← All posts

Best Practice: Labour Roles, Rates & Modifiers

A single flat rate per hour is fine right up until it isn't. The moment a business bills differently for a senior engineer than a junior one, or charges more for a weekend callout than a Tuesday afternoon, one number can no longer describe the price of an hour of work. It depends on who did it, and when.

The obvious fix is to take a base rate and apply a multiplier for seniority, another for after-hours. It turns out to be the wrong one. A 50% after-hours premium doesn't mean the same thing for a junior engineer on a modest base rate as it does for a principal engineer on a high one. Seniority premiums and time-of-day premiums don't stack predictably, because they aren't the same kind of adjustment to begin with. This guide covers the model that holds up: roles and modifiers as separate, named dimensions, with the rate for any real combination of them set explicitly rather than computed on the fly.

A role is what the work is worth, not who did it

A labour role is a priced tier of work: Junior Engineer, Senior Engineer, Project Lead. It is not a person. Several people can share a role, and one person can hold several roles if their work genuinely spans tiers. Keeping that distinction clean matters. A role tied to a specific person stops being a pricing structure and becomes a workaround that breaks the first time someone changes job title or a replacement is hired at a different rate.

Keep the list short enough that it still means something. Four to six roles is a practical ceiling for most teams. That's enough to separate genuinely different value, and not so many that a client reading an invoice can't tell what they're paying for. A narrow gap between two roles' base rates is usually a sign they shouldn't be separate roles at all.

Each role carries a base rate, a billable setting and an agreement scope. The base rate is the starting figure before any modifier applies. The billable setting says whether the role is billable at all, since some roles exist purely for internal time tracking. The scope says whether the role's time counts against block-hours agreements. Modifiers adjust the base rate for context. They don't replace it.

Modifiers describe the context, not the price

A modifier is a tag for a circumstance that can change what an hour is worth. It isn't a number attached to the role. Two dimensions cover the vast majority of real pricing needs:

Location describes where the work happened: remote, onsite, a client site with its own travel or access complexity. Time describes when it happened: standard hours, after-hours, weekend, a public holiday. Keeping these as two independent dimensions, rather than one combined list of "situations", is what makes the system scale. Adding a new location doesn't mean redefining every time modifier from scratch, and vice versa.

A modifier by itself carries no rate. "After-hours" means nothing priced until it's paired with a specific role. It's a label applied to a time entry, not a formula. That separation is deliberate, and it's what sets this model apart from a simple surcharge: the rate for any role-and-modifier combination is set explicitly, once, rather than derived by multiplying a percentage against a base number.

Why the rate is a lookup, not a calculation

A multiplier model assumes every role's premium scales the same way, so that a 50% after-hours uplift applies uniformly across the rate card. In practice that breaks almost immediately. A client-site premium might matter a great deal for a junior engineer's travel burden and barely at all for a role that is remote by definition. A weekend premium that's fair at a modest base rate can look inflated once it's multiplied against a senior rate, or too thin to matter at a junior one. The relationship between a modifier and its effect on price isn't linear, and it isn't even consistent in direction across roles.

The fix is to stop calculating and start recording. Every real combination of role, location and time that the business actually bills gets its own explicit rate, set deliberately. That's a lookup table, not a formula, and it means a senior engineer's weekend-onsite rate and a junior engineer's weekend-onsite rate can reflect two genuinely different judgements about what that combination is worth.

Role Location Time Rate
Senior Engineer Remote Office hours Set explicitly
Senior Engineer Onsite Office hours Set explicitly
Senior Engineer Onsite After hours Set explicitly
Junior Engineer Onsite After hours Set explicitly, and not a multiple of the row above

The same shape of table exists one level up for rates scoped to a specific client agreement: the identical role, location and time key, resolved against that agreement's own negotiated rates before falling through to the main table. A client who has negotiated a different after-hours rate for a specific role gets exactly that, on the agreement's Labour Rates tab, without a parallel rate card built just for them.

Assigning roles to people, not the other way round

A role structure only holds up if logging time against it is actually constrained. Otherwise it's a pricing model on paper that anyone can bypass by picking whichever role happens to be convenient.

Each person is assigned a set of allowed roles, plus exactly one default. Time logging enforces it: an engineer can only log against a role they're assigned to, and the default is pre-selected to keep the common case fast. This does two things at once. It keeps billing honest, because a junior engineer can't accidentally (or otherwise) log time at a senior rate just because the option exists in a dropdown. And it keeps the rate structure meaningful as a statement about who is qualified to do what, not just what something happens to cost.

It also means a role change is a deliberate act. Granting someone a new role, or changing their default, is a visible decision made on their Employment tab, not a side effect of picking a different option one day. Quoted work is the exception: when time is logged against a quoted task, the task dictates the role, as the quote already fixed it.

A rate always resolves to a real number, or fails loudly

When an entry needs a rate, the system checks in strict order. Does the client's agreement have an override for this exact role, location and time? If not, does the main combination table have one? If neither does, time can't be logged, and the form says no rate is configured rather than guessing.

That is a deliberate design choice. A silent fallback, such as quietly defaulting to the role's base rate when nobody ever set a price for "Senior Engineer, onsite, weekend", would hide a genuine configuration gap behind a number that looks plausible on an invoice but was never actually decided by anyone. A hard, visible failure the first time that combination is used means the gap gets filled deliberately, by a person making a pricing decision. A wrong rate that nobody chose is worse than a blocked time entry that makes someone stop and set one.

In practice this means filling in the combination table isn't optional busywork. It's the configuration step that makes labour-role billing function at all. A role with no rate for a given location and time can't be billed under that pair until someone sets it.

Office hours vs after hours: the time modifier doing real work

The clearest everyday example of the time dimension is the office-hours and after-hours split. It's worth setting up properly rather than leaving it to whoever happens to notice a job ran late.

Define office hours once, and the national holiday list alongside it. Without a holiday list, a bank-holiday callout quietly bills at the daytime rate, because nothing told the system that day was different from any other Tuesday.

Pick a rule for the boundary case deliberately. A job that starts inside office hours and runs past them needs a stated policy. Either the whole entry bills at the rate of the window it started in, or it splits into an office-hours part and an after-hours part, each priced at its own rate. Both are legitimate. What matters is that it's a configured policy, not a judgement call made differently by each engineer. A grace period can hold the office rate for a short time past closing.

It's opt-in, and it fails safe. The split only applies to a role and location that has both an office-hours and an after-hours rate configured. If only one exists, the entry stays a single line rather than guessing at a split that was never priced. Quoted work keeps its quoted rate and is never split. Turning the feature on doesn't force every role into it overnight.

Common pitfalls

Defining far more roles than the business prices differently. A dozen roles with only a few distinct base rates between them is complexity nobody asked for. It makes the combination table bigger to fill in without making any pricing decision more accurate.

Assuming a multiplier would have been close enough. It's tempting to backfill the table by eyeballing "roughly 20% more for weekends" across every role. That reintroduces the exact problem explicit rates were meant to solve, and it takes longer to notice because the numbers look deliberate even when they weren't.

Leaving gaps and discovering them at invoice time. A "no rate configured" message during active client work is a worse moment to find a gap than during setup. Walk the realistic combinations, every role against every location and time pairing the business expects to bill, before relying on the structure in production.

Granting roles too liberally "to be safe". Assigning every engineer every role defeats the purpose of per-person assignment. The constraint only does its job if it's actually a constraint.

Using agreement overrides to paper over a missing standard rate. Overrides are for rates a client has genuinely negotiated. If the standard table has a hole, fix the table.

Pre-flight checklist for labour roles, rates and modifiers

Related guides