Strong online exams come from process, not just software. The most secure platform in the world cannot rescue a paper whose instructions were unclear, whose access window was communicated the evening before, or whose results were exported three weeks late. These best practices reduce support tickets, disputes, and malpractice — whether you are running a class unit test or a hiring screen for five hundred candidates. They are organised around the three phases that matter: before the exam, during the exam, and after the exam.
Exam software handles the mechanics: timing, submission, scoring. It cannot handle the human side: candidate anxiety, unclear rules, last-minute changes, and disputed results. Most exam-day crises trace back to a process gap, not a software failure. Documenting your process and running it consistently across every exam session turns chaos into a repeatable operation.
Lock the attendee list and access settings at least 24 hours before the window opens. Late changes — adding candidates, adjusting time limits, toggling integrity controls — introduce inconsistency and give grounds for disputes. Confirm that shuffle is enabled and that the access window matches your advertised schedule down to the minute.
Candidates who know what to expect cause fewer problems on exam day. Send exam details at least 48 hours in advance. Include duration, start time, device requirements, and the support contact for technical issues. A short practice link — three to five questions in the same format — dramatically reduces login issues and "how do I navigate" messages during the real exam.
Many support messages during a live exam come from candidates who did not read the instructions carefully. The solution is to make the invite email self-contained. Every candidate should be able to answer these five questions from the invite alone: When does the exam start and end? How long do I have once I begin? What device and browser should I use? What happens if I lose my internet connection mid-exam? Who do I contact if something goes wrong?
The timer and auto-submit are the most important integrity tools during a running exam. Confirm that auto-submit fires correctly — test it in advance by letting a trial attempt run to zero. If your platform does not auto-submit server-side when time expires, candidates can simply wait out the clock. EasyEvaluate auto-submits on timer expiry regardless of the candidate's browser state, which removes this risk entirely.
Not every exam needs the same security level. A low-stakes formative quiz benefits from shuffle but does not need webcam proctoring. A graded semester exam or a ranking mock justifies enabling fullscreen enforcement and tab-switch limits. Reserve active AI proctoring for exams where results determine access to limited seats, scholarships, or certifications — the overhead of reviewing flagged recordings should be proportional to the decision being made.
When a candidate reports a technical issue mid-exam, the first step is not extending the time — it is diagnosing the issue. Ask for the exact error message or behaviour they are seeing. Most mid-exam issues are browser-specific: switching from Firefox to Chrome resolves roughly half of all compatibility problems. Connection drops are the other common cause. If the platform saves answers progressively, reconnecting restores the attempt; if not, you need a documented exception policy ready.
Keep a short technical troubleshooting guide ready to share on your support channel during the window: clear browser cache, try Chrome incognito mode, check if mobile data is more stable than home WiFi. Most candidates can resolve their own issue within three minutes with the right prompt, which is faster than escalating to the vendor.
Export scores within 24 hours for graded exams. Delayed results generate speculation, support requests, and complaints. For formative quizzes, showing results immediately in the platform is usually fine and helps learning. For summative exams, consider a brief hold period to allow integrity review before releasing.
Keep the exam settings log and the candidate attempt data for at least one week after results are released. When candidates dispute their score, the log shows exactly what rules were in effect. Being able to share a screenshot of the settings page resolves most disputes immediately. If a candidate reports a technical failure, cross-reference their attempt log — the platform records submission time and answer state.
Write a one-page process doc for each exam type you run regularly: unit test, mock, hiring screen. Include the default settings, the communication timeline, the support channel, and the result export steps. Share it with everyone who touches the exam workflow. Consistency across sessions is what makes your data comparable term over term.
Some candidates have documented needs that affect how they interact with a timed online exam — extended time, larger text, or assistance devices. Build an exception process before the first exam cycle, not in response to a complaint. Define who can request accommodation, what documentation is required, and what the alternative arrangement looks like. Candidates with genuine needs who receive fair accommodation perform more consistently and raise fewer post-exam disputes.
