Part 4 of Governance Beyond Compliance — building accountable leadership, resilient organisations and healthy cultures. First published on LinkedIn, 10 June 2026.
Why staff attrition is a governance metric, not a management statistic
Key takeaways
- Staff attrition is a governance metric, not a management statistic.
- People data that reaches the board through management can be filtered or curated; psychological distress is itself a governance signal.
- Independent board oversight of people data needs deliberate safeguards — the article closes with a practical checklist for board chairs.
There is a number sitting in your organisation right now that tells you more about its governance health than your financial statements, your compliance calendar, or your board papers combined.
It is your staff turnover rate.
Not the raw number – though that matters. The pattern. The concentration. The timing. Who left, when they left, under what circumstances, and whether anyone at board level has ever asked those questions independently of the management line that may have contributed to the answers.
In most small NFPs, the answer to that last question is no.
Staff turnover is treated as a management metric. It is reported to the board – if it is reported at all – as a number, stripped of context, framed by management, and filed alongside the operational updates that boards receive and note and move past. It is rarely examined as what it actually is: a governance signal of the first order, carrying information about culture, leadership, compliance risk, and organisational health that no other metric provides as directly or as honestly.
People leave organisations for reasons. Those reasons are data. And a board that is not systematically collecting, independently reviewing, and honestly interrogating that data is not governing the organisation’s most significant asset – its people – and is not protecting itself from the regulatory exposure that follows when people management fails.
This article is about why staff turnover is a governance metric, what the data is actually telling you, and what a board needs to do to hear it clearly.
Why turnover is a governance metric
The instinct to treat staff turnover as a management matter is understandable. Hiring, managing, and retaining staff are operational functions. They sit within the CEO’s mandate. The board does not recruit staff, manage their performance, or decide when their employment ends. These are management decisions, appropriately made by management.
But the pattern of those decisions – the aggregate effect of how the organisation hires, manages, and loses staff over time – is a governance matter. And the distinction is important.
A single staff departure is a management event. A pattern of departures – concentrated in a particular team, a particular tenure, a particular leadership style – is a governance signal. It is telling the board something about the conditions the organisation has created for the people who work within it. And those conditions are the board’s responsibility to understand and, where necessary, to act on.
This is not a theoretical position. It is grounded in the legal and regulatory framework within which Australian NFPs operate.
Safe Work Australia requires organisations to identify and manage psychosocial hazards – including poor organisational justice, lack of role clarity, and poor support – as workplace safety obligations. High staff turnover is both a symptom and a consequence of unmanaged psychosocial risk. A board that is unaware of its organisation’s turnover pattern cannot demonstrate that it is meeting its psychosocial risk obligations.
Fair Work Australia’s jurisdiction extends to the manner in which employment is ended – including whether terminations and redundancies are genuine, whether performance management processes are procedurally fair, and whether the pattern of exits suggests systemic conduct that may constitute adverse action or constructive dismissal. A board that has no independent visibility into how employment ends in its organisation is exposed to regulatory findings it never saw coming.
The ACNC’s responsible persons obligations require board members to act with reasonable care and diligence in the charity’s best interests. A board that ignores persistent signals of cultural and people management failure is not acting with reasonable care – regardless of whether those signals were presented to it in a format it found easy to dismiss.
Staff turnover, in short, is a compliance matter wearing a management costume. And a board that treats it as a management matter alone is accepting risk it does not have the authority to ignore.
What the pattern is telling you
Not all turnover is equal. A board that monitors turnover needs to understand what different patterns mean – because the governance implications vary significantly depending on the nature of the attrition.
Voluntary resignation concentrated in a single tenure. When multiple staff members resign within a relatively short period – particularly when that period corresponds to the appointment of a new CEO or a significant change in leadership style – the concentration is the signal. Individual resignations have individual explanations. Concentrated resignations have a common cause. The board’s job is to find it – not to accept the management explanation at face value, but to seek independent corroboration through exit data, staff surveys, and direct board access to departing staff where appropriate.
Performance management as a pattern rather than a tool. Performance management is a legitimate and necessary management function. But when the frequency of performance management processes increases significantly under a particular leader, or when performance management disproportionately affects staff who have raised concerns or who are perceived as challenging to the current leadership, it ceases to be a management tool and becomes a conduct risk. A board that has no visibility into the frequency, distribution, and outcomes of performance management processes in its organisation cannot distinguish between legitimate performance management and its weaponisation.
