Writing a School Improvement Plan Staff Actually Use

Why so many improvement plans end up doing nothing

Walk into most schools in July and you’ll find last year’s improvement plan somewhere between the filing cabinet and the shared drive, several pages long, colour-coded, and almost entirely untouched since October half term. It was written with real intent. It got signed off by governors. And then the year happened around it rather than because of it.

This isn’t a motivation problem. Heads who write these plans are not lazy or careless, and the staff who nod along in the September briefing are not disengaged. The plan itself is usually the issue: too broad to guide a decision, too vague to check progress against, and often shaped more by what an inspector might want to see than by what a Year 4 teacher actually needs to change on a Tuesday morning.

Too many priorities at once

The most common failure is scope. A plan with ten strategic priorities, each with its own set of sub-actions, isn’t a plan, it’s a wish list with a RAG rating attached. Nick Hart, who writes widely on school strategy, puts it plainly: if everything is a priority, nothing is. When every subject leader, every phase leader and every governor sub-committee gets their pet project into the document, headteachers stop being able to act as what he calls the gatekeeper of collective attention, and staff are left trying to make progress on everything at once, which in practice means meaningful progress on very little.

The Education Endowment Foundation’s implementation guidance, Putting Evidence to Work: A School’s Guide to Implementation, names this same pattern directly. Schools under pressure to improve quickly are the ones most likely to fall into it, because everything genuinely does feel urgent. But the EEF’s research into how change actually embeds in schools is consistent: attention and capacity are finite, and initiatives that are launched without enough of either tend to fade rather than fail outright. Nobody announces they’re abandoning the phonics push or the new marking policy. It just quietly stops being what anyone actually does.

Written for the inspector, not the staffroom

The second failure mode is audience. Plans get written in a register nobody in the building would actually use: “to further embed a culture of academic excellence through the strategic deployment of quality-first teaching.” That sentence can survive a governors’ meeting. It cannot survive contact with a Monday morning staff briefing, because nobody can tell from it what they’re supposed to do differently.

This is worth being clear-eyed about, because the assumption behind it doesn’t hold up. Ofsted’s own position is that schools shouldn’t be creating documentation specifically for inspection at all. The school inspection handbook states that inspectors will not expect self-evaluation “other than that which is already part of the school’s business processes,” and schools are only asked to provide a summary of existing self-evaluation, in whatever form they already use it, by 8am on the first day of inspection. Ofsted has been explicit, going back to 2019, that there is no required format for a SEF, and the newer inspection toolkit guidance reiterates that schools are “not expected to generate new evidence or undertake additional work solely for inspection purposes.” A plan written primarily to be inspection-ready is solving a problem that doesn’t actually exist in the way most leaders assume it does. What inspectors are actually trying to assess is whether leaders understand their own school and are taking effective action, not whether the document reads well.

Ian Frost, a headteacher who has written extensively about school development plans, makes the point more bluntly: length is the enemy. A fifteen-page plan gathers dust. His advice is to be brief, map the destination clearly, and resist the same undisciplined pursuit of more that Hart describes from a different angle. Two experienced practitioners arriving at the same warning from different directions is a reasonable signal that it’s a real pattern, not a personal quirk.

No real ownership beyond the head

The third failure is structural. Plans that list the headteacher as the named lead against every single priority are not distributing responsibility, they’re just documenting who’s accountable for everything, which functionally means no one below SLT owns anything specific enough to be held to. When a middle leader can’t point to their line in the plan and say “that one’s mine, and here’s what I’m doing about it,” the plan hasn’t created ownership. It’s created a to-do list for one person with everyone else as an audience.

Prioritise ruthlessly: an evidence-informed approach, not just “pick three things”

Telling a head to “focus on fewer priorities” is true but not actionable on its own. The more useful discipline is a genuine prioritisation process, and the EEF’s guidance offers one worth borrowing rather than reinventing.

