Bias In Boxes:

What the Nine Box Talent Grid Misses About Neurodivergent Employees

Many of us developed our way into our careers with models and frameworks. Handy tools designed to simplify information and make it easy and accessible, and the nine box talent grid is no exception. It has become part of the furniture of talent management; a clean and apparently objective way of mapping where people are and where they are headed. 

The framework is presented as a tool for having better conversations, not making final judgements. In practice, though, the boxes shape decisions about who gets developed, who gets stretched, who gets managed out, and who gets quietly left where they are, and those decisions have real consequences for real people.

But when did we last check just how objective the model is? If our model is steeped in bias, how can we possibly expect it to support objectivity?  

What the grid treats as fact 

Every framework rests on assumptions and the nine box grid has several that rarely make it into the thought leadership posts. 

Firstly, it treats performance as observable and consistent. That sounds reasonable until you consider that someone whose best work happens at 11pm, or who works in concentrated bursts rather than a steady daily flow, can look unreliable to a framework that only captures what happens during visible working hours and formal check-ins. Work that falls outside of that, or follows a different rhythm, is easy to miss. 

Next, it treats the effort behind the output as irrelevant. Two people can produce the same result. One of them walked to work; the other ran a marathon in unsuitable shoes and has been doing so every day for two years. The grid records the same output. It has no mechanism for seeing what it cost. 

It also treats the person doing the rating as a neutral observer. Ratings are only ever as good as the relationship between the manager and the person being assessed. Where that relationship has friction the manager has not examined, where a manager has not had enough time or opportunity to understand how someone actually works, or where the manager's picture of what capable looks like is narrower than they realise, the rating reflects all of that. Not the person. 

It treats previous ratings as reliable data, carrying them forward into new conversations as though the context that produced them no longer matters. For someone who had a difficult period, an unsupportive manager, or a role that did not fit, those ratings can follow them for years. 

And it treats the starting conditions as equal, but they rarely are. 

The overhead nobody is measuring

Performance and potential are the headlines of this model, but friction is the unmentioned thread running through the whole thing. 

Some neurodivergent people are working in environments that generate unnecessary friction, to name just a few: 

• Unclear communication that takes real effort to decode 
  

• Open plan spaces that cost more cognitively than they appear to 
  

• Social dynamics in meetings that require active, conscious processing rather than instinct 
  

• Instructions that most colleagues absorb without question but that leave others with a list of unanswered questions they are not sure how to raise 

  
  

That friction is not a measure of capability, but a measure of environment design. The person is entirely capable of the job. The route the organisation expects them to take to get there is creating overhead that their colleagues may not experience at all. Address the friction, and you often get better output. Leave it in place, and the grid records reduced output and files it under performance. 

This is also where disclosure matters, and often falls down. Some people have been clear about what they need and nothing changed. So they adapted. 

From the outside, that can be read as inconsistency, uneven delivery, or reluctance. Much of the time, it is someone working around friction that was never acknowledged or addressed. 

The grid cannot see burnout coming 

The model has no axis for sustainability, and instead measures a snapshot in time. It certainly does not account for the fluctuation of burnout. 

By the time burnout appears in someone's performance data, it has usually been building for a long time. The person who delivered consistently for two or three years and then seemed to fall away did not suddenly change. The building pressure of overextending, masking and adapting accumulated, and then the proverbial camel's back gave way. The nine boxes are not shaped in a way that gives space to understand this. Instead we watch someone fall from Current Star to Inconsistent, from Key Player to Average, or all the way to Risk, and treat it as a performance problem rather than what it actually is: a system failure that had been signalling for months. 

The grid records impressive output but it has no way of telling you whether that output is sustainable, or what the trajectory looks like from here. 

Potential is being measured by how it is performed 

Across every box on the grid, potential tends to go to people who can make it visible in the expected ways. 

We exist in a world where bias favours extroversion. Someone who speaks in meetings with confidence and animation, holds eye contact comfortably, networks in the ways the organisation rewards, and makes their ambition legible to the people above them is systematically favoured in this model. If those things come easily, perhaps even automatically, the grid works in your favour. If they are draining, effortful, or genuinely difficult, the bias within the system consistently places you at a disadvantage. 

For many neurodivergent people, hitting all of those markers simultaneously and consistently, while also actually doing the work, is disproportionately hard. Slow processors are at a disadvantage in real-time conversations where the person who speaks first often shapes the room. Those who find writing, spelling, punctuation and grammar difficult are often regarded as less knowledgeable than they are. Introversion, which tells you nothing about capability, regularly gets scored as limited potential. 

In those cases, the grid is measuring communication style and recording it as capability. This is systemic discrimination, underpinned by neuronormative bias: the assumption that there is a right way for brains to process, communicate, and understand.

Progression as a ladder leaves a lot of people out 

The model assumes progression runs upward, in one direction, on evenly spaced rungs, toward leadership. But what about those who are climbing a cargo net? 

