Zurück

Descriptive Research: Definition, Types and Examples

Matthieu Saussaye

Descriptive research measures what is happening: who your customers are, what they do, how often and how many. It describes a population, market or behavior as accurately as possible at a point in time or over time, usually with structured surveys, observation or existing records. It answers what, who, where, when and how much. It does not answer why.

The short version:

  • It is about accuracy, not discovery. You already know what to measure. The job is to measure it well, on the right people, with enough of them.

  • It is usually quantitative. Closed-ended surveys, usage analytics, sales records and structured observation produce counts, percentages and averages.

  • It needs a representative sample. A descriptive figure is only as good as the match between who answered and who you want to describe.

  • It comes in three main designs: cross-sectional (one snapshot), longitudinal (the same measures repeated over time) and case studies (a deep description of one unit).

  • It does not prove causes. Finding that loyal customers use feature X more does not show that feature X creates loyalty. That takes causal research.

  • It is the most common type of business research. Market sizing, customer profiles, usage and attitude studies, brand tracking and satisfaction scores are all descriptive.

Descriptive research is one of the purpose-based types of research. This guide covers its designs, how to plan one, the questions to use and the mistakes that make descriptive numbers misleading.

Types of descriptive research design

The design you pick decides whether you can talk about change over time, about one case in depth, or only about a single moment.

Design

How it works

Answers

Business example

Watch out for

Cross-sectional survey

One structured survey of a sample, at one point in time

What does the population look like now?

Profile of current customers by size, sector and usage

A single snapshot can catch a seasonal high or low

Longitudinal (tracking)

Same questions asked repeatedly, of a new sample (trend) or the same people (panel)

How is it changing?

Quarterly brand awareness tracker

Changing a question's wording breaks the trend line

Observation

Recording behavior as it happens, in person or through analytics

What do people actually do?

Store visit counts, click paths, time on task

Shows behavior without motive

Case study

Detailed description of one organization, account or market

What does this case look like in full?

How a key account rolled out a product across departments

Not generalizable to other cases

Existing records

Describing data already collected: CRM, sales, public statistics

What do the records show?

Revenue by region and segment

Records were built for another purpose and may miss what you need

Most programs combine a cross-sectional survey for the first read and a longitudinal tracker to watch the numbers that matter.

What descriptive research is used for

If a question starts with how many, what share, who, how often or what is typical, it is a descriptive question. Typical uses:

  • Market sizing. How many businesses or households have the problem you solve, and how much do they spend on it today?

  • Customer profiling. Who buys, in what roles, company sizes and regions. The base for segmentation and personas.

  • Usage and attitude (U&A) studies. Which products people use, how often, for what, and how they rate them.

  • Satisfaction and loyalty measurement. CSAT, NPS and effort scores are descriptive metrics. They tell you where you stand, not why.

  • Brand tracking. Awareness, consideration and preference measured wave after wave.

  • Benchmarking. How your numbers compare with an industry figure or a competitor's published data.

How to conduct descriptive research

  1. Write the exact statistics you need. 'Share of SMB customers who use the mobile app weekly, by country' is a descriptive objective. 'Understand mobile usage' is not. If you cannot list the numbers, you probably need exploratory research first.

  2. Define the population precisely. 'Customers' could mean all accounts, active accounts or paying users. The definition decides who is eligible and what your figure means.

  3. Pick the design. One snapshot or a repeated measure? A survey, observation or records you already hold?

  4. Size the sample for precision. Decide how precise each figure must be, including for subgroups you plan to report. A figure for 'all customers' may need 400 respondents; the same figure for each of four regions needs roughly 400 per region. Use a sample size calculator rather than guessing.

  5. Write closed questions with complete answer lists. Every option a respondent might need should be there, plus 'none of these' and, where honest, 'don't know'.

  6. Field and check the sample. Compare the achieved sample with known population figures (region, size, age). Weight if needed, and report that you did.

  7. Analyze and report with the right caveats. Frequencies, percentages, means, medians and cross-tabs. State the base size behind every figure.

How precise is a descriptive figure?

For a simple random sample, the margin of error at 95% confidence for a percentage near 50% is roughly 0.98 divided by the square root of the sample size. The table applies that formula. It is arithmetic, not a benchmark, and it assumes random selection, which online panels and customer lists rarely achieve.

Respondents

Approximate margin of error (95%, p = 50%)

Good enough for

100

plus or minus 9.8 points

Rough direction only

250

plus or minus 6.2 points

Large differences between groups

400

plus or minus 4.9 points

Most single-population estimates

1,000

plus or minus 3.1 points

Published figures, tracking small changes

2,500

plus or minus 2.0 points

Detailed subgroup reporting

The lesson for planning: precision improves with the square root of the sample, so halving the margin of error takes four times the respondents.

