Teacher Retention: Why Good Staff Actually Leave (and What Keeps Them)

What actually predicts a teacher leaving — the data, not the folklore

The Department for Education’s Working Lives of Teachers and Leaders survey is the closest thing England has to a proper national pulse-check on this, and its 2024 wave (fieldwork January–May 2024, over 10,000 teachers and leaders, run by IFF Research and UCL’s Institute of Education) is uncomfortable reading. Thirty-six per cent of respondents said they were thinking about leaving teaching within the next 12 months, up from 25% two years earlier. Job satisfaction fell sharply too: only 46% said they were satisfied “most of the time,” down from 58% the previous year, and 19% said they were rarely satisfied at all.

Ask people what’s actually driving that, and the answers cluster tightly. NFER’s synthesis of the same working-lives data found that 90% of teachers who were considering leaving in 2023/24 cited high workload as a factor. The DfE survey itself found stress or poor wellbeing cited by 84%, and feeling undervalued by 83%. Pay matters — more on that below — but it isn’t the headline reason teachers give when you actually ask them. Workload, stress and feeling undervalued are.

What’s genuinely useful for a head is that the “why” differs by phase, and that difference tells you where to look first. In primary schools, 78% of those considering leaving cited pressure relating to pupil outcomes or inspection. In secondary schools, 59% pointed to pupil behaviour. In special schools and PRUs, 54% cited a lack of support from superiors and 44% a lack of progression opportunities. That’s not one problem with one fix — it’s three different working environments producing three different pressure points, and a retention strategy that treats them as identical will miss most of what’s actually going on in your building.

There’s a wellbeing gap between leaders and classroom teachers worth noting too. The same survey found 63% of teachers said work had negatively affected their mental health, against 55% of leaders — but leaders were more likely to feel supported (78% agreed their line manager supported their wellbeing, versus 62% of teachers) and more likely to have access to support schemes (79% versus 54%). Heads often assume their staff have roughly the access and support they themselves have. The data says that’s usually not true.

NFER’s most recent Teacher Labour Market in England report adds a cautious note of progress: 26% of teachers described their workload as acceptable in 2024/25, up from just 17% in 2021/22, and vacancy rates eased slightly. But the same report flags pupil behaviour as one of the fastest-growing contributors to workload since the pandemic, and recommends an independent review of behaviour-related workload and pastoral care specifically — which lines up closely with what secondary teachers are telling the DfE survey directly.

Leaving teaching altogether is a different problem from leaving your school

It’s worth being precise about a distinction that gets flattened in most retention conversations: a teacher who moves to the school down the road and a teacher who leaves the profession entirely are not experiencing the same problem, even if both show up as “attrition” in your workforce figures.

Research on teacher movement consistently shows that changing schools is driven mainly by local, immediate working conditions — leadership style, workload distribution, team culture, how supported someone feels day to day. Leaving the profession altogether tends to be driven by broader, more structural dissatisfaction: workload as a whole way of working, pay relative to other graduate careers, and a sense that the job itself has changed for the worse. Research summarised by the Chartered College of Teaching notes that a meaningful proportion of teachers who leave state schools — around half — don’t leave education at all: they move into the independent sector or into teaching-related roles rather than exiting the profession outright. That’s a real signal. It suggests that for a large group of departures, the problem isn’t “teaching” as a vocation, it’s the specific conditions of the specific school they were in.

That distinction should change how you read your own leavers list. Someone moving three miles down the road to a school with a calmer behaviour system and a lighter meeting load is telling you something fixable and local. Someone leaving the sector entirely after fifteen years is telling you something bigger, and while you can’t single-handedly fix national pay policy, you can still influence whether your school was the reason they gave up, or the reason they stayed as long as they did.

What you can actually change, even though pay and Ofsted aren’t up to you

It’s tempting to treat retention as something that lives entirely outside a headteacher’s control — set by pay review bodies and inspection frameworks neither of which you get a vote on. That’s true of the two biggest national levers. It is not true of the four things the data says matter most day to day.

Workload culture

Workload doesn’t mean “the job is hard.” It means the accumulation of marking policies, planning expectations, meeting frequency and administrative asks that a school chooses to run, largely independently of what any external body actually requires. Auditing what you genuinely ask of staff against what you’re actually obliged to provide is usually the single highest-leverage exercise a leadership team can do, and it costs nothing but time. Ready-made, curriculum-aligned planning and marking resources — something like Learnaroo Hub’s teacher library — can shave a modest amount off the weekly grind for some staff, but it’s one small lever among several, not a substitute for genuinely reducing what a school asks of its teachers.