Their implementation model runs through four phases: Explore, Prepare, Deliver and Sustain. Most schools spend most of their energy on Deliver, because that’s where the visible activity happens, and treat Explore as a formality that takes an afternoon in July. The EEF’s guidance argues the opposite: the Explore phase is where the real work of prioritisation should happen, and it should draw on more than instinct. Their approach to evidence-informed improvement asks leaders to work through what they actually want to achieve, what the realistic range of solutions is, and what would give an idea the best chance of succeeding in this specific school, before committing resource to it, rather than adopting an approach because it worked somewhere else or because it’s what everyone is talking about at the moment.

In practice, ruthless prioritisation looks less like a brainstorm and more like triage. For each candidate priority, ask three questions honestly. First, what is the actual evidence this is a problem here, not just a general concern? Progress data, work scrutiny, pupil voice, attendance patterns and behaviour logs all count as evidence; a hunch from a single lesson observation doesn’t, on its own. Second, is this something the school can meaningfully move within a year with the staff and time available, or is it actually a three-year change being squeezed into an annual document? Third, and this is the one plans usually skip, what would have to stop happening for this to get proper attention? If the answer is nothing, the priority isn’t really a priority, it’s an addition to an already full plate.

Ian Frost’s advice runs the same direction from a practical angle: before adding targets or accountability structures, get the fundamentals right first, which for him means clarity of direction and genuine simplicity, achieved by cutting rather than adding. A three or four priority plan that staff can hold in their heads will outperform a comprehensive twelve-priority document every time, not because the twelve issues aren’t real, but because a plan that tries to fix everything gives permission to make progress on nothing in particular.

Turning vague priorities into specific, measurable actions

“Improve outcomes in writing” is not a priority a teacher can act on Monday morning. It’s a direction of travel that still needs translating into something concrete enough to check. The gap between the two is where most plans lose their usefulness.

A workable action has three things a vague one doesn’t: it names what will actually change in practice, it says who will do it and by when, and it states what evidence will show whether it worked. “Improve outcomes in writing” becomes something closer to “all KS2 teachers model extended writing using the agreed structure strip in at least two sessions a week, led by the English lead, checked through half-termly book looks, with the target that 70% of Year 5 pupils are working at the expected standard by the spring assessment point.” That’s longer to write but shorter to act on, because there’s nothing left to interpret.

Success criteria deserve particular care. A criterion like “pupils make good progress” can’t be falsified, which means it can’t actually be checked, only asserted. Better success criteria are things you could point to and say clearly whether they happened or not: a percentage, a piece of work, a completed audit, a change observed in a set number of lesson visits. If a success criterion could plausibly be marked “achieved” regardless of what actually happened in classrooms, it isn’t doing its job.

Assigning ownership beyond the head

Distributed ownership only works if it’s genuine, not decorative. Naming a subject leader against an action without giving them the time, budget line or decision-making authority to actually move it is a common way plans quietly fail while looking properly delegated on paper.

The NGA’s joint guidance with ASCL and NAHT on strategic governance, Being Strategic, frames this as a year-round discipline rather than a September exercise. Strategic priorities need to be identified, properly resourced, and monitored consistently through the year, not simply agreed once and then assumed to be progressing. That resourcing question matters more than it usually gets credit for: an action with a named owner but no protected time is an action that will lose out to whatever’s urgent that week, every single time, because urgent always beats important when nobody’s actively defending the important thing’s place in someone’s diary.

Real ownership also means the named owner has genuine input into how the action gets done, not just responsibility for reporting that it happened. If a phase leader is handed a fully specified action with no scope to shape the approach, they’re not an owner, they’re a project manager for someone else’s idea, and that distinction shows up quickly in how much energy they bring to it.

Building in progress-checking that isn’t just an annual review

An improvement plan reviewed once a year, typically in the same July slot it was written, isn’t being monitored, it’s being archived twice. By the time anyone checks whether a priority actually moved, three terms of teaching have already happened without the correction that early warning would have allowed.

Frost’s advice here is to keep the plan a genuinely live document rather than a filed one: reviewed at set points through the year with senior leaders and governors, RAG-rated honestly rather than optimistically, and used as a working reference in the meetings that already happen anyway, line management conversations, team briefings, governor committee agendas, rather than as a separate exercise bolted on top of them. A plan that only gets opened for its own scheduled review meeting has already lost the argument for being useful day to day.

