Real Time Exam Supervision Needs More Than Extra Staff

Large online exam cohorts are easier to manage when capacity is planned around peak demand rather than headcount alone. For assessment directors, that means aligning staffing with technical capacity, support routes, escalation processes and contingency planning before hundreds of candidates arrive within the same short window.

Staffing ratios still provide a useful baseline for coverage. Their limitation is that several demands may converge at once, from authentication and technical questions to incidents requiring academic judgement. The more useful capacity question is therefore whether the wider supervision operation can absorb those demands together.

Capacity Planning Starts With Peak Demand

An effective capacity plan begins with the shape of the examination itself. Cohort size matters, but so do expected concurrency, exam duration, authentication requirements, support demand and the types of incidents that may require intervention.

Average workload offers limited guidance when hundreds of candidates may authenticate within the same short window. A session that appears manageable when demand is spread across two hours may behave very differently when logins, identity checks and technical questions converge within the first ten minutes. Capacity planning therefore needs to consider peak concurrency, not simply total candidate numbers.

Preparation emerged as an operational control when researchers examined the introduction of digital invigilation in a 2025 International Journal for Educational Integrity study. Lecturers used advance communication, guidance, practice opportunities, mock assessment and real time support as part of implementation, giving students opportunities to become familiar with the process before the higher stakes assessment took place.

For assessment teams, that suggests usable capacity can be created before supervisors begin monitoring anyone. Familiarisation may reduce avoidable questions during the live session. System checks can expose access difficulties while there is still time to address them. Clear instructions establish what candidates should do when something unexpected occurs.

Preparation does not remove peak demand. It reduces the amount of routine uncertainty competing with events that genuinely require professional attention.

Clear Routing Protects Professional Attention

Once an exam begins, different types of events should not automatically enter the same queue.

An authentication difficulty does not necessarily require the same response as suspected misconduct. A lost connection is different again. Some situations may need technical support, some require a supervisor to communicate with the candidate, and others need an educator or assessment specialist to make a consequential judgement.

That division of responsibility is also visible in a 2025 systematic narrative review published in Open Praxis, which mapped online proctoring across authentication, monitoring, immediate intervention and post exam review. Responsibility could move between automated processes and human judgement at different points, while potential rule violations still required human review.

Assessment leaders can use those distinctions as an operating framework. Before the exam window, teams can decide which events supervisors resolve directly, which move to technical support and which require academic escalation. They can also define what information should accompany an escalation so the next person does not need to reconstruct the incident from the beginning.

Clear routing becomes particularly valuable when several demands occur together. Routine technical events can continue along one path while professional attention remains available for incidents requiring interpretation or academic judgement.

The benefit is not simply a faster response. It is a better allocation of human attention across the examination.

Infrastructure Has to Scale With Supervision

Human capacity and technical capacity also need to grow together.

Adding supervisors has limited value if concurrent video streams, authentication services, monitoring systems or communication channels are already approaching their operating limits. The same principle applies to support tools. A large supervising team still depends on being able to see relevant information, communicate consistently and move an issue to the right person without unnecessary delay.

Assessment teams therefore need to consider expected concurrency across both people and systems. A larger cohort may increase the number of live feeds, simultaneous authentication events, support requests and records being handled at once. Planning needs to account for those combined demands rather than assuming that additional staffing automatically creates equivalent operational capacity.

Technical support should be considered within the same model.

A candidate who loses connectivity, cannot restore a camera feed or experiences an access problem creates an immediate operational decision. The supervisor may need to establish what happened, communicate the next step and determine whether another team needs to become involved.

Defining technical support routes alongside supervision procedures helps keep those responsibilities clear. Common technical events can have agreed responses, support responsibilities can be visible to supervisors, and escalation criteria can be established before the examination begins.

That is why planning the infrastructure for live online proctoring is better understood as an exercise in coordinating supervision, support and incident handling, rather than simply connecting more people to candidate video feeds.

Additional supervisors increase human availability. Infrastructure determines whether that availability remains usable when the cohort and the number of simultaneous events increase.

Readiness Testing Turns Capacity Assumptions Into Evidence

Capacity planning becomes more useful when assessment teams test assumptions before a live examination rather than discovering their limits during it.

A rehearsal does not need to recreate every possible failure. It can focus on the moments most likely to generate concentrated demand: candidates joining at once, authentication delays, several support requests arriving together, connectivity interruptions or an incident that needs to move from a supervisor to an assessment decision maker.

The same emphasis on preparing for different assessment risks appears in Australia’s Tertiary Education Quality and Standards Agency resources. Its 2026 Academic Integrity Toolkit section on assessment design and security brings together frameworks and approaches institutions can consider when managing assessment security, including material addressing online invigilated examinations.

For educators, the useful principle is proportional preparation. A small examination with familiar participants may require a different response model from a large simultaneous sitting where supervision and support demands could rise within minutes.

Testing also gives teams an opportunity to confirm contingency decisions before they are needed. Assessment leaders can determine in advance how supervision should continue if connectivity deteriorates, support queues rise unexpectedly or a wider technical interruption affects part of the cohort.

The purpose is not to predict every disruption. It is to avoid making fundamental continuity decisions for the first time during a live examination.

Teams can establish when an assessment should continue, when a candidate should move to another support path, who has authority to pause or resume activity, and how information should be communicated if circumstances change. Those decisions create usable capacity because supervisors know what to do when normal conditions no longer apply.

A rehearsal can then reveal where the model needs adjustment. If technical questions quickly consume supervisor attention, the support route can be changed. If escalation depends too heavily on one person being continuously available, another pathway can be defined. If several supervisors interpret the same event differently, response guidance can be clarified before results depend on it.

Capacity Is Better Measured Under Pressure

Once supervision is treated as a coordinated operation, assessment leaders gain a more useful way to evaluate readiness.

Candidate to supervisor ratios still matter, but they can sit alongside operational questions. How does the system perform when authentication demand peaks? How quickly can a routine support request reach the right person? Which events consume the most supervisor time? Where does academic judgement enter the process? What happens when several incidents occur together?

Those questions turn capacity from a theoretical number into something teams can observe.

A session with very few incidents may show that preparation worked particularly well. Another may reveal that staffing levels were adequate but technical support became concentrated at the beginning of the exam. A third may show that infrastructure remained stable while an escalation process created unnecessary delay.

Reviewing those patterns allows assessment teams to change the part of the operation that actually constrained performance rather than automatically changing headcount.

That is the deeper distinction behind real time supervision capacity. More staff increase the number of people available to act. Infrastructure, role clarity, escalation pathways, contingency procedures and readiness testing determine how effectively those people can act when demand reaches its peak.

For assessment directors, the meaningful measure is therefore not simply how many candidates a team can supervise under ideal conditions. It is how many simultaneous situations the assessment operation can resolve consistently while keeping professional judgement focused on the decisions that genuinely require it.