Margin Architecture
Synthetic data used for illustration purposes
Worked Example · Field ServicesThe contracts all looked the same. They weren’t.
A $22.4M mechanical services company is heading into renewal season guided by a margin report showing every maintenance contract inside a comfortable 32–39% band. Honest attribution of overtime, callbacks, travel, and fleet cost puts the real band at −5% to +43%.
Illustrative scope: one $9.2M service line with 12 contracts inside a $22.4M company. Illustrative fees comprise an $8k Financial Signal Diagnostic and a $26k Core Fix. Real engagements are scoped to the decision, evidence readiness, systems, and complexity. The $558k represents modeled opportunity, not realized or guaranteed client results.
Gross margin by contract
Switch the view. Same dollars, same twelve customers — the only change is where the cost lands.
Margins this uniform are usually a sign of allocation policy, not operational reality. Twelve customers with different buildings, densities, and service intensities do not naturally land inside a six-point band.
The change, contract by contract
Ordered by how far the reported number sat from the truth. Shaded rows fall below overhead recovery.
Where the margin went
Four cost pools sat below the line, attributed to no contract. Nothing was hidden — it was simply never assigned.
Same dollars, honestly placed. Overtime premium on after-hours callbacks sat in a payroll pool. Warranty rework was booked to a general account. Travel ran on a flat 4% assumption and fleet was charged per technician rather than per route-hour.
Size is not profitability
Revenue on the horizontal, true margin on the vertical. The second-largest contract in the book sits well below overhead recovery.
What the work meant
The diagnostic took three weeks and cost $8,000. The rebuild that followed — a contract-level true-margin model, a repricing pack for renewal negotiations, and a monthly margin pack the controller now owns — cost $26,000. Total investment: $34,000.
The decisions it unlocked model to roughly $558,000 in the first year. About $427k comes from repricing or restructuring five contracts that were quietly underwater. Another $131k comes from route-density scheduling and warranty root-cause recovery. That is a 16.4× ratio of modeled opportunity to fees, counting year one only, while the pricing corrections persist for the life of each contract.
The analysis also stopped an indiscriminate 5% renewal discount campaign ten days before launch — a cut that would have fallen hardest on the contracts already below overhead recovery, and that would have reduced gross profit dollar-for-dollar absent offsetting volume, retention, or cost effects. Any associated avoided downside is excluded from the headline economics above.
What leadership can act onNone of this was visible in a report that showed every contract between 32% and 39%. The report was not wrong about the dollars. It was wrong about where they belonged.
Start with the report or number leadership no longer fully trusts.
If a margin, profitability, pricing, cost, or revenue-unit signal is important enough to act on but not trusted enough to act on confidently, that is the conversation.
This worked example uses a field services book because contract-level attribution failures are easy to see there. The same pattern appears wherever cost pools sit above the contract, product, or customer that caused them.