Behaviour support

Given that pupil behaviour is the single most cited reason secondary teachers consider leaving, and one of the fastest-growing drivers of workload nationally, a consistent, visibly-backed whole-school behaviour system is a retention intervention as much as a discipline one. The specific thing staff report valuing isn’t a stricter policy on paper — it’s knowing that when they call for support, someone senior actually shows up. A behaviour policy nobody enforces is worse for retention than having no policy at all, because it signals the gap between what leadership says and what leadership does.

Career development

This mattered most acutely in special schools and PRUs, where 44% of those considering leaving cited a lack of progression opportunities, but it isn’t confined to that phase. Career development doesn’t have to mean a promotion pipeline everyone can’t realistically climb — it can mean protected CPD time that isn’t quietly cancelled under pressure, honest conversations about what a TLR or a subject lead role would actually involve, and making sure good teachers know you’ve noticed them before they start looking elsewhere for someone who will.

Feeling heard

Eighty-three per cent citing “feeling undervalued” as a factor in considering leaving is not a small number, and it’s not fixed by an end-of-term email thanking staff for their hard work. It’s fixed by staff seeing a visible link between something they raised and something that changed. A staff survey that never produces a visible outcome trains people to stop bothering to answer it honestly — which then quietly poisons your ability to read the room at all.

The real cost when a good teacher walks

The cost of losing a teacher rarely appears as one number in a budget, which is part of why it’s easy to underweight against more visible pressures. It sits across several lines: the direct cost of recruitment and advertising, the cost of supply cover while a post sits vacant or a replacement gets up to speed, and the much harder to quantify cost of lost institutional knowledge — the relationships with pupils and parents, the informal understanding of “how this department works,” the mentoring capacity that walks out with an experienced colleague.

On supply cover specifically, one widely cited (if now dated) analysis put annual spending on supply teachers across England’s schools at around £821m, roughly £168 per pupil — and schools sourcing cover through an agency typically pay 20–30% above the base daily rate once agency margins are added. Many schools now set aside dedicated budget lines specifically to absorb this, sometimes running into tens of thousands of pounds a year for a single school, precisely because unplanned absence and vacancy cover has become a predictable, recurring cost rather than an occasional shock.

The harder-to-cost part is what happens to the colleagues who stay. Every vacancy left unfilled, or filled by short-term supply, means someone else absorbing cover lessons, extra planning, or a class that’s now unsettled by a run of different adults in front of it. That’s a direct line back to the workload data above: unfilled vacancies don’t just cost money, they actively worsen the workload pressure that was probably part of why the last person left.

Running an exit interview that actually surfaces the truth

Most schools already do something called an exit interview. Fewer schools do one that produces honest information, because the format most commonly used — a form handed over on someone’s last week, or a conversation with the line manager they may have specific grievances about — is almost engineered to produce polite, non-committal answers.

A few practical changes make a genuine difference:

• Someone other than the direct line manager should conduct it. A teacher leaving partly because of how a department is run will not tell the head of that department the real reason, however well-intentioned that person is.
• Timing matters. A conversation held weeks before departure, once a new job is secured and there’s nothing left to lose by being candid, tends to produce far more useful information than one held on the final day when the instinct is simply to leave on good terms.
• Confidentiality has to be real, not just stated. If staff have any reason to believe specifics will be traced back to them, or fed straight to the person they’re describing, they’ll default to vague, safe answers.
• Ask about specifics, not generalities. “Was there a particular week, meeting or decision that made you start looking?” tends to surface more than “Why are you leaving?”
• Look for the pattern across leavers, not the story in any one interview. One person’s account might be about a personality clash. Three people independently describing the same meeting structure or the same unsupported behaviour incident is a systemic finding, and it needs to be treated as one.
• Close the loop. If exit interviews consistently flag something and nothing visibly changes, current staff will hear about it through the grapevine just as reliably as they’d hear about a positive change — and it will cost you the credibility of every future survey and interview you run.

It’s also worth running the mirror version of this with staff who stay: a short, low-pressure “stay interview” for people a few years into post, asking what’s keeping them and what would tip the balance the other way. It costs almost nothing and often surfaces the same themes as exit interviews, but early enough to act on them.

Retention is a hundred small decisions, not one big one

There is no single lever that fixes teacher retention, and any advice that promises one should be treated with suspicion. National pay and inspection policy sit outside any individual head’s control, and pretending otherwise sets an impossible bar. But the data is consistent, from the DfE’s own national survey through to NFER’s labour market analysis: workload, behaviour support, feeling valued and a credible route for career development are the things staff actually cite, and every one of them is shaped, in practice, by decisions a school makes about its own culture.

The heads who keep good staff longest aren’t necessarily running the calmest schools or the best-resourced ones. They’re the ones where the gap between what leadership says it values and what it actually asks of people day to day is smallest — and where, when someone does eventually walk, the school is honest enough with itself to find out why.