Every institution has a version of the story. The online exam window opens. Within minutes, the support inbox floods. Students are locked out, frozen mid question, or staring at error screens. By the time it’s over, leadership has already drafted a strongly worded message to the platform vendor.
It’s an understandable reaction. But in most cases, it’s the wrong one.
The truth, uncomfortable as it is for assessment teams to hear, is that the majority of online exam failures are not platform failures at all. They are governance failures, design failures, and preparation failures dressed up as technology problems.
And until institutions are willing to look inward, they will keep repeating the same expensive mistakes, with different vendors and the same outcomes.
The Vendor Blame Cycle Is Costing Institutions More Than They Realise
When an online exam goes wrong, the platform takes the first blow. This is partly institutional habit; there is always a contract to point to, a service level agreement to invoke, a helpdesk to escalate to. Blaming the vendor is administratively convenient.
But the pattern that follows is where the real cost lies. Institutions conduct a post incident review, conclude that the platform was inadequate, and begin procurement for a replacement. New vendor. New implementation. New training. New costs. And then, typically within a year or two, a variation of the same incident occurs again.
This cycle repeats because the root cause was never addressed. The platform changed. The governance did not.
Open Praxis research on online assessment integrity noted that technical failures (including software crashes and malfunctioning proctoring tools) do occur, and that when they do, they generate appeals and additional workloads for faculty that platform upgrades cannot resolve.
But what the research makes equally clear is that these technical events are far less frequent than the institutional conditions that surround and amplify them.
What Actually Goes Wrong: Three Failure Modes Institutions Own
1. Exam Design That Was Never Tested at Scale
Many institutions design online exams the way they once designed paper ones, sequentially, by subject matter experts, with little consideration for how the assessment will behave under concurrent load, across different devices, or in conditions the exam team cannot control.
As Open Praxis findings on LMS assessment capacity observe, learning management systems are widely adopted but often lack the capacity to support innovative and authentic forms of assessment, not because the technology is inherently limited, but because the assessment design placed around them was never stress tested.
The design problem compounds when institutions treat the exam format as fixed. Question types, timing structures, and submission pathways that made sense in a physical hall frequently create friction online. When they do, it looks like a platform problem. It is, in reality, a design problem that the platform was never equipped to solve.
2. Candidate Preparation That Exists Only on Paper
This is perhaps the most consistent blind spot in institutional assessment practice. Candidates are sent into high stakes online exams having received, at best, a PDF of instructions and a link to a FAQ page.
The assumption is that students who navigate digital environments daily will navigate an unfamiliar exam interface without difficulty. That assumption is wrong with remarkable regularity.
Jisc’s 2023/24 Digital Experience Insights survey of UK higher education students found that while 72% rated their institution’s support for learning with technology as above average, skills and training opportunities were not always felt to be available across the academic year or throughout a course of study. That gap matters most precisely when it matters most, during assessment.
The same pattern plays out in vocational and professional training contexts. Students may be experienced learners. They are not necessarily experienced online exam takers. Familiarity with a smartphone does not translate to comfort with a timed, proctored, browser locked assessment environment.
When something goes wrong during the exam (a warning message, a flagged behaviour, a slow upload), a candidate who has never encountered the interface before is far more likely to panic, close the browser, or take an action that invalidates their session.
That is not a platform failure. It is a preparation failure.
3. Institutional Governance That Nobody Owns
Behind most online exam incidents, if you follow the thread far enough, you find a governance gap. Nobody with authority over both the technology and the assessment process was in the room when the decisions were made. IT managed the platform. The academic team set the exam. Student services handled queries. Nobody coordinated across all three.
VerifyEd’s analysis of examination failure in UK universities drew on over 50 interviews with educational staff, from course leaders to pro vice chancellors. It found that the makeshift solutions adopted during the pandemic have since become permanent fixtures.
Practices that were excused as emergency measures (loosened invigilation standards, fragmented oversight, and policies assembled under pressure) were never revisited once the crisis passed. Institutions normalised the dysfunction rather than dismantling it.
This structural disconnect is not unique to the United Kingdom. In Australia, ASQA’s November 2024 provider roundtable findings made clear that the integrity challenges facing registered training organisations are not primarily technological but process based, centred on how training and assessment are designed, delivered, and governed.
The regulator’s focus on provider diligence and holistic operational review reflects a growing recognition that platform sophistication is irrelevant if the human systems around it are not fit for purpose.
The Proctoring Paradox
Online proctoring deserves its own mention, because it sits at the intersection of all three failure modes above, and it generates more misplaced blame than almost any other component of online assessment.
Security researcher Bruce Schneier, writing on the security failures of online exam proctoring, cited researchers who found a significant gulf between what proctoring technology promises and what it delivers on the ground, with systems that assume everyone looks the same, takes tests the same way, and responds to stressful situations in the same way.
When proctoring software flags a student unfairly for adjusting their glasses, for ambient background noise, or for an atypical eye movement pattern, institutions typically escalate the complaint to the platform vendor.
But who configured the sensitivity thresholds? Who decided which behaviours would trigger a flag? Who briefed candidates on what to expect? In almost every case, those decisions were made institutionally or not made at all, leaving default settings in place that were never reviewed for the institution’s specific context.
Technical problems such as connectivity issues, software glitches, and hardware failures do disrupt exam processes and add to candidate anxiety. These are real. They are also, in most cases, foreseeable, and the difference between a foreseeable problem and a crisis is whether the institution built a contingency process before the exam began.
What the Evidence Suggests Institutions Should Do Instead
The shift required is not primarily technological. It is cultural and operational.
First, institutions need to conduct an honest audit of where their online exam failures actually originate, not where they surface, but where they begin. A student locked out at the login screen may be experiencing a platform error.
Or they may be using credentials that were never properly provisioned, on a device that was never checked against system requirements, for an exam they received no meaningful preparation for. The surface failure and the root cause are rarely the same thing.
Second, candidate readiness needs to become a structured part of exam delivery, not an afterthought. Practice exams on the actual platform, in conditions that mirror the real assessment environment, are not a luxury. For high stakes assessments, they are a baseline requirement.
The challenges of online exams that consistently surface in post incident reviews (access issues, interface unfamiliarity, submission errors) are, in large part, addressable through structured pre exam preparation that most institutions simply do not provide.
Third, assessment governance needs a single point of accountability. Someone must own the full chain from exam design through to results release, with authority over the decisions made at each stage.
Without that, every incident becomes a jurisdictional dispute between teams, and the platform becomes the default target because it is the one party that cannot push back in the internal meeting.
The Platform Is a Tool. The Institution Is the Craftsman.
No online exam platform, however well engineered, can compensate for poorly designed assessments, underprepared candidates, or fragmented institutional governance. The best platforms provide stability, flexibility, and support. What they cannot provide is the institutional will to use them well.
The question worth asking after the next online exam incident is not “what did the platform do wrong?” It is: “What did we fail to build around it?”
That shift in framing, from vendor accountability to institutional ownership, is not comfortable. But it is where improvement actually begins.