Redundancies that follow a pattern. Restructures and redundancies are legitimate organisational responses to genuine operational need. But a redundancy that eliminates the role of a staff member who has recently raised a complaint, escalated a concern, or challenged a management decision warrants careful board scrutiny – not because it is necessarily improper, but because the confluence of timing and circumstance creates a risk that the board has an obligation to examine. The question is not whether the restructure is operationally justified. The question is whether the operational justification is the complete explanation.
Exit interview data that never reaches the board. Exit interviews – when conducted honestly, analysed systematically, and reported independently – are among the most valuable intelligence sources available to a board. They provide direct access to the reasons people leave, the conditions they experienced, and the patterns that individual management explanations may have obscured. In most small NFPs, exit interview data either does not exist, exists but is not analysed, or is analysed by management and reported to the board in a form that management has shaped. A board that does not receive exit interview data independently – or that has never asked why it doesn’t – is missing the most honest feedback its organisation produces.
The filtered information problem – specific to people data
In the previous articles in this series, I have discussed the general vulnerability of boards in small NFPs to filtered information – the tendency to rely on the management line as the primary source of intelligence about the organisation’s health.
People data is particularly vulnerable to this filtering – for reasons that are both structural and psychological.
Structurally, staff information is held by the HR function, which in a small NFP typically reports to the CEO. Exit interview data, performance management records, complaint histories, and turnover analyses are all generated within the management line and travel upward through it. By the time people data reaches the board, it has passed through the filters of the people who generated it – and who have the greatest interest in the interpretation it receives.
Psychologically, people data carries emotional weight that financial data does not. A board paper showing a budget variance is uncomfortable. A board paper showing that six staff have left in eight months, that three of them raised concerns before they departed, and that two of them are now alleging constructive dismissal is devastating. The instinct to manage the presentation of that information – to frame it, soften it, contextualise it – is entirely human. It is also a governance failure when it results in the board receiving a picture of its people situation that is materially incomplete.
The remedy is independent data access. Not distrust of management – but structural independence of the information the board uses to govern its people obligations. This means exit interview data reported to the board directly, not through the CEO. Staff survey results provided to the board in full, not summarised by management. Turnover rates and patterns included as standing items in board papers, not reported selectively when management judges them relevant.
Psychological distress as a governance signal
Staff turnover is the most visible indicator of people management failure. But it is a lagging indicator – it tells you what has already happened, not what is currently happening or what is about to happen.
The leading indicators are subtler and require active board attention to detect.
Psychological distress among staff is one of the most significant. When staff are operating under conditions of sustained uncertainty, role ambiguity, poor organisational justice, or inadequate support, the psychological cost accumulates before it becomes visible in turnover data. Staff become disengaged. Sick leave increases. Presenteeism – the phenomenon of staff attending work but functioning at significantly reduced capacity – becomes prevalent. The quality of service delivery declines. Relationships between staff members become strained as the stress of the environment distributes itself across the team.
In a small NFP, where staff are often deeply committed to the mission and reluctant to leave despite difficult conditions, psychological distress can reach significant levels before turnover data begins to reflect it. Staff stay because they believe in the cause – while the conditions they are working under erode their wellbeing, their performance, and ultimately their capacity to serve the people the organisation exists to help.
A board that waits for turnover data to signal a people management problem has already waited too long. The governance question is not only: why are people leaving? It is: what are the conditions being experienced by the people who are staying?
Psychosocial risk assessment – systematic, independent, and reported to the board rather than managed within the management line – is the instrument that answers that question before the turnover data does.
What independent board oversight of people data looks like
Building independent board oversight of people data does not require a large investment of time or resources. It requires three structural decisions.
First – make turnover a standing board agenda item. Not a discretionary item that management includes when it judges the data relevant. A standing item – appearing in every board paper, in a standard format, showing the turnover rate for the period, the reasons recorded for each departure, the tenure of departing staff, and any patterns the data reveals. Management provides the data. The board reviews it independently and asks its own questions.
Second – establish an independent exit interview process. Exit interviews conducted by management – or by HR reporting to the CEO – are not independent. The departing staff member knows who controls the information and adjusts their candour accordingly. An independent exit process – conducted by the board chair, a designated board member, or an external party – provides the board with unfiltered access to the reasons people leave. In a small NFP this need not be elaborate. A simple, confidential conversation with departing staff, conducted independently of the management line, is sufficient to reveal patterns that curated management reporting will not.
Third – commission periodic independent staff surveys. Not surveys designed and administered by management – but independent surveys, commissioned by the board, with results reported directly to the board rather than through the CEO. A simple annual survey covering psychological safety, role clarity, management fairness, and organisational culture provides the board with a leading indicator of people management health that turnover data alone cannot supply.
