Recruit actual or likely users whose experience fits the task you need to study. Define eligibility and relevant variation before screening, ask neutral questions about behavior, arrange access needs, and document who was missing. A filled calendar or a small sample does not establish that your findings represent all users.
Recruit for a real situation, not a flattering persona
A team can run a polished usability session with the wrong person and learn very little about the people who will use the service. Recruiting only colleagues is an obvious example. A subtler one is recruiting frequent online users to evaluate the first attempt at a task: their existing knowledge can hide the very problem you need to see.
Start with the decision the study will inform. This guide follows a fictional public library redesign. The team wants to know whether borrowers who have never attempted the library's online renewal flow can find the right book, request more time, and understand when a renewal is unavailable. It needs people who actually borrow books and may need more time, rather than people who merely say they like libraries.
GOV.UK's participant guidance emphasizes actual or likely users, a mix of access needs, and recruitment criteria tied to the study. Its recruitment-brief guide gives a practical way to communicate those requirements. The example below turns those principles into a filled brief, a neutral screener, candidate decisions, an invitation, and a coverage check. All people and recruitment outcomes shown here are invented teaching examples.
Write a brief another person could recruit from
The library study includes adults who borrowed a physical book in the past year and have never attempted a renewal through this particular library's online account. People who renewed at a desk or by phone can qualify. That distinction matters: experience with the underlying need is useful; prior knowledge of the exact interface would change the first-use question.
Do not add requirements simply because they make scheduling easier. Owning a recent laptop is not necessary for this service, so it should not become an eligibility rule. Nor should being available during working hours. Arrange session options around the relevant users, and record what remains uncovered if the available resources cannot support an important group.
The fictional first round below plans six sessions to investigate specific barriers and revise the design. Six is a scheduling decision for this example, not evidence that all problems will be found or that the sample represents the library population. The team will examine the findings and the gaps before deciding the next round.
| Field | Example brief |
|---|---|
| Decision | Revise account navigation and renewal feedback before the next prototype iteration |
| Eligible experience | Adult borrower of a physical library book within the past 12 months; has never attempted online renewal in this library account |
| Relevant variation | Include people who use a phone, people who use a computer, and people who normally seek help with online tasks |
| Access coverage | Arrange participation with people using relevant access tools or support; ask about practical requirements, not a diagnosis |
| Exclusions and reason | Library staff and project team: insider knowledge; people who previously attempted this exact renewal flow, whether successful or not: different research question |
| Round scope | Six planned qualitative sessions; not a representative completion-rate survey |
| Format | Moderated use of a fictional account on an accessible prototype; remote or suitable supported venue |
| Participant burden | Allow 45 minutes; explain recording choices, breaks, compensation, and cancellation terms before booking |
| Recruitment routes | Library noticeboard and newsletter opt-in, plus a community access partner; retain route in the recruitment log |
| Revisit condition | Missing access contexts, repeated instruction confusion, or new user groups revealed by the round trigger further planning |
Use a screener that does not reveal the desired answer
A screener should distinguish relevant experience without teaching candidates how to qualify. 'Are you an inexperienced borrower who struggles with renewals?' reveals both the target and a story the participant may feel expected to tell. Ask about what they have done, then apply the eligibility rules privately.
The questions below are a complete small screener for the example. Show the participant-facing wording and choices; keep the decision column for the recruiter. Do not display a running eligibility score. The short open response helps clarify what a person means by 'renewed online' without asking them to upload records or reveal a library account number.
Support needs are a planning input, not an automatic rejection. Someone who uses a screen reader may be exactly the person the round needs, while the research tool itself may need adaptation. If a candidate does not want to answer an optional support question, offer a way to discuss practical arrangements privately before the session.
Pilot the screener too. In this case, the team should check whether borrowers understand the difference between reserving a new book and extending an existing loan. If the question causes confusion, clarify the wording for everyone rather than interpreting each answer in whichever way fills the schedule.
| Participant-facing question | Answer choices or response | Recruiter decision |
|---|---|---|
| Are you 18 or older? | Yes / No / Prefer not to say | This particular round uses an adult protocol; do not enroll under that protocol without confirming eligibility |
| Which library activities have you done in the past 12 months? | Borrowed a physical book / Borrowed an ebook / Used a study space / Attended an event / None of these; select all that apply | Physical-book borrowing is relevant; other activities alone do not establish this round’s task context |
| How have you tried to keep a borrowed physical book for longer, including attempts that did not succeed? | At a desk / By phone / In this library’s online account, whether or not it worked / Another library’s website or app / I have not done this; select all that apply | Keep desk, phone, other-library, and no-renewal experience eligible; any previous attempt at this exact online flow is outside the first-use group |
| Please briefly describe your most recent experience, if any. | Short response; no account numbers, book records, or personal documents needed | Resolve ambiguous renewal versus reservation answers, and whether an online attempt occurred, with a neutral clarification; keep eligibility pending until confirmed |
| What do you normally use for online library or similar tasks? | Phone / Tablet / Computer / Shared or public device / Someone helps me / I do not usually do these online; select all that apply | Plan relevant device and support variation; do not reject solely for needing help |
| Do you work for this library or on this website project? | Yes / No | Exclude insider familiarity from this first-use round; consider it separately if needed |
| What would help you take part comfortably? | Optional response about contact method, device, access tools, breaks, venue, or other practical support | Arrange support and discuss options; do not require a medical history |
| If invited, which session arrangements could work for you? | Offer the actual dates, remote or venue options, and a way to suggest alternatives | Schedule after eligibility; do not treat availability as evidence of user relevance |
Make recruitment decisions visible before the calendar fills
The candidate log below demonstrates how the brief changes selection. Candidate A is relevant despite never renewing a book before. Candidate D is useful for a different study but not this first-use round. Candidate E is not eligible merely because they use the library frequently; their experience is with ebooks and study space, not the physical-loan task.
