Challenging the Status Quo of Performance Assessments
Rethinking Performance Assessments: Why the Ritual Survives Its Own Failure
Few practices are as deeply embedded in organisational orthodoxy as the formal performance assessment.
It remains the backbone of decisions on pay, bonuses, promotions, improvement plans, and feedback; and it endures despite decades of evidence that it does none of those jobs well.
That persistence is the interesting part, and it’s the part most critiques miss.
More than thirty years ago, W. Edwards Deming warned about the corrosive effects of performance assessment. Lived experience inside organisations and a steady stream of scholarship have echoed him ever since. And yet, past incremental tweaks to the annual compliance interview, the ritual survives – usually defended on the grounds that there’s no viable alternative. But a practice that fails on its own terms and survives anyway isn’t surviving because it works.
It’s surviving because it does something else, and it’s worth naming what.
Why the ritual persists
The formal assessment endures for the same reason other hollow managerial tools endure: it lets an organisation perform accountability without building the structure that would actually produce it. A scheduled review, a rating, a documented conversation: These look like rigour, and they let leadership point to a process while the architecture that actually shapes performance stays untouched. The assessment isn’t the mechanism that drives performance. It’s the artefact an organisation reaches for when it wants the appearance of managing performance without the harder work of designing a system that produces it.
The substitution problem at the centre of it
There’s a deeper reason assessment fails, and it explains why even reformed, continuous, well-intentioned versions struggle. The moment you assess someone against a measure – a KPI, a target, a rating – that measure stops being a description of performance and becomes the thing people optimise. The proxy replaces the purpose.
Attach a customer-satisfaction score to the review, and you get engineered surveys; Attach a sales number, and you get pipeline theatre;
Attach any metric to a consequential rating, and people manage the metric, including the parts of it that have drifted away from the actual outcome it was meant to track.
This is why linking assessment to pay is so corrosive.
It isn’t that pay-for-performance schemes fail to motivate. The argument doesn’t need that contested claim. It’s that the instant a rating determines a bonus, the rating stops measuring performance and starts measuring how well someone can manage their rating. Honest self-appraisal becomes irrational; defensiveness becomes rational. You have converted an instrument of judgement into a negotiation. Deming’s core point holds: the assessment should describe performance, and the moment it’s yoked to compensation, it can no longer do even that.
What performance actually is
Part of why assessment goes wrong is that organisations never define the thing they’re assessing. Reduced to a KPI tally, “performance” becomes compliance with assigned numbers. Dressed up with psycho-social evaluation, it conflates personality with contribution, measuring what someone is like rather than what they produce, which belongs to development, not assessment.
Real performance is a matter of competence, and competence is more specific than either. It’s the ability to do the actual work well: to keep a process reliable and stable rather than firefighting it, to reduce unnecessary complexity so the work becomes more dependable, and to improve and innovate at the point where the work meets its stakeholders. Judged against that, 360-degree feedback and KPI tallies don’t just fall short — they distort the signal, rewarding visibility and metric-management over the quieter competence that actually produces value.
What replaces the ritual
If the formal assessment is an artefact that performs accountability rather than producing it, the answer isn’t a better-designed ritual – a smarter form, a more frequent cycle, a cleaner rating scale.
It’s to build evaluation into the architecture of the work itself, so it stops being a scheduled event and becomes a property of how the work runs.
That means assessment that is continuous and contextual, happening in real time as work is done, or when a specific incident warrants inquiry, rather than on a calendar.
It means judgement of output quality and genuine stakeholder value, not compliance with a target.
And it means the evaluation is owned by the people closest to the work: Process owners and the stakeholders the work actually serves, rather than performed by a manager once a year as a standalone duty.
The manager’s real role
This changes what a manager is for. In this model, the manager doesn’t conduct the assessment; the manager builds and maintains the architecture that makes continuous evaluation possible, the tools, the information flows, the clarity about what good work is, and the decision rights that let the people doing the work judge and improve it. The shift is from performing the review to designing the system that makes the review unnecessary as a separate event.
The point
The case against the formal assessment isn’t that it’s an outdated tool waiting for a better one. It’s that the ritual survives precisely because it lets organisations look like they’re managing performance while the structure that actually produces performance stays unbuilt, and because the moment it’s tied to a rating or a bonus, it stops measuring the thing it claims to measure at all.
Retire the ritual.
But don’t replace it with a cleverer ritual.
Replace it with an architecture where evaluation is embedded in the work, owned by the people closest to it, and never yoked to the pay decision that would corrupt it.
Everything short of that is the same performance of rigour the annual review always was.