The Quarter Nobody Could Explain
A shop owner runs two machines that could not be more different: a manual bandsaw that costs almost nothing to own and run, and a five-axis vertical machining center that cost more than a house. Both get quoted at the same shop rate, because that's the number that's always been used — the one from the accountant's spreadsheet, blended across the whole floor, unchanged for years.
At quarter-end, the pattern repeats. Jobs that were mostly manual saw work and deburr came in fine, sometimes better than expected. Jobs that spent real hours on the five-axis came in soft, sometimes underwater, and nobody could say exactly why — the labor hours matched the quote, the material matched the quote, but the job still lost money.
This is the blended shop rate trap, and it's structural, not a fluke of one bad quarter. A single average rate applied across multiple work centers with genuinely different costs will always overprice the cheap stations and underprice the expensive ones. The shop doesn't find out which jobs it's losing money on until the mix of work that quarter happens to lean toward the five-axis. This article walks through why the average breaks down, works through the arithmetic on paper, and lays out what replacing one blended shop rate with per-work-center rates actually looks like.
What a Blended Shop Rate Actually Averages
A blended shop rate is usually built one of two ways: total shop overhead and labor divided by total shop hours, or a number inherited from a prior owner, accountant, or competitor's rate card, adjusted occasionally for inflation. Either way, it treats every hour on the floor as interchangeable — an hour on a manual mill "costs" the shop the same as an hour on a CNC lathe with live tooling, a coolant system, and a much larger depreciation base.
That averaging isn't wrong as an accounting exercise; the total dollars in and total dollars out can reconcile just fine at the shop level. The problem shows up at the job level, because no individual job runs across the shop in the same proportion as the shop's overall hour mix. A job that's 90% saw and deburr time absorbs almost none of the real cost driving that blended number up, and gets quoted with a rate padded for machines it never touched. A job that's 90% five-axis time absorbs a fraction of the real cost of that machine, and gets quoted with a rate diluted by cheap stations it never touched either.
Why the Average Breaks Down Job by Job
The mechanism is simple once it's laid out, but it's easy to miss because both errors are invisible in isolation — they only show up as unexplained margin drift when the job mix shifts.
Consider what actually differs between work centers:
- Machine cost and depreciation. A five-axis VMC or a CNC lathe with live tooling carries a materially higher capital cost than a manual mill or bandsaw, and that cost has to be recovered somewhere.
- Power, tooling, and consumables. High-speed spindles, coolant systems, and specialty tooling cost more to run and replace than manual equipment.
- Skill premium. Operating a five-axis program or a wire EDM commands a different pay grade than manual deburr or saw work, and that shows up in the labor portion of the rate.
- Utilization. A machine that sits idle waiting for programs or fixtures a third of the time needs a higher rate on the hours it does run, just to recover its fixed costs — a point covered in more detail in work center utilization tracking.
A single blended number folds all of that into one figure. It can't distinguish a job that spends four hours on a manual saw from a job that spends four hours on a five-axis center, even though the real cost to the shop of those four hours is nowhere close to equal.
A Worked Example: Same Rate, Two Very Different Jobs
The following is an illustrative example for a representative shop, not a benchmark — the point is to show the mechanism, not to assert real-world figures.
Say a shop calculates a blended rate of $85/hour across the whole floor. Two jobs come in for quoting:
- Job A — 8 hours total, almost entirely manual saw cutting and hand deburr. Quoted at $85/hr = $680.
- Job B — 8 hours total, almost entirely five-axis VMC time. Quoted at $85/hr = $680.
Now suppose the shop worked out real per-work-center rates instead — again, illustrative figures only:
- Manual saw/deburr work center: $45/hr real cost to run
- Five-axis VMC work center: $140/hr real cost to run
Job A's real cost is 8 × $45 = $360, but it was quoted at $680 — overpriced relative to its actual cost, which might just mean the shop is winning less saw-heavy work than it should, or leaving margin on the table it doesn't realize it has.
Job B's real cost is 8 × $140 = $1,120, but it was quoted at $680 — underpriced by $440 relative to its actual cost. That's not a rounding error; it's a job that loses money every time it runs, disguised by a blended rate that looks perfectly reasonable on paper.
The blended rate isn't lying about the shop's average cost. It's lying about every individual job's actual cost — in both directions, simultaneously.
Building Per-Work-Center Rates Instead
The fix is to calculate a rate for each work center separately, based on that station's own burden — its share of depreciation, power, consumables, floor space, and the labor grade typically assigned to it — divided by its own realistic run hours. That's the method covered step by step in work center rate calculation.
It also means being precise about what's actually being priced: a machine rate (the cost of owning and running the equipment) is a different number from a labor rate (the cost of the person operating it), and the two get combined per work center rather than folded into one shop-wide average. The distinction and how to combine them is worked through in machine rate vs labor rate.
For a shop building this from scratch, the foundational math — the same logic applied to a single average — is laid out in how to calculate machine shop hourly rate, and applying it separately to every station is what actually closes the blended-rate gap.
Utilization: The Variable Blended Rates Ignore Completely
A per-work-center rate is only half the fix. The other half is knowing how many hours each station is actually running versus sitting idle, because a rate calculated against theoretical full-time hours understates the true cost of an underutilized machine. A five-axis center that's only cutting chips 60% of available hours needs its fixed costs recovered across fewer real hours, which pushes its true rate higher than a naive calculation would suggest. Tracking that utilization by work center — not just by the shop as a whole — is what makes the rate defensible rather than another guess, and it's covered in work center utilization tracking.
This is also where the blended-rate trap connects back to actual-vs-quoted tracking on the floor: a routing built with per-work-center rates only holds up if time is actually logged by operation and work center, not lumped into one job total. WorkTickets builds routings by work center, clocks time against each operation with setup/run/waiting/rework reason codes, and rolls actuals up against the quoted standard per operation — so a five-axis overrun shows up against the five-axis rate, not buried in a shop-wide average.
Putting Per-Work-Center Rates Into Practice
Moving off a blended rate doesn't require a full ERP rebuild. It requires listing every work center, pricing its real burden and labor separately, and checking utilization against those numbers — then applying the resulting rates at quoting time instead of the one inherited figure. For a structured starting point, the Multi-Work-Center Rate & Utilization Model walks through the calculation station by station and includes a utilization tracking layer so the rates stay grounded in real hours, not assumptions. It's a reasonable next step alongside the broader methodology in the job costing resource hub.