Descriptive survey questions you can use

Descriptive questions are closed, specific and time-bound. Grouped by what they measure:

Who (profile)

  • Which of the following best describes your role? (list of roles, plus Other)

  • How many employees work at your organization? (1-9, 10-49, 50-249, 250-999, 1,000+)

  • In which country is your organization headquartered?

  • Which age group are you in? (18-24, 25-34, 35-44, 45-54, 55-64, 65+)

What and how often (behavior)

  • In the past 30 days, how many times did you use [product]? (0, 1-2, 3-5, 6-10, more than 10)

  • Which of these tools do you currently use for [task]? Select all that apply.

  • When did you last buy [category]? (this week, this month, 1-3 months ago, longer, never)

  • Approximately how much does your team spend per year on [category]? (ranges)

How much (attitude and satisfaction)

  • Overall, how satisfied are you with [product]? (5-point scale from very dissatisfied to very satisfied)

  • How likely are you to recommend [product] to a colleague? (0-10)

  • How easy or difficult was it to [task]? (very easy to very difficult)

  • Which of these brands have you heard of? Select all that apply.

Where and when (context)

  • Where do you usually buy [category]? (online, in store, through a distributor, other)

  • On which device do you mostly use [product]?

  • At what time of day do you typically [task]?

Keep wording identical from wave to wave if you plan to track. One changed word can make a trend line meaningless.

Example: a usage and attitude study

A meal-kit company wants to know who its subscribers are and how they use the service. Illustrative design:

Element

Choice

Objective

Share of subscribers by household size, order frequency, cooking time and satisfaction, by country

Population

Active subscribers with at least one delivery in the past 90 days

Design

Cross-sectional online survey, repeated every six months

Sample

400 per country, random draw from the active list, weighted by plan type

Output

Profile tables, cross-tabs by plan type, satisfaction scores with base sizes

The study will show, for example, that larger households order more often and rate value lower. It will not show why value scores are lower. That is a job for explanatory research: driver analysis and follow-up questions that unpack the pattern.

Descriptive research vs other research types

Exploratory

Descriptive

Explanatory

Causal

Question

What is going on?

How many, who, how often?

Why does it happen?

Does X cause Y?

Starting point

No hypothesis

Known variables to measure

A measured pattern

A specific hypothesis

Typical method

Interviews, open questions

Structured survey, analytics

Driver analysis, mixed methods

Experiment, A/B test

Sample

Small, varied

Large, representative

Medium to large

Randomized groups

Common mistakes in descriptive research

  • Reporting causes from correlations. 'Customers who attend webinars renew more' is a description. Write it as one.

  • Hiding base sizes. A 60% figure from 35 people in a subgroup is not the same as 60% from 1,000.

  • Describing the wrong population. Surveying only people who opened your newsletter describes engaged readers, not your market.

  • Incomplete answer lists. If 'Other' gets 25%, your list was built without enough exploration.

  • Changing questions in a tracker. New wording means a new series, not a continuation.

  • Ignoring non-response. If only satisfied customers answer, the satisfaction score describes them, not the customer base.

Describe the what, and capture the why in the same survey

Descriptive surveys tell you that value ratings are lower among large households. Most teams then commission a second study to find out why.

SmartInterview lets you run the structured descriptive questions and add AI follow-up probing on the answers that need explaining, by voice or text and in each respondent's language. Open answers are coded automatically into themes, so the profile tables and the reasons behind them come from the same respondents, in the same fieldwork.

Start free and build your first descriptive survey.

Frequently Asked Questions

What is descriptive research in simple terms?

Descriptive research measures and describes something as it is: who your customers are, what they do, how often and how many. It uses structured methods such as surveys and observation to produce counts, percentages and averages. It describes the situation but does not explain why it happens.

What are the three main types of descriptive research?

The three most cited are surveys, observation and case studies. Surveys collect structured answers from a sample, observation records behavior as it happens, and case studies describe one organization, person or event in depth. Surveys can be cross-sectional or longitudinal.

Is descriptive research qualitative or quantitative?

It is usually quantitative, because it aims to measure frequencies and averages. Qualitative methods such as case studies or observation notes can also be descriptive when their goal is to document what something looks like rather than to explain it.

What is an example of descriptive research?

A customer survey that reports what share of users are in each industry, how often they log in and how satisfied they are is descriptive research. So are a national census, a quarterly brand awareness tracker and a market sizing study.

Can descriptive research show cause and effect?

No. Descriptive research can show that two things occur together, for example that frequent users are more satisfied, but not that one causes the other. Establishing cause requires a design that controls for other explanations, usually an experiment.

How large should a descriptive research sample be?

It depends on the precision you need and the subgroups you will report. For a simple random sample, around 400 respondents gives a margin of error of about plus or minus 5 points on a percentage, and around 1,000 gives about plus or minus 3 points. Each subgroup you report needs its own adequate base.

Table of Contents

Verwandte Artikel

Kostenlos anmelden

Kostenlos anmelden

Kostenlos anmelden