Jenny Segal, workplace culture expert, author, qualified actuary and investment professional discusses how DEI training can overlook a crucial aspect of delivery.
As a CIO, I am used to thinking about the risk and return of an investment. And when I look at most UK organisations, the return on DEI investments don’t look great: the gender pay gap is still alive and well, with men continuing to outnumber women significantly in the plum senior roles. Yet women outnumber men in the population as a whole, so if we can’t get this right for the ‘diverse’ majority, what hope is there for a diverse minority? My diagnosis is not that the project has failed, but that the investment has been misdirected. Perhaps it’s time to move it.
Across UK organisations, DEI spend has risen steadily for a decade, manifesting in training programmes, awareness sessions and board commitments, with HR teams building out programmes to deliver against them. But the metrics that should be moving, from women’s progression to minority retention to sentiment data, have barely shifted. What is causing the disconnect?
Training as a token gesture
There are numerous conventional explanations, and all of them feel plausible and ring true. Even before the US-led DEI reversal, some leadership teams were half-hearted, resulting in training as a token gesture, inadequately supported by a convincing business case championing the decision-making and social benefits of a more representative workforce. None of them, in my view, fully explains why the gap between investment and outcome remains so stubborn.
A more useful explanation is structural, that DEI architecture as currently designed, is frequently aimed at the wrong layer. This results in money being spent at the top and the bottom of the organisation, whilst the layer that actually determines daily employee experience receives the least support, the least training and the least attention. That layer is middle management.
Often, the DEI order of events starts with senior leaders attending bespoke programmes, designed to build commitment, sponsor change and frame culture for the rest of the organisation. Employees receive awareness sessions, online modules and the occasional away day. The middle is treated as a relay station, expected to absorb messages from above and transmit them downwards, with little dedicated investment of its own. The problem is that culture does not move that way.
In the 250+ interviews I have conducted with senior professionals across financial services for my On Motivation series, the most consistent finding is this: when people describe their experience of culture at work, it is directly shaped by their line manager. Not the values poster in reception, or the online DEI module they completed last quarter. Their manager. The person who responded, or did not respond, when something happened in a meeting. The person who put their name forward in a closed-door conversation – or did not. The person who treated their bias concern as a real issue – or as background noise. That layer is sitting on the thinnest budget per head in the organisation, with the heaviest operational load, the least protected development time, and the least psychological safety to admit it does not know what to do. We then ask it to be the engine of inclusion. That architecture is wrong.
How bias shows up at work
My work is to address under-investment in management, and the need to do so has sharpened over the years. To see why this matters, let’s consider how bias actually shows up at work.
I find it helpful to map incidents along two axes. The first is intention. Was the behaviour deliberate, or unintended?
The second is degree. Was it mild, or egregious? The four resulting quadrants give very different prescriptions.
For example, intentional egregious behaviour – a deliberate insult, a hostile remark, a targeted exclusion – tend to be escalated, often correctly. HR processes are designed to handle these, escalating to formal complaints and disciplinary records. Conversely, mild and unintended behaviour, such as an assumption, a casual aside, a moment of being overlooked, tend to be absorbed silently. Few employees raise them and they sit below the threshold for HR action.
Bias should be treated as a skills issue, not a values issue.”
The interesting space, and the one where culture is formed, is the territory in between. The intended but mild. The egregious, but unintended. These behaviours are not severe enough to escalate, yet not mild enough to ignore. The ones where someone needs to make a judgement call about whether and how to intervene. That judgement call almost always sits with the line manager. Whether a comment in a meeting gets named or absorbed is a manager’s call.
Whether a misattributed idea gets corrected is a manager’s call. Whether a junior woman gets put forward for the project that builds her profile is a manager’s call. Whether the bias she experiences gets framed as a real issue or as a sensitivity is a manager’s call.
These are not values problems: the managers I have interviewed often know what is right and fair and what good looks like. But they cannot always do it: because they do not have the language, or the time, or the skills. Or the explicit criteria for what good performance looks like in this dimension. They have not seen senior leaders model the behaviour and survive it intact. They are not sure whether intervening will make them look weak, political, or both.
HR has spent a decade telling them what to value, but has spent far less time equipping them with the specific capabilities required to act on it.
There is no shortage of useful interventions. The architectural shift, in my view, is to stop treating middle managers as the audience for DEI training and start treating them as its delivery mechanism. And that changes what you invest in.
What to look out for:
Bias should be treated as a skills issue, not a values issue. Most managers know what they should not do, but lack practise for the specific actions that disrupt bias in the moment: redirecting a misattributed question, naming a pattern without making it personal, raising someone’s contribution in a forum where they were not credited. These can be taught, rehearsed and assessed. Most DEI training does not do this.
Most managers have no framework for great, fair sponsorship, and no expectation that they will be held to account for it.”
Sponsorship should be a measurable management responsibility, not an informal cultural ask. When who progresses in any organisation is largely linked to who has been sponsored behind closed doors, it can be hard to get transparency on this. Most managers have no framework for great, fair sponsorship, and no expectation that they will be held to account for it.
Feedback structures should be explicit and criteria based. The single most reliable intervention against both imposter syndrome and bias in performance review is making the criteria visible, comparable and consistently applied. Administratively boring, it can be culturally transformational.
Managerial uncertainty should be normalised. The phrase ‘I don’t know how to handle that’ should be a sign of capability, not of weakness. In most organisations, it is seen as the latter.
The organisations that move on this will see measurable results. Retention will improve, because employees leave their manager and not their company, and a better-equipped manager produces better-treated employees. Progression data will start to shift, because the gatekeeping decisions made in closed-door conversations will be more consistent. Sentiment data will follow.
These are not soft outcomes. They are commercial ones. The cost of replacing a senior employee, the cost of a damaged team, the cost of a public DEI failure, are all measurable. The cost of getting middle management right is measurable too, and a lot, lot lower.
