Flat Overhead Allocation Is Distorting Your Job Profitability Report

Almost every project-based business produces a job profitability report, and almost every one of them ranks jobs by margin. It's one of the more useful documents a business can have — provided the overhead allocation underneath it reflects how overhead is actually consumed.

In most businesses, it doesn't.

An accounting convenience becomes a management input

Flat percentage allocation exists for a good reason. It's simple, it's consistent, and for statutory reporting it's perfectly adequate. Total overhead divided by total revenue gives a rate, that rate is applied to every job, and the numbers reconcile.

The problem arises when a figure designed for reporting gets used for decision-making.

Revenue is a poor proxy for overhead consumption. Overhead in an operational business is mostly management time, supervision, administration, procurement, safety, scheduling and the finance function. Those resources are consumed by complexity, duration and interaction count — not by contract value.

Who subsidises whom

Consider two jobs in the same year.

The first is a $1.8m fitout for a repeat client with a competent superintendent, running on one site, with a clear scope and monthly claims. Your project manager touches it twice a week. There are four variations across five months, all agreed inside a fortnight.

The second is a $180,000 package of remedial work across nine separate sites for a client who has never run a project before. It generates twenty-two site attendances, fourteen variation discussions, a stack of access coordination, three separate mobilisations and weekly phone calls to the office.

Under flat allocation, the first job carries ten times the overhead of the second. In reality, the second job probably consumed more.

Now run that across a portfolio. Your large, well-run jobs are absorbing overhead they don't cause. Your small, complex jobs are being credited with margin they haven't earned. The ranking is inverted at the edges, and the edges are exactly where you'd want to make portfolio decisions.

The decisions that follow

This matters because the report gets used.

It gets used to decide which sectors to pursue, which clients to prioritise, whether to keep a service line, how to set minimum job sizes, and how to reward project managers. If the underlying allocation is wrong, every one of those decisions is being made against distorted information.

The typical drift is toward high-volume, low-value, high-touch work — because it consistently reports well. Revenue grows. Overhead grows faster. Net margin compresses, and nobody can point to the job that caused it, because no single job looks like the problem.

Allocate on the driver

The correction is straightforward. Identify what actually constrains your business and allocate against that.

For most operational businesses it's one of three things:

Direct labour hours — appropriate where supervision and administration scale with crew activity. Overhead per productive hour is a number every operator should know anyway.

Supervision or management days — appropriate where the constraint is senior attention rather than crew capacity. Requires project managers to log rough time by job, which is less onerous than it sounds if it's weekly rather than daily.

Plant or equipment days — appropriate where owned plant, transport and maintenance form a material part of the cost base.

You don't need perfect precision. You need an allocation basis that correlates with consumption. Anything that correlates beats a basis that doesn't.

What changes when you get it right

Two things, usually.

First, the job profitability ranking reorders — and in most businesses at least one job type moves from the top quartile to the bottom. That single insight is often worth more than a year of cost-cutting.

Second, your minimum viable job size becomes a defensible number rather than a gut feel. When you know what a job actually consumes, you can decline work with confidence instead of taking it because the report said it made money.

You can rebuild twelve months of history on a new basis in an afternoon. It's one of the highest-return pieces of analysis available to an operational business, and it doesn't require a single new system.