Voice of Customer
Customer experience research
Plan customer experience research around a clear decision, choose suitable evidence, recruit relevant customers and report findings with their limits.
Customer experience research investigates what people need from a service, what happens as they use it, and whether they achieve their goal. Start with a decision the organisation must make. Choose evidence that can inform it, recruit people with relevant experience, and separate what customers report from what researchers observe or infer.
Start with a task and a decision
“Improve customer experience” is too broad to guide a study. A more useful question: what leads customers to contact support after submitting an address change? The answer could inform the form, its confirmation or the work after submission.
Define whose experience matters, where the task begins and what counts as a successful outcome. Include people who could not finish where possible. Research limited to customers who completed the task may miss the main barrier.
A short brief can name the customer group, research question, decision owner and evidence needed. Record assumptions separately. A complaint trend may suggest a question worth investigating; it does not establish the cause.
Match the method to the question
| What you need to learn | Useful evidence | What it cannot establish alone |
|---|---|---|
| How people understood a recent experience | Interviews about specific incidents | Every action they took, or how common the problem is |
| Whether people can complete a proposed task | Observation of relevant people attempting it | How often the issue affects all customers |
| Where recorded contacts or errors occur | Service records, interpreted in context | The customer’s reasoning or activity outside recorded channels |
| How people felt about an interaction | A question asked after that experience | Whether the service outcome occurred |
Methods can be combined around one question. If customers are uncertain whether an address change worked, interview them about the confirmation they expected, observe attempts to make the change, and check the resulting record where access is authorised. Each source answers a different part of the question.
Plan the research before sessions begin
Place the study in the design process before choosing activities. Discovery research can help uncover people’s motivations and needs, while validation research can test a prototype, proposed solution or existing product. The stage of work helps determine which methods can answer the question usefully.
A research plan can record the project background, objectives, methods, target audience, timeline and outputs. The Victorian Government recommends consulting internal stakeholders and reaching agreement before research begins, so there is shared understanding of the study’s scope and intended outcomes.
Prioritise objectives according to which are most likely to deliver business value; a project may raise more questions than its time and resources can cover. Use the chosen objectives as the basis for research questions and activities, rather than letting interesting but lower-priority questions expand the study.
Recruit for the experience
Define eligibility by the task and customer role. In a renewal study, people who renewed, abandoned the process or needed help may reveal different barriers. A support list alone will miss customers who did not contact support.
Plan for variation relevant to the decision, such as device, channel and access needs. Record recruitment routes and gaps with the findings. A small qualitative study can reveal barriers, but repeated observations within it do not establish a population rate.
Explain the session, any recording and the intended use of material before it begins. Follow the organisation’s applicable privacy processes for collection, access and retention.
Check the whole outcome
Research the customer’s goal across the relevant service journey. An order query may begin with a delivery message, pass through self-service and support, and end when the customer receives an accurate answer. A submitted form or closed ticket does not by itself prove resolution.
Record the task, actions, customer comments, any help given and the result. Keep interpretation separate: returning to a form after submission is observable; uncertainty about the confirmation is an explanation to check. Define success before observation when the result can be verified, and record partial or unknown outcomes rather than treating every submission as success.
Make a bounded decision
Group evidence by the question it answers. Keep examples, contrary cases, participant context and sample gaps alongside each finding. State what the team should change or investigate and what would show improvement.
If participants look for an updated address in their accounts and still see the old one, that supports checking the update and status process. It does not establish that every customer has the same experience. Estimating prevalence requires suitable records or quantitative research with a defined population.
After a change, check whether customers can achieve the intended outcome.
Analyse and document findings
Start analysis soon after sessions, while details are fresh. Compare observations with the research objectives and consider how far the evidence answers each key question.
One way to find patterns is affinity mapping: group raw data into common themes, then consider what each theme means and why it matters. People who observed the research can contribute to the analysis, helping build group consensus and guard against focusing only on what one researcher considers important.
Document findings in a form that suits the project and team. Include the research methods, who took part and how information was collected, so readers can understand the basis for the insights; where appropriate, prioritise recommendations so critical findings are distinguishable from minor ones.
In this guide
- Choosing a customer experience question worth investigatingTurn a broad CX concern into a research question with a defined customer, task, uncertainty and decision.
- Customer satisfaction versus task successUnderstand the difference between customer satisfaction and task success, and how to interpret cases where the two disagree.
- Combining interviews with observed customer behaviourPlan interviews and task observation around one question, compare what customers say and do, and keep interpretations traceable.
