Why your comp bands break at the VP line

A salary band is a statistical object before it's a policy. It works when a role has enough people doing roughly the same job that their pay forms a distribution you can trust: a midpoint that means something, a spread that reflects real variance, edges you can defend to a comp committee. For an engineer, an analyst, a manager, that population exists. You have dozens internally and thousands in the market, and the band sits on top of them like a well-fitted curve.

Above the VP line, the population thins to almost nothing. A company might have one VP of Engineering, one Head of Data, one General Manager for a new line. The market for each is a few hundred credible people, and no two of them are carrying the same scope. One ran a 400-person org through a re-platforming. Another built a function from 12 people to 80. A band asks you to average those into a single midpoint, and the average describes neither of them.

So the band you quote at this level is a guess dressed as a benchmark. The spread is wide because the roles are genuinely different, and the wideness is the tell: when the band has to be that loose to fit the people in it, it has stopped doing the job a band is supposed to do.

Hold the band firm at the VP level and the structure hands you two exits, both bad. The first is that you lose the candidate. The person you spent 4 months finding sits above the top of your range because their last scope was larger, or the market for their specialism has moved, or a competitor already offered them a title you won't match. You walk because the band says walk, and the requisition stays open for another quarter while the cost of the vacancy runs quietly in the background.

The second exit is the exception. You want the person, so you get an approval to go off-band. It closes the hire. It also puts someone into the org at a number the structure doesn't explain, which means the next senior offer has a precedent it has to reckon with, and the one after that. Each exception is reasonable on its own. Together they're the erosion of the exact system the band existed to protect.

Both outcomes come from the same root. You're applying a tool built for populations to a hire that's closer to a negotiation between two specific parties. The band failed because at this level there's no number that's simultaneously firm enough to hold the structure and flexible enough to win the person.

The reason the population is small is only half the problem. The other half is that senior compensation is a function of scope, and scope is the least standardised thing in a job. Two people can hold the same title, report to the same level, and be doing work that differs by a factor no band captures: budget owned, headcount underneath, whether the mandate is to run a machine or to build one, how much of the P&L moves when they get a decision wrong.

This is sharpest in GCCs, where a single title can span a very wide reality. A “VP, Engineering” at one centre owns a maintenance mandate for a mature platform. At another, the same title owns a greenfield charter to stand up a capability the parent doesn't yet have. The market rewards those differently and should, but the band the parent hands down often reflects a global grade that was set against neither.

The senior market is also thinner in a way that compounds the idiosyncrasy. Executive search exists as a distinct discipline precisely because these roles aren't filled from an open pool. When the credible field for a role is small and every candidate's value is tied to a specific track record, an averaged band is the wrong instrument. You're pricing what this particular person's scope and proof are worth to this particular mandate.

The instinct when a band won't stretch is to reach for more base. It's the most legible lever and the most dangerous one, because base is permanent, visible to the next hire, and the easiest thing for a competitor to beat by a few percent. A cash bidding war is a fight you can win and still lose, since the number you won on becomes the floor someone else has to clear next year.

The senior candidates who move rarely move for base alone. What they're weighing is the shape of the role: how much they'll own, what the title signals inside and outside the company, how their equity compounds, and whether there's a real path to the level above. These are the levers that close the hire, and most of them cost less than the base you were about to over-commit. The table below is how we structure that conversation.

The structural answer is to stop pretending the senior tier is a band and design it as a range with defined latitude. A band has an edge you either respect or violate. A range with latitude has an approved zone of movement built into it, so the flexibility every senior hire requires is inside the system. The exception stops being an exception because the structure anticipated it.

In practice that means a wider published range at the top, and a named authority who can move within it against written criteria: the scope the person will own, the scarcity of the specialism, the strength of the proof behind their track record. The latitude is real, but it's governed. A comp committee can look at any senior offer and see which criterion moved the number and by how much, which is exactly the legibility a string of ad-hoc exceptions destroys.

This isn't a licence to pay whatever it takes. Governed latitude is more disciplined than a rigid band, because it forces the reason into the open. When someone lands near the top of the range, the file says why: this mandate, this proof, this market. That record is what lets you defend the offer to the board and explain it to the next person who asks how the number was set.

The move is to find the line in your own structure where the population gets too small to trust a midpoint, and treat everything above it differently. For most organisations that line sits around the VP tier, though the exact grade depends on how flat the org runs and how many incumbents a level actually holds. Below the line, keep the bands: they work because the distribution is real. Above it, switch to ranges with latitude and a defined set of non-cash levers you're prepared to use before base.

Done well, this spends the same budget more accurately. You stop losing senior candidates to structures more willing to move than yours, and you stop breaking your own bands one reasonable exception at a time. The senior offer becomes a designed instrument the committee can defend, which is what a band was always trying to be and, above the VP line, could never quite manage.