What research does a website need before launch?

Start with the decisions that the website needs to help people make and the tasks it needs to support. Research should establish more than job titles or broad personas: it should explore audience questions, triggers, objections, alternatives, decision-makers and the proof people need before they act. Stakeholder discussions and desk research may be enough to clarify a focused brief, while a new market, audience segment or proposition may need deeper market analysis, competitive audits, social listening, persona development or customer journey mapping.

The aim is not to conduct every possible research activity. It is to replace assumptions with a strategic foundation that gives design, writing and development a clear direction. Define what needs validation before choosing methods, then record which findings are sufficiently reliable to guide messaging, information architecture and user journeys.

Behavioural data can indicate where visitors click, view, leave or complete a task, but direct feedback adds context about what audiences understand, need and choose to share. Asking for feedback can also show that an organisation is listening, while first-party data gives marketers information collected directly from users.

Use these sources for different purposes rather than treating one as a substitute for another. Audience conversations can clarify the language, level of detail and framing people need. Usage data can help identify patterns worth investigating and measure whether a changed experience is behaving differently. Stakeholder input can supply operational and commercial context, but it should not be the only basis for decisions about audience needs. Bring the findings together into a small number of prioritised problems, rather than a long unranked list of observations.

Pre-launch testing should reflect the conditions in which people will actually use the website. Accessibility is not limited to a single user group or device. People may be using a slow connection, a keyboard rather than a mouse, a screen reader, a mobile phone in bright light or a setting where concentration is limited. These conditions can change whether content is understandable and whether a task is possible to complete.

Test key journeys with people who reflect relevant access needs and use the technologies they rely on. Review whether navigation, contrast, timing, interaction and content structure work without making assumptions about users. This also improves the quality of general usability testing, because it shifts attention from whether an interface looks complete to whether people can use it effectively in context.

Prioritise findings according to their effect on the audience and the website's purpose. A finding should lead to a decision: clarify a proposition, change content structure, adjust a journey, simplify an interaction or test a competing approach. Creative work is stronger when it is grounded in a clear articulation of the challenge, audience expectations and the insight behind the decision.

For complex audiences, avoid assuming that one format or level of detail works for everyone. Research for Siemens Digital Industries Software found a spectrum from technical engineers to senior executives, each needing different depth, language and framing. The resulting experience used accessible entry points and allowed visitors to choose between detailed technical papers and executive summaries. This is a useful pattern when research identifies materially different information needs across an audience.