How to structure salary bands by seniority: the nine-box method

Most bands are built the same broken way: take a market median, apply a fixed percentage spread above and below it, and call the result a band. That method works passably at one seniority rung and falls apart across a whole ladder, because the thing that actually varies band to band is not the percentage spread. It is the structure underneath it.

Why flat percentage spreads fail

A flat spread, say plus or minus 15 percent around the median, treats every rung the same. In practice, junior bands are naturally narrow (there is not much room between a weak and a strong graduate hire) and senior bands are naturally wide (a Strong Vice President and an Exceptional one can be worlds apart in scope and pay). Force both onto the same percentage spread and the junior band overpays its bottom while the senior band underpays its top.

The deeper problem is that a percentage spread answers the wrong question. It tells you how far a number can move. It does not tell you why one person in a role earns more than another at the same level, which is the question a hiring manager or a comp committee actually needs answered.

The two axes that matter

A band that survives scrutiny is built on two axes, not one.

The first axis is seniority rung: analyst, associate, manager, senior manager, director, and so on up the ladder, each rung defined by scope, not by title (title inflation makes titles an unreliable proxy for level, a problem covered in full elsewhere in this cluster).

The second axis is calibre within the rung: Standard, Strong, and Exceptional. A Standard performer at a given rung is meeting the bar for the role. A Strong performer is exceeding it on scope or impact. An Exceptional performer is operating meaningfully above the rung, often taking on responsibility that belongs to the rung above. Crossing these two axes produces a nine-box: three rungs' worth of calibre distinctions, if you are looking at three adjacent levels, each with its own defensible pay position rather than one blended number.

This is the structural difference between a band and a range. A range says "this role pays between X and Y." A nine-box says "this role pays X for a Standard performer, Y for a Strong one, and Z for an Exceptional one," which is the actual decision a manager is making when they set an offer or approve a raise.

Setting the width of each rung

Junior rungs should run narrow. There is limited room for a Standard-to-Exceptional spread when the role itself has limited scope to vary; a wide band at entry level usually signals an undefined role rather than real performance differentiation.

Senior rungs should run wide, and the widening should be deliberate rather than accidental. At director level and above, an Exceptional performer is frequently running a function an Exceptional-minus-one performer is not, and the pay gap should reflect that gap in scope, not tenure alone.

The gap between rungs matters as much as the width within them. If a Strong performer at one rung earns close to a Standard performer at the rung above, the ladder is compressed and gives people no reason to grow into the next level rather than staying put and negotiating harder within the current one.

When a band needs a track split

Some sectors need more than one ladder. Legal & Compliance is the clearest case: a Private Practice associate, an In-House counsel, and a Compliance officer at the same nominal seniority sit in genuinely different markets, with different peer sets and different pay ceilings. Forcing all three onto one ladder produces a band wide enough to be meaningless.

The Legal & Compliance sector hub shows this in practice: a Partner in Private Practice in Dubai sits at a median total cash of AED 248,625 across 6 verified sources, a figure that belongs to a different market entirely than an in-house Senior Counsel role at a similar tenure. A single ladder covering both would either overstate the in-house number or understate the private-practice one. The test for whether a sector needs a track split is simple: if two roles at the "same" level have different peer sets, different hiring pools, and different pay ceilings, they need separate ladders, not one wide band.

Keeping bands honest with source-counted data

A nine-box is a structure, not a source of numbers. Every cell in it still needs to be anchored to verified, source-counted market data, or the structure is an opinion organised into a grid, nothing more. An Associate cell built on 6 verified sources and a Partner cell built on the same 6 sources are both defensible; a cell built on a single recruiter's estimate is not, regardless of how well the grid around it is designed.

Build the grid first to decide what you are measuring, then fill every cell from a dataset that states its source count, not from a market memory or a single data point dressed up as a range.

Putting it into practice

Setting compensation bands walks through the workflow for building this structure against your own roster, mapping each person to a rung and a calibre tier rather than a single blended range. For the defensibility side, once the bands exist, setting defensible comp bands in the Gulf covers how to keep the source-count discipline in place as the bands age.

If you are building your first structured ladder, book a demo and we will walk through the nine-box against your actual sectors and levels rather than a generic template.