A bonus is a design, and the design tells you what it is for

Why a bonus exists in the first place

Paying part of compensation as a variable amount isn’t primarily generosity. It does three useful things for an employer. It converts a fixed cost into one that falls when results are poor, which protects the business in a bad year. It attaches a reward to something the organisation wants more of. And it creates a reason to still be there on the payment date, which is a retention device whether or not anyone describes it that way.

Those three purposes pull in different directions, and most schemes are a compromise between them. That compromise is why bonus rules so often feel incoherent from the inside: they are trying to be an incentive, an insurance policy and a set of handcuffs at once, and no single design does all three well.

Formulaic, discretionary, and the pool in between

At one end sits the purely formulaic bonus, common where individual output is measurable — sales, billable work, defined targets. It is predictable, which is its virtue, and it reliably produces behaviour aimed at the measure rather than the goal, which is its cost. Anyone who has watched a quarter end in a target-driven environment knows the shape of that.

At the other end sits pure discretion, where a manager assesses the year and a number appears. It is flexible and it can reward the things no formula captures, such as the person who quietly prevented a disaster. It also depends entirely on the fairness and memory of the person deciding, and memory is the weak part: an assessment made in one month frequently reflects the previous three rather than the previous twelve.

Most schemes sit in the middle. A pool is generated by company performance, allocated to departments, then distributed by managers according to individual assessment. This structure explains the most common source of confusion, which is an excellent personal year producing a disappointing payment. If the pool shrank, individual performance only determines your share of a smaller amount. Understanding that in advance turns an insult into an arithmetic outcome.

Read your own scheme for the specific clauses

Whatever the shape, a handful of details do most of the work and they are usually written down somewhere. Whether the payment is discretionary or contractual, which is a legally meaningful distinction in many places and worth checking properly rather than assuming. Whether you must be employed on the payment date, and whether being under notice disqualifies you. How a partial year is treated for joiners and leavers. Whether the company component has a threshold below which nothing pays at all.

The employment-on-payment-date clause is the one that changes decisions, because it can make the timing of a resignation worth a substantial amount. That isn’t a reason to stay in a bad job, but it is a reason to know the date before you start a job search, since interview processes take longer than people plan for and a small amount of sequencing can be worth more than a negotiation.

What a bonus is worth compared with base pay

Base salary and variable pay aren’t equivalent even when they add to the same total. Base compounds, because future raises and often future offers are calculated from it. It is also certain. A bonus is uncertain, is frequently reset each year, and in many schemes disappears entirely in a bad year for reasons that have nothing to do with you.

This matters when comparing two offers or weighing a rise in one against the other. A larger variable component is a transfer of risk from the employer to you, and it’s reasonable to want compensating for that rather than treating the headline total as the same thing. Whether it is a good trade depends on how stable the business is, how the scheme has actually paid in recent years — a fair question to ask — and how much variability you can tolerate. There is no general answer, and anyone offering one is selling something.

The behaviour a scheme actually produces

It is worth watching what your scheme causes people to do, including you. Measures attached to money get optimised, and the optimisation is rarely what the designer intended. Targets set at team level can quietly discourage helping another team. Individual targets can make sharing information costly. A scheme heavily weighted to one metric will, over a couple of years, reshape what people consider important, regardless of what the strategy says.

None of this means you should refuse to engage with the scheme. It means you should know what it is rewarding, be deliberate about the cases where doing the right thing and maximising the measure diverge, and be sceptical of your own reasoning at those moments. It is also fair, and occasionally effective, to point out to a manager where a measure is producing a perverse result. Schemes do get changed, usually slowly, and usually because somebody described the perverse case concretely rather than complaining that the scheme was unfair.