None of these three mechanisms is complicated. None is expensive. All of them are within the reach of the smallest NFP board. And all of them will tell the board things about its organisation that it currently does not know – and that the people it exists to serve need it to know.
The curation problem – and how to safeguard against it
Establishing the three mechanisms described above is necessary. It is not sufficient.
The most sophisticated governance failure I have observed in people data management is not the absence of oversight mechanisms. It is the curation of those mechanisms – the quiet management of independence itself, so that the board believes it is receiving unfiltered information while the filtration continues uninterrupted.
This curation can operate at every stage of the data collection and reporting process.
The survey provider is selected by management – with a preference, conscious or otherwise, for instruments that produce reassuring results or that frame questions in ways that reduce the likelihood of critical responses.
The survey is administered by management – with the implicit or explicit message that responses are not truly confidential, or that participation is observed, or that the purpose of the survey is affirmation rather than honest assessment.
The raw data is received by management – and summarised, contextualised, or selectively presented before it reaches the board. Themes that are uncomfortable are attributed to isolated individuals or external factors. Patterns that implicate leadership are dissolved into aggregates that obscure their concentration.
Exit interview data is collected by HR – which reports to the CEO – and reaches the board as a management narrative about why people left, rather than as the direct accounts of the people who left.
Turnover data is reported as a raw number – without the contextual detail of who left, when, under what circumstances, and whether the pattern warrants independent inquiry.
In each case, the mechanism exists. The independence does not. And a board that has not specifically designed against this curation risk will not know the difference – because the curated version of independence looks, from the board’s perspective, exactly like the real thing.
The safeguards – designing for genuine independence
Genuine independence of people data requires the board to control the critical variables in the data collection and reporting chain – not management.
The survey provider must be selected by the board – not management. The board chair or a designated board member – not the CEO or HR – identifies, engages, and instructs the external provider. The provider’s terms of reference come from the board. Their reporting obligation runs to the board. Management is informed that the survey is occurring but does not control its design, administration, or output.
In a small NFP where cost is a genuine constraint, this does not require an expensive engagement. A simple, well-designed instrument administered by an independent third party – a governance consultant, an external HR advisor, or a board member from a different organisation acting in an independent capacity – is sufficient. The independence of the relationship matters more than the sophistication of the instrument.
The survey must be genuinely confidential – and demonstrably so. Staff who do not trust the confidentiality of a survey will not respond honestly to it. Demonstrable confidentiality requires that staff can see that the survey provider is independent of management, that individual responses cannot be identified, and that results go directly to the board rather than through the CEO.
This means communicating to staff – directly, from the board chair rather than from management – that the survey is a board initiative, that responses are confidential to the independent provider, and that the board will receive the results without management intermediation. That communication, coming from the chair rather than the CEO, is itself a signal to staff that the mechanism is genuine.
Exit interviews must be conducted independently – and staff must know this before they resign. An exit interview conducted by HR after a resignation has been tendered is not independent – and the departing staff member knows it. The information they provide will be shaped by what they believe will happen to it, who will see it, and what the consequences of candour might be for references, entitlements, or ongoing relationships.
Genuine independence requires that departing staff have access to a board-level exit conversation – separate from and in addition to any HR process – that is explicitly confidential to the board. This conversation should be offered proactively, conducted by the board chair or a designated board member who is not the chair, and framed as the board seeking to understand the organisation’s health rather than to assess the departing staff member’s performance.
In a small NFP, this is a fifteen to thirty minute conversation. Its value to the board’s understanding of its organisation is disproportionate to its cost.
Turnover data must be reported in disaggregated form – not as a summary statistic. A board that receives turnover as a percentage has received a number. A board that receives a list of every departure in the period – role, tenure, stated reason, circumstances of exit, and whether any concerns were raised during employment – has received intelligence. The difference between these two forms of reporting is the difference between a board that is watching its people situation and a board that has been shown a number and told not to worry.
The disaggregated report should be a standing board agenda item, prepared in a standard format, and provided to the board before the meeting rather than tabled at it – so that board members can review the detail independently and arrive at the meeting with their own questions rather than being led through a management presentation of the data.
The board must have access to independent people expertise – with one important qualification. A well-governed small NFP should have access to independent HR or employment law expertise that the board can consult directly when people data raises questions that management is not positioned to answer honestly. This does not mean the board employs its own HR function. It means the board has a known, accessible, independent source of advice on people and employment matters – a relationship that exists outside the management line and that board members know they can use without the CEO’s knowledge or approval when the circumstances require it.
