OperationsFramework / modelModelIntermediate

Qualitative surveys

How can qualitative surveys improve people, teams, or organisational effectiveness?

IntermediateOperationalTeam3 min read

Contents

Where a quantitative survey seeks to ‘quantify’ a topic through numbers and statistics, a qualitative survey looks to ‘qualify’ the response with more subjective opinion.

A qualitative survey uses open responses to explore how people interpret an experience, why they behave as they do and which meanings or emotions sit behind an opinion. Unlike a quantitative survey, its primary purpose is depth and discovery rather than numerical estimation.

When to use it

Use qualitative surveys when the research question is exploratory, the relevant answer categories are not yet known or respondents need room to explain an experience in their own terms.

They are particularly useful for investigating:

What customers think and feel about a product or service.

How customers choose among offers and competitors, and what motivates the decision.

How branding, design or packaging shapes interpretation and preference.

Which aspects of a message or advertisement resonate, confuse or repel.

How price is understood and how it enters the buying decision.

A survey is only one qualitative method. If context, interaction or follow-up probing is essential, interviews, observation or facilitated groups may be more suitable.

Origins

Qualitative surveys emerged from the broader traditions of qualitative social research, including open-ended interviewing, field research and interpretive analysis. They have no single inventor. Their contemporary form combines the reach and standardised administration of a questionnaire with the freedom of respondents to answer in their own words.

What it is

Suppose an organisation is considering a new identity. Closed ratings might show which version people prefer, but open questions can reveal what the existing brand signifies, which associations a proposed design evokes and why reactions differ. Qualitative survey data can therefore expose the reasoning behind an observed percentage or behaviour.

This depth is useful for defining a problem, generating hypotheses, discovering unanticipated needs and developing the language for later quantitative research. It comes with limits: responses are self-reports, the sample may not be representative, and open text is more demanding to interpret. Text Analytics or Sentiment Analysis can support large collections, but automated classification should not replace careful reading, contextual coding and review of ambiguous cases.

Why it matters

Qualitative surveys let people describe what matters to them rather than forcing every experience into categories chosen in advance.

That perspective can reveal unmet needs, misunderstood processes, alternative explanations and ideas for improvement. The result is analytical insight, not a statistically projectable share of the population unless a separate probability-based design supports that inference.

How to use it

Start with a focused research question and identify whose experience is needed. Write neutral, open prompts that invite concrete description, examples and reasoning. Avoid questions that presume the problem, combine several topics or ask respondents to speculate beyond their knowledge.

Guidance for Quantitative Surveys still applies to clarity, relevance, accessibility, consent and pilot testing. Online tools can simplify distribution and capture, while phone, postal and face-to-face methods may reach different populations or allow richer engagement. Choose the mode for the research population, not merely for administrative convenience.

Plan analysis before launch. Define a coding process, preserve contradictory and minority views, use more than one reviewer when interpretive risk is high, and link themes back to de-identified evidence. State the sample and method so readers do not mistake exploratory findings for population estimates.

Possible data sources

The primary data are the written or transcribed responses to a purpose-built qualitative survey.

The survey may be administered online, through a mobile tool, by post, over the phone or face to face. Relevant existing open comments may supplement the study if their original context, consent and sampling limitations are understood.

How difficult or costly is it to collect?

Cost depends on reach, mode, accessibility requirements, response length and analytical depth. Interviewer-led, telephone and postal approaches require more labour, while online collection reduces administration.

Collection is often cheaper than analysis. Researchers must clean responses, develop and test codes, interpret context and report variation without erasing nuance. Automated assistance can improve scale, but it introduces validation and governance work of its own.

Practical example

A company notices critical posts about service and needs to learn whether they reflect isolated incidents or a broader process failure. A short qualitative survey can ask affected customers to describe what happened, what they expected, where the experience broke down and what resolution would have restored trust.

A vague prompt such as “How was your last customer experience?” may produce “Good” or “Terrible,” neither of which explains the event. Asking respondents to describe their most recent purchase, the steps they took, any obstacle they encountered and its consequence elicits more actionable evidence. Existing complaints can guide areas to probe, but the wording should still allow customers to disconfirm management’s assumptions.

Top practical tip

Keep the study focused and pilot the prompts with people similar to the intended respondents. A small set of questions that elicits specific experiences is more valuable than a long form that produces shallow or abandoned answers.

Top pitfall

Do not present themes from a convenience sample as population prevalence. Report who responded, how they were recruited, how coding decisions were made and which important views did not fit the dominant pattern. A mixed survey can add closed measures, but it does not remove sampling or interpretation bias.