The plans are being written now. Across the sector this month, in the weeks between one admissions cycle closing and the next cohort arriving, teams are assembling the documents that will govern the coming year, among them the annual monitoring returns, the responses to survey results and the plan a committee asked for in June. Most will be competent. A good number will commit an institution to actions on matters whose causes it has not established, and will do so without anybody involved behaving carelessly.
There is a familiar complaint that universities hurry towards action and neglect diagnosis, and it has been made often enough to have stopped doing much work. Institutions do skip diagnosis, rarely because anyone believes it unimportant and usually because the document does not ask for it, the committee date is fixed and a plan arriving without actions is not a plan and will be sent back. The argument here is narrower, and rather less comfortable for being so. It concerns the form of the plan itself, and what that form can hold.
The form and what it can hold
It is worth pausing on what the document looks like. Almost everywhere, the improvement plan takes the shape of a table with four or five columns: the action, the owner, the target date, the current status and, in more elaborate versions, the evidence that will demonstrate completion. This is a sensible design with obvious virtues. It renders commitments legible, allows a chair to see at a glance what has moved and identifies in advance whoever will be asked about an item in twelve months’ time.
The design carries a cost, and it is not the cost usually identified. The requirement that each action carry a single named owner is the load-bearing convention of the whole form, since without it nothing can be tracked. That requirement performs a subtle selection before anyone has written a word. A problem sitting inside one department, or one professional service, can be given an owner, and so becomes an action, and so enters the plan. A problem arising between a timetabling office, an admissions team and a school, belonging wholly to none of them, cannot be given an owner without somebody being made accountable for what they do not control. What tends to happen instead is that it is decomposed into the parts that can be owned, each addressed conscientiously, whilst the interaction which produced it is addressed by no one.
This may partly explain the item that keeps returning, familiar to anyone who has read a long run of these papers. It appears, it is actioned, it is marked complete and eighteen months later it comes back in altered wording under a different owner. The usual explanation is weak implementation, or attention drifting once the review has passed, and sometimes that is right. An item recurring across several plans, several owners and more than one restructure has arguably demonstrated that its cause was never inside the boundary of any action taken against it. The handbook was revised, and revised well. What made a revised handbook seem necessary was not itself a handbook problem, and nothing in the plan reached it.
What another system concluded
It would be reassuring to treat this as a peculiarity of universities, though another part of the public sector has examined much the same question with more at stake. The health service in England spent much of the last decade reviewing how it responded to safety incidents, and in 2022 replaced a framework which had stood since 2015. Root cause analysis, its supporting guidance states, prompts simple linear cause-and-effect analysis and has consistently failed to deliver benefits of the scale and quality needed. The replacement declines to mandate investigation as the only route to learning, and accepts that for some incidents it is appropriate to record what occurred and go no further. Its governing principle is that safety arises from the interaction between components rather than from any single one.
Two elements of that judgement bear on a university’s position, and they pull in different directions. The first is the systems point, which supports the argument advanced here, since a method resolving every problem into an owner and an action is committed in advance to a linear account of cause. The second is more awkward for anyone inclined to call for more diagnosis. That system had not investigated too little. It had investigated a great deal, at considerable expense, producing plans in volume without the improvement the volume was meant to secure. Analysis applied indiscriminately appears to have been part of the difficulty rather than its remedy, which suggests the useful question is less how much diagnosis an institution does than which problems it spends that effort on.
None of this should be taken as an argument that institutions ought to act less and deliberate more. A cohort will not wait whilst a university satisfies itself about causation, and a student whose experience is poor this term is owed something better next term. There are people in every institution who call for further diagnosis as a way of declining to act, and colleagues are usually quite good at telling which is which. The pressure under which these plans are written is real, and the teams producing them behave reasonably given what is asked of them. The concern is narrower than a complaint about institutional habit. It is that the plan as designed offers nowhere to put a problem with no owner, so such problems tend not to be written down, which is a different thing from their not being understood.
The plans drafted this month will mostly be approved, and will mostly be adequate to what was asked of them. Whether they are adequate to what an institution faces is a separate question, and not one the document can raise about itself. Could a leadership team say, without going and looking, which items on this year’s plan also appeared on the last three? Where is a problem supposed to go when it belongs to the space between two functions and can be assigned to neither? And is it possible that the discipline of naming an owner, which has served the sector well and nobody would wish to lose, has been mistaken for the discipline of understanding a cause? These appear to be the questions most likely to repay a governing body’s attention, though they are harder to place on an agenda than a plan is.