The important qualification is this: an independent expert is only as independent as the relationship with the paymaster permits. An expert engaged and paid by management is not independent of management. An expert engaged and paid by the board is more independent – but not unconditionally so. True independence requires not only independence of appointment but independence of obligation: a documented commitment to report findings honestly regardless of what those findings reveal about the people who engaged them. And it requires a board culture that can receive honest findings rather than managing them. An expert who delivers uncomfortable findings to a board that then ignores, minimises, or suppresses them has not produced independent governance. They have produced independent advice that governance then failed to act on. The structural mechanism and the cultural resolve are both required.
When the data tells a story the board does not want to hear
The hardest moment in people governance is not collecting the data. It is acting on it when the story it tells is uncomfortable.
A board that commissions an independent staff survey and receives results showing widespread psychological distress, low trust in leadership, and significant role ambiguity has two options. It can investigate independently, name what the data is telling it, and take the governance action the findings require. Or it can accept a management explanation of the findings and defer the action the findings demand.
The first option is difficult. The second option is a governance failure – and one with consequences.
Staff who have responded honestly to a survey in the hope that the board will act, and who then observe the board accepting a management explanation and moving on, do not respond honestly to the next survey. They leave. Or they stay and stop speaking. Either outcome serves no one – not the staff, not the organisation, and not the people the organisation exists to serve.
The board’s obligation is not to make governance comfortable. It is to make it real. Real governance acts on what it finds. It asks the questions that management has not answered. It seeks the corroboration that management explanation alone cannot provide. And it accepts that the discomfort of acting on difficult data is a far smaller cost than the consequences of not acting on it.
Those consequences – regulatory investigation, Fair Work proceedings, WorkSafe intervention, reputational damage, mission disruption – are borne by the organisation and the people it serves. The board that chose comfort over governance does not bear them alone. But it bears responsibility for them. And that responsibility, once the data was available and the governance action was not taken, cannot be explained away.
The practical checklist that follows offers a set of prompts for board chairs assessing their people governance. But beneath every item on that list, one question applies – to every mechanism, every report, every piece of data the board receives about its people:
At each stage of the data collection and reporting chain – design, administration, receipt, analysis, and presentation – who controlled this step?
If the answer at any stage is management – or an entity that reports to management – the independence at that stage is compromised. The board may still receive useful information. But it should not assume it is receiving unfiltered information.
That question – asked consistently, applied rigorously, and answered honestly – is the difference between a board that believes it is governing its people obligations and a board that actually is.
A practical checklist for board chairs
- Turnover rate not reported as a standing board agenda item
- Turnover data reported as a summary statistic rather than a disaggregated record of individual departures
- Exit interview data absent from board reporting – or summarised by management rather than provided in full
- Exit interviews conducted by HR without a separate board-level exit conversation offered to departing staff
- Staff survey results that are consistently positive – particularly in organisations experiencing visible operational difficulty
- Survey provider selected by management rather than the board
- Survey results reaching the board through the CEO rather than directly from the independent provider
- Departures concentrated in the tenure of a particular leader – with explanations that consistently attribute responsibility to the departing staff rather than the organisational conditions
- Performance management frequency that has increased significantly without a corresponding board-level explanation
- Redundancies that follow complaints, concerns, or challenges to leadership – without independent board scrutiny of the confluence of timing
- Board members receiving informal accounts from staff that differ materially from the formal management reporting
- No independent people or employment law expertise accessible directly to the board outside the management line – or independent expertise whose terms of engagement have not been reviewed by the board
- A board culture where asking questions about staff welfare feels intrusive rather than appropriate
- Each of these signals is a prompt for independent inquiry. Not for accusation. Not for crisis. For the calm, persistent, evidence-based questioning that good governance requires – and that the people who work in your organisation are depending on you to provide.
They cannot depend on management to raise these questions on their behalf. They are depending on the board.
That is what the board is for.
Dennis Chew is a Corporate Governance specialist, Company Secretary, and Director of Operations professional with forty years of experience across NFP, faith-based, educational, and corporate organisations in Australia, Malaysia, New Zealand, and Hong Kong. He writes on governance, institutional resilience, and people development.
He is open to conversations about Director of Operations, Non-Executive Director, and governance advisory engagements in Melbourne.
Previous: Part 3 — Why Good Governance Protects Everyone – Including the CEO
Next: Part 5 — What Information Should Your Board Be Receiving and From Whom
All articles in Governance Beyond Compliance
Working through a governance challenge? Dennis is available for board and non-executive director roles, interim executive engagements and advisory conversations. Get in touch or see Governance & Leadership.