The grid has no language for the specialist who is extraordinary at what they do and has no interest in managing people, whose development has effectively stalled because the only recognised form of growth goes toward a job they do not want. It does not prompt consideration for a lateral move into a better-fitting role as a genuinely positive outcome, and it does not account for depth as a form of progression, or for the person whose career ambitions look nothing like the grid's definition of potential. 

It also positions us to assume we are aiming for everyone to reach the high performance end of the scale, with some models showing this in green and moderate and low performance in amber and red respectively. Consider the box of low potential, moderate performance. Across the many versions of this model available, you will find it labelled as Average Performer, Effective but Stunted, Adequate, Blocker, and even Up or Out. In many organisations, the people sitting here are steady, consistent, reliable, prepared, delivering quietly and creating the conditions for the high flyers to fly. They can be the invisible load-bearing structure of an entire function, and the grid significantly underrepresents their contribution. It has no way of showing you what would happen if they left. 

Then there is the low performance, moderate potential box: The Inconsistent Player, The Predicament, In the Wrong Role. Uneven output is information. It is often telling you that something in the conditions is not working: too much context-switching, ambiguous communication, an energy deficit that does not show up in a performance review. The recommended response to all of that is to set stretching goals, which is almost exactly the wrong thing to do. A more considered approach would be to look at what the barriers to performance actually are, and what can be done about them. 

What the boxes are not asking 

At the low performance, low potential end, the Risk box, you will find recommendations for close management, a PIP, and potential exit. What that box rarely prompts anyone to ask is whether this person is undiagnosed, without the adjustments they need, working from instructions that were never clear enough, in an environment that costs them significantly more than it costs their colleagues. Or whether they were a solid, consistent performer eighteen months ago, and what looks like low performance now is the end stage of a burnout that had been building long before anyone noticed. 

The Possible Gem box sits just above it: high potential, low performance. The grid points toward right role and training. A more useful starting point is asking what support was requested and not provided, and what has been quietly built in the gap. Compensatory strategies look a lot like inconsistency from the outside. They are not the same thing. 

The Current Star box has a different problem. Strong performance, growth potential, recommended for reward and motivation. What the grid cannot see from here is what it is taking to sustain that output. Whether this person is working late to stay on top of something that looks effortless to everyone else. Whether every meeting, every social interaction, every ambiguous brief is costing more than it appears to. The grid records the output. It has no way of knowing whether the person producing it is six months from a wall. 

At the top, Top Talent, we tend to assume arrival here happened on a level playing field. Some of these people had the right manager at the right time. Some had adjustments that worked. Some landed in a role that happened to fit how they think, through luck rather than design. Promoting from this box without asking how people got here embeds exactly the same advantages into the next layer of leadership, and the model never prompts that question. 

And then Trusted Expert. High performer, reached potential. For someone who has worked considerably harder than the people around them to reach consistency in a role that genuinely suits them, that phrase does a lot of damage quietly. Development has stopped. The organisation will use what this person knows and is not curious about what else they might do. That is not recognition. It is assumption with a ribbon on it. 

What this is actually costing 

This is a fairness argument. It is also a business one, and the two are not in competition. 

Every time a performance improvement plan is started before anyone has examined whether the environment, the role design, or the absence of support is the actual problem, the organisation is spending time and money heading in the wrong direction. A formal process takes significant management time to run. It takes HR resource. It creates disruption for everyone around it. And if the underlying issue was always friction in the system rather than the person, the process resolves nothing. It accelerates an exit that should never have happened. 

Those exits are expensive. Recruitment costs. Onboarding time. Lost knowledge. Lost relationships. The accumulated understanding of how things actually work that walks out with the person and takes months to rebuild, if it gets rebuilt at all. 

Meanwhile, the specialist sitting in the middle of the grid, muddling along in a role that does not quite fit, with a manager who does not have the visibility to see what else they could do, is delivering a fraction of what they are capable of. The organisation may never know, because the grid is not designed to surface that question. 

Fixing the environment, getting role design right, making sure adjustments are actually in place: these things rarely only help the one person they were intended for. They tend to improve conditions for everyone around them too. The return on that investment is almost always broader than the problem that prompted it. 

What to do with all of this 

The nine box talent grid is not useless. But it needs to be used with a clear understanding of what it is actually measuring: how someone has performed in the conditions they have been given, in a role that may or may not fit them, with a manager who may or may not understand how they work, from a starting point that may or may not have been comparable to their colleagues. 

Before placing anyone in a box, take a genuinely holistic approach. Have the barriers to performance been examined? Has this person had the adjustments or support they need? Does the person doing the rating have enough of a real relationship with them to assess fairly? Is the role designed in a way that lets them do their best work? And is what is being measured actual capability, or the circumstances that have surrounded it? 

Neuroinclusion is not about adjusting neurodivergent people so they fit a system that was not built for them. It is about looking at the system, and being honest about what it was built to see. 

Hayley Brackley is a neuroinclusion consultant, executive coach and speaker. She works with HR teams, OD leads, and senior leaders on making workplace systems genuinely fair, and with neurodivergent individuals navigating the ones they are already in. Get in touch to find out more.