All of the remaining user research methods, including surveys, focus groups, interviews, and ethnographic studies, require the selection of representative samples of users to participate in the research studies. With the possible exception of surveys, it’s rarely possible to study every user of a web site.
The definition and prioritization of intended and actual audiences for the site is obviously a critical factor. As we discussed earlier, there are myriad ways of slicing and dicing these audiences. Just as you define a primary hierarchy for your web site, you also need to define a primary hierarchy for participant selection. This hierarchy should strike a balance between the traditional ways that an organization views its customers (e.g., home users, business users, value-added resellers) and the distinctions an information architect is interested in (e.g., people familiar with the old site, people unfamiliar with the old site).
For large projects, the information architect should consider working with a traditional market research firm that has experience defining audience categories, developing profiles of participants within those categories, recruiting participants, and handling logistics like facilities, incentives, and note taking.
Surveys are a broad-and-shallow research tool that provide an opportunity to gather input from a large number of people relatively quickly and inexpensively. Surveys can be done via email, web, telephone, ...