Qualitative research explores why: it captures meaning, motivation, and experience in people’s own words, usually through small numbers of participants studied in depth. Quantitative research measures what and how much: it captures counts, rates, scores, and durations from larger samples, so patterns can be compared and tracked over time (Rohrer, 2022). Neither disproves the other; they answer different questions.
Common qualitative methods include:
- Interviews: Structured or semi-structured conversations that surface a person’s context, motivations and language.
- Observation and contextual inquiry: Watching people use a service in their own setting, uncovering workarounds they would not think to mention.
- Diary studies: Participants record experiences over days or weeks, revealing patterns a single session would miss (Flaherty, 2024).
- Usability testing: Watching someone attempt real tasks to see where a service breaks down.
- Focus groups: Group discussion that surfaces shared attitudes and disagreement between users.
- Feedback forms, comments and complaints: Unprompted accounts of what went wrong, useful for spotting recurring issues.
Common quantitative methods include:
- Surveys: Structured questions asked of larger samples, producing comparable scores.
- Net Promoter Score (NPS): A single likelihood-to-recommend question, tracked over time as an indicator of loyalty (Bain & Company, n.d.).
- Service performance data: Satisfaction scores and other tracked measures, such as the UK Customer Satisfaction Index (Institute of Customer Service, 2026).
- Demand data: How many people use a service, when, and through which channel.
- Contact centre volumes and waiting times: Call numbers, answer speed and abandonment, showing where pressure is building (Pearce, 2018).
Quantitative data shows where a problem exists and how big it is; qualitative data explains why it is happening and what might fix it. A spike in contact centre waiting times becomes actionable once interviews or complaints reveal the cause, such as a confusing online form pushing people to call instead.
Evidence only has value if it is collected, reviewed and acted on continuously, not gathered once and filed away. A feedback loop connects frontline data, such as complaints and NPS scores, back to the people who can act on it, so changes can then be tested and their effect measured. The Design Council’s Double Diamond frames this as ongoing movement between discovering evidence and defining what it means, not a one-off exercise (Design Council, n.d.).
Service designers triangulate across methods to identify user needs, pain points, opportunity areas and improvement priorities, checking that a pattern holds across more than one source before acting on it. A single comment is an anecdote; a pattern confirmed across interviews, complaints and demand data is evidence worth designing around.
Research should involve more than a research specialist. Delivery staff, frontline colleagues who deal with users daily and often notice patterns first, and people from across the service should all help plan research and review findings, not only people with a research job title.
Recruitment should be planned deliberately, since how people are approached shapes who takes part. Recruitment briefs should set out diversity requirements, and any incentives, channels or agency screeners should be checked so they do not narrow the pool without anyone noticing (Government Digital Service, n.d.).
Samples that reflect the full user population, not just age, gender and region but also disability, digital confidence and ethnicity, surface needs that would otherwise stay invisible (Market Research Society, n.d.). Common sampling biases include self-selection bias, where only motivated or opinionated people respond (Pew Research Center, 2018), and convenience bias, where only users who are easy to reach or already engaged are heard. Relying only on these voices risks designing a service that works well for people who already cope, while excluding those the service most needs to reach.
Action Point
Pick one piece of qualitative feedback, such as a complaint, and one quantitative signal, such as a service metric, that relate to the same part of your service. Check whether they tell the same story. If not, plan what further evidence you need, and who currently isn’t being heard, before deciding what to investigate further.