Money At Work
A salary band is set before anyone has met you
Compensation for a role is usually decided as a structure months in advance, which explains why individual negotiation has less room in it than people expect.
By Julien Perrot3 min read
Pay is a structure, not a per-person judgement
Most people imagine their salary as the output of a negotiation about them: their skills, their interview performance, their nerve on the day. In organisations above a certain size, that is close to backwards. The number is drawn from a band that already existed, that was set by people who will never meet you, and that applies to everyone doing broadly that job at that level.
The reason is administrative rather than sinister. Paying every person a bespoke number creates an unmanageable problem: two colleagues doing identical work on very different salaries is a retention risk, a fairness problem and in some jurisdictions a legal one. Bands exist to make pay defensible, and defensibility is worth more to an employer than flexibility in any single case.
Where the numbers come from
A band is typically built from two inputs pulling against each other. The first is external market data — surveys of what comparable roles pay in comparable places, usually bought from firms that specialise in collecting it. The second is internal equity, meaning what the organisation already pays people at adjacent levels, which constrains how far any one band can move without disturbing everything around it.
Then someone chooses where in the market to sit. An employer might aim at the middle of the range for its sector, or deliberately above it to compete for scarce skills, or below it while offering something else — stability, flexibility, a mission, a shorter week. That positioning decision is strategic and slow to change, and it is generally the largest single factor in what you will be paid.
None of this produces a fixed point. A band has a floor, a midpoint and a ceiling, and the spread between them is often substantial. That spread is where the actual conversation happens.
Position in band matters more than most people realise
Where you sit within your band is the number with the longest tail, because percentage increases compound from wherever you start. Someone hired near the bottom and someone hired near the middle, both performing identically, can stay apart for years, since annual increases are usually applied as a percentage of the current salary rather than as a correction toward a target.
Employers are aware of this and many track it deliberately — the ratio of your salary to the band midpoint is a standard measure. Being low in band is sometimes used as an argument for a larger increase, on the grounds that it is a retention risk. But it is equally often invisible unless somebody raises it, and it is a reasonable thing to ask about in plain terms.
What this means when you are negotiating
The practical consequence is that arguing about the number in isolation is arguing with the wrong object. The more productive questions are about the structure: what level is this role, where does the band for that level start and end, and where in it does this offer sit. Some employers publish bands, some will describe them if asked, and some treat the whole thing as confidential.
Level is the lever with the most travel. Moving from the top of one band to the bottom of the next is a bigger change than anything you will win by haggling within a band, and it also changes the ceiling for everything that follows. If you have a genuine case that the role is scoped above the level advertised, that case is worth making before the offer is drafted rather than after.
The honest caveats
Not every employer works this way. Small organisations frequently have no bands at all, and pay whatever the founder decided seemed right, which cuts both ways: there is far more room to negotiate and far less protection against arbitrariness. Startups may weight compensation toward equity, which is a different kind of promise with a different risk profile. Public sector and unionised environments often have published scales with almost no individual discretion.
It also varies enormously by country. Norms around whether pay is discussed openly, whether bands are disclosed, and how much annual movement is normal differ so much that advice from one market can be actively misleading in another. Anything specific — tax treatment, contractual entitlements, what a particular scheme actually pays out — is worth checking against local rules or a qualified adviser rather than assumed.
What travels is the shape of the thing. Pay is mostly set by structure, structure is set in advance, and the questions that move a number are about which box you are in rather than about how deserving you sound inside it.
Common questions
Should I ask what the band is?
Yes, and asking is far more common than it used to be, particularly where pay transparency rules have arrived. Phrase it neutrally — "what is the range for this level, and where does this offer sit within it?" — because that is a factual question rather than a demand, and a recruiter can usually answer at least part of it.
Does being underpaid relative to the band get corrected automatically?
Sometimes, through what is often called a market adjustment, but you cannot rely on it. Correcting somebody upward costs budget that has to come from somewhere, so it tends to happen when there is a visible risk of losing the person. Raising it explicitly, with the internal comparison rather than an external one, is usually the faster route.
Is discussing pay with colleagues a bad idea?
It depends where you are, and the legal position differs by country — in some places contractual pay-secrecy clauses are unenforceable, in others the practice is simply discouraged socially. It is worth knowing that information asymmetry about pay is not neutral in its effects, and it does not favour the person with less experience of the market.
Deputy editor, After the First Job
Julien covers first months, managing up, money at work and the questions readers actually send in and would rather show the working than assert the conclusion.
