The Full Desk: What Cognitive Load Does to Performance, and What Managers Can Do About It.

A person can look completely fine and still be operating at half their capacity.

They haven't become less capable. They haven't stopped caring about the job. In most cases, they don't even know it's happening to them.

This isn't about ordinary work pressure — a tight deadline, a difficult customer, a full inbox. Most people handle that and switch off when they go home. It's about what's sitting underneath it that most managers never see: grief, caregiving, divorce, untreated depression, PTSD, financial trouble — and, just as often, something the workplace itself caused directly. Sustained bullying from a manager or colleague is linked to depression, anxiety, and PTSD symptoms specifically.^[1] Being repeatedly pushed to act against your own values — moral injury, first studied in soldiers, now documented in ordinary customer-facing jobs^[2] — and straightforward job insecurity, sharpened by the live reality of AI-driven restructuring, both do the same thing: they occupy real mental capacity before the workday even starts.

Any one of these consumes capacity on its own. The problem is what happens when ordinary work demands land on top of a mind already carrying one of them. That's cognitive load: the total demand on a person's limited mental capacity at any given moment, from everything they're carrying — not just what's on their to-do list.

The desk

Think of working memory — the part of the brain that holds what you're actively thinking about — as a small desk. There's only so much room on it before things start falling off the edges.

Serious, sustained stress — grief, caring for someone who's dying, untreated depression, PTSD, a genuine crisis — takes up a large, fixed amount of that desk before the workday even starts. Ordinary work demands get piled on top of that. For someone not carrying anything serious, those demands fit fine. For someone already carrying something heavy, the same ordinary demands are what finally push things off the edge.

Once the desk is full, something gives. Not because the person is failing. Because there's nowhere left to put the next thing.

What happens once the desk is full

Short-term thinking is affected quickly — in two separate waves, not one. Research on acute stress and working memory shows impairment doesn't arrive as a single, steady decline. It comes in two windows: an immediate dip in the first few minutes, driven by a fast-acting stress chemical (noradrenaline), followed by a second, more pronounced dip around 25 to 50 minutes later, once cortisol — the slower stress hormone — peaks in the brain.^[3] In practical terms: someone who's just been hit with a stressful demand may actually seem fine for a few minutes, then get noticeably worse a little while later, once the second wave lands.

Under stress, thinking shifts from careful to reactive. The brain's arbitration between "think it through" and "react on instinct" tips toward instinct — deliberate, prefrontal-cortex-driven reasoning gives ground to faster, more emotion-driven responses.^[4] This is well established. What's less certain is exactly how much worse this gets specifically when time pressure is added on top — the direct evidence for that particular combination is thinner than commonly assumed, so it's fair to say it's likely, not proven.

Attention locks onto the wrong thing — but only when there's a real threat, not just excitement or busyness. Controlled research using eye-tracking found that attention narrows sharply around threat-relevant content specifically — not around arousal in general.^[5] A tight deadline that feels threatening will narrow focus. A busy but non-threatening day may not.

It's rarely one bad day. It's what's already stacked up. Most people picture overload as one dramatic event tipping someone over. In practice, it's usually built up slowly — steady workload, not enough people to share it, no real gap between demanding tasks, and having to look composed on the outside regardless of what's happening internally. That last one is the one that catches managers out.

"But I do my best work under pressure"

There's some truth in this, but it's more specific than the popular version suggests. On simple, well-practised tasks, more pressure keeps improving performance. On complex, unfamiliar tasks — most knowledge work — performance rises with moderate pressure, then drops sharply once pressure goes too far.^[6] That pattern describes a short burst of pressure with a known end point. It doesn't describe pressure that never really switches off, which is the version this article is actually about.

The person who says "I do my best work under pressure" is often the person who hasn't hit their limit yet — performing well for a stretch while quietly running down their reserves. Building a culture around that phrase usually means mistaking a short-term human ability for a permanent way of working.

Why you often don't see it happening

The better someone is at appearing calm and in control, the less visible the load actually becomes — to their manager, their team, and sometimes even to themselves.

This is worth sitting with, because it inverts the usual assumption. The employee who seems the most composed under real pressure isn't necessarily the one handling it best. They may simply be the best at masking it — because whatever they're actually carrying (a dying parent, a divorce, a depression nobody at work knows about, being bullied by someone senior to them) isn't something most people bring into the office. Composure, in that situation, isn't evidence the load isn't there. It's often evidence of exactly how much effort is going into keeping it invisible.

This is especially true when the source of the load is inside the workplace itself. Someone being bullied by a manager, or being pushed to act against their own values, is often the least likely person to report it directly — the source of the harm may be the same person they'd have to report it to. What other people notice instead is quieter and less specific: performance quietly dropping off, someone becoming withdrawn, a person who used to speak up in meetings going quiet.