Candidate C should not disappear from the plan because their usual setup takes more effort to accommodate. Check whether the prototype and research platform can be used with that setup, and arrange an appropriate session. If the setup is inaccessible, that is a limitation in your research preparation, not evidence that the participant is unsuitable.
These decisions preserve the question the team chose to study. They do not imply that excluded groups are unimportant. Existing online users may face different problems, and staff may provide valuable operational context. Plan those conversations under their own questions instead of mixing them into one sample and calling it representative.
| Candidate and context | Decision for this round | Reason or next action |
|---|---|---|
| A: borrowed print books; never renewed; phone user | Eligible | First-use task is plausible; include phone context |
| B: borrowed print books; renews by phone; usually asks a family member for online help | Eligible | Relevant task and support context; arrange session without assuming independent computer use |
| C: borrowed print books; uses a screen reader; renewed at the desk | Eligible, arrange access setup | Confirm research platform and prototype work with the usual setup |
| D: renews weekly in this library’s online account | Not this first-use round | Invite only to a separate experienced-user round if that question is in scope |
| E: uses library study space and ebooks, no physical borrowing | Not this round | Does not meet the declared physical-loan experience |
| F: library employee who helped write the renewal instructions | Not this round | Insider knowledge would change the task |
| G: borrowed print books; renewed on another library’s site; laptop user | Eligible | Other-site familiarity is recorded, not mistaken for knowledge of this flow |
| H: borrowed print books; no online renewal; public computer user | Eligible, arrange venue | Shared-device context matters; do not impose home-device ownership |
Send an invitation that explains the commitment clearly
An invitation should let an eligible person decide whether taking part fits their circumstances. It needs the purpose, activity, duration, location or connection method, compensation arrangement, recording choices, contact, and an easy way to decline. The exact privacy and consent information must match what your organization actually does.
Here is an editable invitation for the fictional library. Replace every bracketed field with a confirmed detail before sending: 'We are inviting borrowers to help us improve the library website. In a 45-minute session, you will try a few tasks using a fictional library account and tell us what you expect to happen. We are evaluating the website, not your skills. The session is [remote link or venue] on [date, time, and time zone]. We offer [amount and payment method] for your time under [clear payment and cancellation terms]. We would like to record [screen/audio and purpose]; [explain the actual recording choice]. You can ask questions, take a break, or stop. Please read [information sheet] and contact [research contact] about access needs or any questions. Reply by [date] if you would like to take part; declining will not affect library services.'
Keep administrative details in the invitation and study information, not hidden until a participant arrives. Do not make compensation depend on finding problems, praising the design, or successfully completing a task. If the team cancels or a technical fault prevents the session, follow the terms communicated in advance rather than improvising based on how useful the data seemed.
Send a confirmation that repeats the agreed time and practical setup. A short optional connection check can help with unfamiliar remote software, but it should not train the renewal task itself. If someone needs another contact method or venue, resolve that with them before deciding they are a no-show.
Track coverage without pretending to have a representative sample
A full calendar is not the same as adequate coverage. Imagine the team books A, B, C, G, H, and one additional phone borrower for six sessions. If C cannot attend and is silently replaced with another confident laptop user, the total stays six while the access context disappears. The round's report should make that change visible.
Keep a small coverage log separate from contact details. Use participant codes in analysis material; retain the minimum administrative information needed to manage attendance and compensation according to the study's declared arrangements. Do not copy personal screener answers into every stakeholder presentation.
The fictional log below shows why outcome categories matter. An eligible person who declines, a person who cannot find an accessible session, and a person who attends but cannot use the prototype are three different situations. Combining them into a generic dropout number hides information about who the research is reaching.
| Record | Why it matters | Example note |
|---|---|---|
| Recruitment route | Channels can reach different people | Newsletter, noticeboard, or community partner |
| Relevant task history | Explains fit with the study question | Desk renewal; no use of this online flow |
| Device and access context | Shows what was and was not examined | Public computer session arranged |
| Invitation and attendance | Separates recruitment loss from task performance | Eligible; declined offered time; alternative pending |
| Practical barrier | Identifies a research-process gap | Recording tool did not work with usual setup; session not treated as a usability failure |
| Coverage gap | Keeps missing perspectives in the next decision | No screen-reader session completed this round; recruit and prepare it before claiming that context was examined |
Carry the recruitment limits into the findings
A useful report might say: 'This round examined first use of physical-book renewal with six adult borrowers recruited through library and community channels. Participants included phone, computer, and supported-use contexts. The table lists the exact contexts completed. We identified navigation and feedback problems to revise; we did not estimate the prevalence of those problems among all borrowers.' Replace those example details with what actually happened.
Avoid reporting '80% of users find renewal confusing' from a small convenience sample without a design that supports that population claim. It is more informative to describe the observed behavior, task state, and people involved: for example, which participants interpreted a declined renewal as successful and what wording led them there. The interface evidence helps the team act; a broad percentage can conceal the recruitment limits.
Before commissioning the next round, revisit the original decision. If navigation changes substantially, recruit fresh first-use participants for that question. If the unresolved issue is expert efficiency, recruit experienced borrowers under a new brief. If a relevant access context was missed, complete that work rather than treating the six-session target as a stopping rule.
To reuse this plan, begin with one sentence: 'We need people who have experienced [situation] so we can decide [design question].' Then fill the brief, screener, invitation, candidate log, and coverage note. Each item should preserve that connection. Recruitment succeeds when the people and conditions allow the study to answer its question, not when every appointment slot is filled.