surveys and digital analytics. They may also support mixed-methods research by structuring quantitative findings that are later interpreted with qualitative interviews or focus groups.
The core components of descriptive statistics include measures of central tendency and measures of dispersion. Measures of central tendency identify the value that best represents the centre of a distribution. The most commonly used measures are:
- Mean – the arithmetic average, calculated by adding all observed values and dividing their sum by the number of observations.
- Median – the middle value after all observations have been ordered from lowest to highest. At least half of the results are at or below it, and at least half are at or above it.
- Mode – the most frequently occurring value or response category.
Measures of dispersion show whether individual observations are concentrated around the centre or spread widely across the scale. Common examples include the range, which is the difference between the lowest and highest observed value, and the standard deviation, which indicates how far results typically vary from the mean. Percentiles, quartiles and the interquartile range are also useful when distributions are uneven or contain extreme observations.
For example, a customer satisfaction report may present the mean rating for service quality, the median rating, the share of respondents selecting each answer category, and the standard deviation. Together, these indicators show not only whether the overall assessment is positive, but also whether that assessment is consistent across customers or divided between distinct groups.
Application of descriptive statistics in practice
Descriptive statistics are used whenever a business needs a reliable, transparent description of collected data before interpreting causes, forecasting outcomes or making comparisons between groups. They are relevant to researchers, marketing teams, product managers, customer experience specialists, sales analysts and decision-makers who need to understand the current state of a market or audience.
In a brand tracking study, descriptive statistics can show awareness, consideration, usage and preference levels for each brand in the category. Frequency tables indicate how respondents answer closed questions, while measures of central tendency summarise scale-based evaluations such as perceived quality, innovativeness or value for money. Segment-level results can then reveal whether the overall average conceals substantial differences between, for example, current customers and non-users.
In B2B research, descriptive statistics may be used to profile decision-making units, describe purchasing cycles or summarise satisfaction with account management. For instance, median contract value can be more informative than mean contract value where a small number of very large accounts strongly affect the average. The question of when to use median instead of mean is therefore practical rather than purely technical: the median is preferable when data are skewed, include influential outliers or represent ordered categories with unequal distances between response options.
In consumer research, descriptive statistics support the reporting of:
- purchase frequency and category penetration,
- product concept ratings and stated purchase intent,
- Net Promoter Score components and customer satisfaction distributions,
- media consumption patterns and channel preferences,
- demographic, behavioural and attitudinal profiles of target groups.
Averages should always be interpreted with the shape of the distribution in mind. A mean score may suggest moderate satisfaction, while the response distribution may reveal that respondents are split between highly satisfied and highly dissatisfied groups. In such cases, reporting only one average can obscure an important business issue. Hume’s Institute uses descriptive statistics as a standard basis for interpreting quantitative findings and for identifying questions that require deeper segmentation or qualitative explanation.
Measures of central tendency and dispersion in a research report
Measures of central tendency are selected according to the level of measurement, the distribution of values and the decision context. The mean is most useful for numerical variables measured on an interval or ratio scale, such as declared monthly spending, delivery time or a composite satisfaction index. It uses every observed value, but is sensitive to extreme results.
The median is particularly valuable for variables such as income, expenditure, order value or time needed to complete a process. In these cases, a small number of unusually high values can shift the mean upward and create an impression that is not representative of the typical respondent or customer. The mode is useful for nominal variables, including the most frequently selected brand, preferred channel, primary purchase motivation or most common response to a single-choice question.
Measures of dispersion add necessary context to measures of central tendency. A low standard deviation around a mean satisfaction score suggests relatively consistent evaluations. A high standard deviation indicates greater diversity of opinion and may justify further analysis by customer segment, region, product type or customer tenure. Similarly, a wide range can flag potential outliers, though it should not be interpreted alone because it depends on only two observations.
In a clear research report, descriptive statistics should be matched to the variable being analysed and presented in a way that supports correct interpretation. Useful reporting formats include tables with bases and percentages, distributions of response categories, mean and median values, standard deviations, and charts that make variation visible. Each measure should answer a specific question rather than add technical detail without decision value.
Descriptive statistics and related methods
Descriptive statistics differ from inferential statistics. Descriptive statistics describe the data observed in a particular sample or database. Inferential statistics use sample data to estimate characteristics of a wider population and to assess whether observed differences or relationships may reflect more than sampling variation. Statistical tests, confidence intervals and significance testing belong to inferential analysis; regression models may be used for descriptive, predictive or inferential purposes.
The distinction matters in market research. If one customer segment reports a higher average satisfaction score than another, descriptive statistics show the size and direction of that observed difference. Inferential statistics help assess whether the difference is likely to persist beyond the surveyed sample, assuming the study design and sampling method permit such inference.
Descriptive statistics are also related to data visualisation, cross-tabulation and segmentation analysis. Visualisations such as bar charts, histograms and box plots make distributions easier to interpret. Cross-tabulations describe how responses vary across categories, for example by age group, region or customer status. Segmentation uses multiple variables to identify groups with distinct profiles, often beginning with descriptive statistics that reveal where relevant variation exists.
In qualitative research, descriptive statistics have a narrower role because the primary material is textual, visual or observational rather than numeric. They may nevertheless be used to report the composition of participants, the frequency of coded themes or the distribution of selected responses. Qualitative interpretation remains necessary to explain meaning, context and mechanisms that numerical summaries alone cannot establish.
Used correctly, descriptive statistics provide the factual foundation for credible market analysis. They show what has been observed, how typical the observed result is and how much variation exists within the data. Their value lies in disciplined description: selecting suitable measures, showing distributions where needed and avoiding conclusions that go beyond the evidence available.