Employment Tribunal Compensation: Why £15,000 Is the Wrong Number
The average compensation award across 129,000+ employment tribunal decisions is £15,000. That number appears in legal guides, HR risk assessments, and settlement negotiations. It's the first figure people reach for when they need a benchmark.
The median is £2,284.
Half of all claimants who win at tribunal receive less than £2,300.
These two numbers describe the same dataset. They tell completely different stories. And when practitioners build quantum advice on the mean instead of the distribution, they are systematically misleading their clients.
What Employment Tribunal Compensation Statistics Actually Show
The gap between mean and median is not a rounding error. It's structural.
A small number of very large awards pull the mean upward. A £500,000 disability discrimination case, a £1 million sex discrimination outlier, these are real decisions in the dataset. Each one shifts the average for thousands of cases. The median, by contrast, is immune to that effect. It tells you what the person in the middle actually received.
| Statistic | Value | |---|---| | Mean compensation (all ET decisions) | £15,000 | | Median compensation (all ET decisions) | £2,284 | | Mean discrimination award (local authority employer) | £52,000 | | Mean discrimination award (NHS employer) | £24,000 | | Vento injury to feelings (lower band) | £1,100 to £11,200 | | Vento injury to feelings (upper band) | £33,700 to £56,200 |
The Vento guidelines give tribunals considerable discretion. The lower band applies where the discrimination is less serious. The upper band covers the most serious cases, including where there has been a significant impact on the claimant's health. The gap between a lower-band award and what the mean compensation figure seems to promise is enormous.
Employer type alone shifts the expected outcome significantly. Local authority discrimination cases average £52,000. NHS cases average £24,000. The same legal framework, the same judges. But employer category changes the number by more than 100%. That context is invisible if you're working from a single summary statistic.
Why Mean-Based Quantum Advice Fails Most Clients
A client told "average payouts are £15,000" has a fundamentally different expectation from one who knows most successful claimants receive under £2,300. The practical consequences run in both directions.
For claimants, mean-based framing creates expectations the evidence rarely supports. A straightforward unfair dismissal claim with no aggravating features, basic award plus compensatory award, is unlikely to produce anything close to £15,000. When settlement offers come in below what the client was led to expect, the gap becomes a crisis of confidence in the advice.
For respondents, the problem runs the other way. HR departments using £15,000 to assess litigation risk are overestimating exposure on most claims while potentially underestimating it on the outlier cases that actually drive the mean.
What useful quantum advice requires is not a single number. It's the distribution: median, interquartile range, 90th percentile, filtered by claim type, sector, discrimination ground, and employer type. That combination tells you what realistic outcomes look like for a specific fact pattern. The mean tells you almost nothing on its own.
The Cognitive Bias Keeping This Problem in Place
The persistence of mean-based benchmarks isn't purely a data literacy problem. There's a cognitive explanation.
Tversky and Kahneman's foundational work on heuristics identified the availability bias: people assess probability based on how easily examples come to mind. High-value tribunal awards are memorable. They generate press coverage, professional discussion, and online attention. The thousands of £800 injury to feelings awards are not. They are decided quietly, published without ceremony, and never referenced again.
The figure that comes to mind when someone says "employment tribunal payout" is shaped by what's available to recall, not by the underlying distribution. That's why the mean persists as a benchmark even among practitioners who should know better. The large cases are vivid. The typical case is invisible.
And this is not a minor distortion. The availability of dramatic cases systematically inflates intuitive estimates of average outcomes across the board. It's the same mechanism that makes people overestimate plane crash risk and underestimate car accident risk: frequency of coverage, not frequency of occurrence, drives the intuition.
The fix requires seeing the whole population. Not the cases that get talked about, but all 129,000+ decisions, sorted by compensation, filtered by relevant characteristics, showing the actual shape of outcomes.
How to Apply This in Practice
The practical implication is straightforward: stop using the mean as a primary reference point for quantum advice.
For claimant solicitors: Filter by claim type and discrimination ground first, then look at the median and 75th percentile for that category. That gives a realistic central estimate and a credible upper range for well-evidenced claims. Use the mean only to illustrate upside where genuine aggravating factors are present.
For respondent advisers: The median is your baseline exposure assessment. The 90th percentile, filtered by employer type and discrimination ground, tells you where serious risk concentrates. Local authority employers face structurally higher awards because the Public Sector Equality Duty creates an expectation that they should have known better. That's not theoretical risk. It's in the data.
For settlement negotiations: The gap between mean and median is negotiating territory. A claimant anchoring on the mean in a case that looks median or below is making a quantifiable error. Showing the actual distribution for comparable claims puts that anchor in context without requiring a lengthy expert report.
For any quantum advice at all: Sector, employer type, discrimination ground, and claim type all shift the distribution meaningfully. No single number summarises that variation accurately. The distribution does.
The data is there. 129,000+ decisions, fully searchable, with compensation figures, claim types, sectors, and outcomes. The question is whether you're looking at a summary stat or the actual shape of what tribunals award.
Search employment tribunal decisions by claim type, sector, and discrimination ground on TribDB. You can also search fitness to practise hearing outcomes across six UK healthcare regulators. Free 14-day trial, no card needed.
Data source: 129,000+ employment tribunal decisions from GOV.UK, plus 15,000+ fitness to practise decisions from HCPC, NMC, MPTS, GDC, GPhC, and GOC. Updated weekly.
Reference: Tversky, A., & Kahneman, D. (1974). "Judgment under Uncertainty: Heuristics and Biases." Science, 185(4157), 1124-1131.