By the time overload shows up somewhere clearly visible — a missed deadline, a resignation letter, a formal complaint — it's usually been building for a long time.

What you might actually notice

This comes with an important limit, stated plainly: the research does not show that managers can reliably spot cognitive overload just by watching someone. Every marker below has been measured in a lab, under controlled conditions — not proven to be something an untrained manager can correctly read in the wild. Each one can just as easily come from illness, distraction, a system outage, an unrelated personal issue, or simple task difficulty. None of them, alone, means anything for certain.

With that limit clearly stated, here's what the evidence does show, from controlled studies using standardised cognitive-load tasks:^[7]

• Previously routine tasks start taking noticeably longer. In one lab study, task completion time rose by roughly 20% under high cognitive load. Accuracy often didn't drop — people compensated by slowing down rather than making more mistakes.
• Visible physical tension — shoulder hunching, restlessness, fidgeting — measurably increases under cognitive load in controlled studies, though the effect is subtle and looks the same as ordinary physical discomfort or anxiety.
• Self-reported frustration rises sharply under high cognitive load — but only shows up if someone actually tells you, which many people won't do unprompted.
• Frequent task-switching and constant interruption is a cause of overload, not a symptom of it. Someone constantly jumping between tasks is a sign their workload design needs attention — not necessarily a sign that person is personally struggling.

When the underlying cause is something the person is actively keeping hidden — bullying, moral injury, a personal crisis — the signs a manager or colleague is more likely to actually see are quieter: a person who used to contribute in meetings going quiet, someone previously reliable starting to miss details with no obvious cause, withdrawal from colleagues they used to be close to. None of these on their own confirms anything. But taken together, and especially as a change from how someone normally operates, they're often the only visible trace of something serious that isn't being said out loud.

What not to do

• Don't diagnose. None of the signs above confirm a specific cause.
• Don't assume laziness or disengagement. A change in performance is a signal to investigate, not a verdict.
• Don't turn a behavioural change into evidence against the person. This list exists to help you reduce someone's load, not to build a case for managing them out.
• Don't demand disclosure. What's actually going on in someone's life isn't owed to you, and asking still leaves you no better placed to help.

It's also worth being precise about where responsibility actually sits. You may not know what's filling someone's desk, and it may have nothing to do with you — a dying parent, a divorce, an illness has nothing to do with how the business is run. But you do control what the organisation puts on top of it, and how you respond when someone's performance changes. That's a narrower, more honest claim than "this is always a workload design problem" — sometimes it isn't. What's consistently true is that the deadline, the workload, and the response are yours to control, regardless of what caused the underlying load.

The practical takeaway: observation alone isn't reliable. A direct conversation about someone's actual workload — hours, task volume, interruptions — will tell you more than watching for behavioural cues ever will.

What managers can actually do

Predictable scheduling — Gap Inc. retail stores. In a two-year randomized controlled trial across matched store locations, Gap Inc. moved part-time retail staff from unpredictable, last-minute scheduling to schedules set one to two weeks in advance, with a guaranteed minimum of 20+ hours a week. The result: turnover fell 20–30% in the stores that made the change, sales per labour hour rose 5–7%, and customer service scores improved.^[8] The mechanism isn't mysterious — not knowing your schedule, your income, or whether you can plan your own life is itself a constant item sitting on the desk. Removing that uncertainty freed up real capacity, and the business results followed.

Schedule control — a Fortune 500 IT division. In an 18-month randomized trial, an IT division gave employees real control over when and where they worked, backed by manager training on supporting the change and setting realistic expectations. Turnover dropped. So did measurable cardiometabolic risk markers — blood pressure, cholesterol — tracked physiologically, not just self-reported.^[9] Job satisfaction and engagement improved too.

Neither study directly measured "cognitive load" with a lab instrument — a fair limitation, and true of almost all real-world organisational evidence in this space. But both are rigorous, randomized, peer-reviewed, and the mechanism connecting the change to the outcome is clearly stated and defensible, not just assumed.

This works best when it's built into how the role runs for everyone by default, not granted individually. Research on coworker reactions to flexible-work policies found that perceived unfairness — not the policy itself — was what predicted lower satisfaction among people not using it.^[10] A case-by-case grant, decided in the moment, invites both resentment and inconsistency. A policy everyone already has access to doesn't have that problem.

Why this matters to the business

That figure is specific to diagnosed depression and shouldn't be stretched further than that. But the theme holds beyond this one figure: most of the cost of overload comes from people who are at work but not really functioning, not from people who are off sick. That's far harder for a manager to spot than an empty chair — there's no line item for it.