People leave for reasons that have nothing to do with study

Withdrawal is recorded as an academic event and is usually a financial, health or family one. Support services exist for that gap and are resourced as though the gap were small.

Ask a student who withdrew why they left and the answer is rarely about the course. Their hours changed at work. A relative became unwell and they became the person who drives to appointments. The car failed and the campus is forty minutes away. The relationship ended and the housing went with it. Money ran out in week nine and the choice was between a textbook and the electricity. Each of these ends a course as effectively as failing every assessment, and none of them appears anywhere in the academic record.

The institutional response is shaped by what is visible. Results are visible, so intervention triggers on results, which means the system responds after the consequence rather than to the cause. By the time somebody has failed two units, the financial or personal situation that produced the failure has usually been running for months and has become considerably harder to address.

The second problem is that the student is the least likely person to raise it. Disclosure requires believing that the institution will respond usefully, that the response will not carry a consequence, and that the problem is the kind of thing that counts. Students consistently underestimate all three, and they particularly underestimate the last: a large proportion believe that financial difficulty, caring responsibilities or ordinary distress are not what support services are for, and that raising them would be an imposition.

This produces the pattern that anybody working in this area recognises. The student who arrives at the counter in week eleven has been in difficulty since week three, has already borrowed from family, already missed two assessments, and is now asking for something the institution cannot easily provide because the window for the cheap intervention has closed. Everybody involved would have preferred the week three conversation and nothing in the system produced it.

The third feature is that support functions are frequently structured around problems rather than around people. There is a counselling service, a financial office, a disability service, an academic skills unit and an international student adviser, each with its own referral route and intake process. A student in difficulty usually has an interlocking set of problems rather than one, and being referred sequentially to four services, each requiring them to explain their circumstances again, is a strong deterrent to persisting.

The fourth is resourcing. Support services are almost universally the function measured last and cut first, because their output is a non-event: the student who did not withdraw generates no record. A service that prevented forty withdrawals cannot demonstrate it, while the cost of the staff is entirely visible in a budget, and that asymmetry operates every year in the same direction.

None of this argues that institutions are indifferent. Most people working in this area are acutely aware of all of it. The argument is that the design of the system consistently detects the wrong signal at the wrong time and routes it through a structure built for the institution's convenience rather than for a person in difficulty, and that this is fixable without any change in resourcing by altering when the contact happens and who makes it.