A workable rhythm for most schools is termly at minimum: a short, honest check against each priority’s success criteria, with the explicit option to say an action isn’t working and needs to change rather than just restating the original target and hoping the next term goes better. RAG ratings are only useful if amber and red are allowed to appear without triggering panic. A plan where everything is permanently green either means the targets were set too low to fail, or nobody’s being straight about what’s actually happening, and both are worse than an honest amber.

Involving staff in a way that builds buy-in, not just compliance

There’s a real difference between staff who were told the priorities and staff who helped set them, and it shows up in how much energy people bring to delivering them. Compliance gets you the minimum: the marking happens because it’s checked, the intervention runs because it’s on a timetable. Buy-in gets you the version where a teacher notices something isn’t working and adapts it, because they understand why it matters and feel some ownership over whether it succeeds.

Frost’s practical suggestion is to bring senior and middle leaders into the process early, before the plan has a first draft, asking them in the summer term to identify what they think the real priorities are, rather than presenting a finished document for comment in September. Written first, discussed second is consultation. Discussed first, written second is closer to genuine involvement, and staff can generally tell the difference between the two even when the final document looks similar.

The NGA’s strategic guidance makes a related point about culture: a governing board’s job includes being genuinely aware of how the plan feels to the people delivering it, not just whether the paperwork is in order. That’s a useful check for headteachers too. If nobody outside SLT could explain, in their own words, what the school’s top priority is this year and why, the plan hasn’t actually landed yet, whatever the document says.

Communication also needs to be pitched differently for different audiences, and that’s not a failure of consistency, it’s just accurate targeting. Senior leaders need to understand the full plan and how the pieces connect. Most other staff need something much shorter: what changed, what it means for them specifically, and how they’ll know if it’s going well. A twenty-page strategic document and a one-page summary of the same three priorities can both be accurate. Only one of them gets read by most of the staffroom.

How the SEF and the improvement plan should actually relate to each other

A lot of confusion here comes from treating the SEF as a separate compliance document that has to be reconciled with the improvement plan afterwards, when the more useful relationship runs the other way. Self-evaluation, done honestly, is what identifies the priorities. The improvement plan is what you’re doing about the priorities self-evaluation surfaced. If those two documents don’t clearly connect, one of them is probably decorative.

This matters because Ofsted has deliberately stepped back from prescribing either document. There is no required SEF template, no expected format, and no separate exercise schools are meant to run purely for inspection purposes. What inspectors are actually trying to establish is whether leaders understand their school accurately and are taking effective, well-judged action in response, which is precisely the self-evaluation-to-plan relationship a genuinely useful improvement plan already embodies. A SEF that’s kept honestly up to date as a working document, alongside a plan that visibly responds to what it says, does more for an inspection conversation than either a polished standalone SEF or an ambitious plan with no clear evidence trail behind it.

The practical implication is to stop treating self-evaluation as an annual write-up and start treating it as the ongoing evidence base the plan draws from: the data reviews, work scrutiny, pupil voice and governor monitoring visits that happen through the year, kept somewhere accessible enough that when it’s time to review or refresh the plan, the honest answer to “how do we know this is still the right priority” is already sitting there, rather than needing to be reconstructed from memory in a rush before the next planning cycle.

What this actually looks like in September

A plan that gets used tends to share a few unglamorous features. It has three or four priorities, not ten. Each one has an owner with real time and authority attached, not just a name in a column. Every action has a success criterion specific enough that two different people would agree on whether it happened. It gets checked properly at least termly, with permission to say something isn’t working. It was shaped by the people who’ll deliver it, not just announced to them. And it was written as a working tool for the staffroom, not as evidence for a visitor who, by Ofsted’s own account, was never expecting a polished document in the first place.

None of that requires more hours than writing the fifteen-page version. It mostly requires deciding, honestly and early, what’s not going to be a priority this year, and having the discipline to protect that decision through a full three terms of everything else competing for attention.