The Monday Morning Report Nobody Acts On
Every shop has one: a printout or spreadsheet tab that gets updated Friday afternoon, glanced at Monday morning, and forgotten by Tuesday. It usually has a machine utilization percentage, maybe an on-time-delivery number, sometimes a scrap count. Nobody in the room can point to a decision that report ever caused. It's not that the numbers are wrong — it's that they don't connect to anything the owner or supervisor actually controls that week.
Meanwhile the real question — did job 4471 make money, and if not, which operation ate the margin — sits unanswered until the books close for the quarter, at which point it's too late to fix anything about that job. It can only inform the next quote, and only if someone remembers to look.
This is the split worth naming out loud: there are machine shop KPIs to track because they change what you do next, and there are machine shop KPIs that exist because someone once put them on a template. This article covers both piles, and how to tell which is which.
What Makes a KPI Worth Tracking
A KPI earns its place on a report if it meets three tests. First, it's tied to a decision someone in the shop actually makes — requote this customer, rebalance this work center, retrain this operator, drop this material. Second, it's measured at a level of detail that supports that decision — a shop-wide number can't tell you which operation on which job caused the problem. Third, it's cheap enough to produce that someone will actually keep it current; a KPI that requires an hour of manual reconciliation every week dies within a month.
Most shop KPI lists fail test two. They report shop-wide utilization or overall on-time percentage — numbers that feel informative but point nowhere specific. The fix isn't more metrics. It's fewer metrics, measured at the job and operation level instead of the shop level.
Job Profitability: Actual vs. Quoted, Per Operation
The single highest-leverage number a job shop can track is actual labor and machine time against the quoted standard, broken out by operation, not just by job. A job that comes in "about on budget" overall can still be masking one operation that ran 40% over and another that ran comfortably under — and if that pattern repeats across jobs, it tells you exactly where the quoting is wrong or the process is unstable.
This is the mechanism behind job profitability analysis: every logged actual gets compared against the routing's standard time for that operation, on that part number, and the variance rolls up into a job-level margin picture. Done at the shop level only, this collapses into a single vague "we made 12% on the quarter" statement that nobody can act on. Done at the operation level, it becomes "setup on this fixture runs long every time we quote it at 20 minutes" — a decision you can make before the next quote goes out, not after.
The worked mechanics are simple. Take a job quoted at 40 hours of labor across four operations. If the actual logged time comes in at 46 hours, that's a variance — but the number that matters is which operation carried it. If three operations landed within a few percent of standard and one ran 30% long, that's not a general labor problem; it's a specific process or setup problem, and it's fixable.
Work Center Utilization — Measured Honestly
Utilization gets a bad name because it's usually reported as a single shop-wide percentage that tells you nothing about where the bottleneck actually is. The more useful version breaks it down per work center: which machines or stations are running near capacity, which are sitting idle waiting on upstream operations, and which are absorbing rework that shouldn't be there in the first place.
Work center utilization, tracked this way, becomes a scheduling and quoting input rather than a scoreboard. If one work center is consistently the constraint, that's the one whose burden rate and lead time matter most in the next quote. If a work center shows low utilization but high waiting time logged against it, the bottleneck isn't the machine — it's the flow feeding it.
The distinction that matters here is between busy and productive. A work center can log high hours and still be absorbing rework or waiting time that never should have counted as productive capacity. That's why a downtime reason code — setup, run, waiting, rework — attached to every clocked hour matters more than the raw utilization number itself.
OEE vs. Job Costing — Different Questions, Not Competing Ones
Shops that come from a discrete-manufacturing or lean background sometimes try to import Overall Equipment Effectiveness wholesale, and it doesn't map cleanly onto low-volume, high-mix job shop work. OEE — availability × performance × quality — was built for repetitive, high-volume lines where a single part number runs for hours or days at a stretch. A job shop running a different part number every few hours doesn't have a stable "ideal cycle time" to measure performance against in the same way.
That doesn't mean the underlying question — is this equipment producing value during the hours it's scheduled — is wrong for a job shop. It means the measurement has to shift from equipment-centric to job-centric. OEE vs. job costing comes down to this: OEE asks how well a machine performed against its own theoretical capacity; job costing asks how well a specific job performed against its own quote. For a job shop, the second question is almost always the more actionable one, because the decision it feeds — requote, rebalance, retrain — is one the shop can act on this week, not one that requires months of stable production data to interpret.
Labor Efficiency Variance — Where the Story Actually Lives
Labor efficiency variance is the connective tissue between the operation-level actual-vs-quoted number and the payroll conversation nobody likes having. It's the gap between hours paid and hours the routing said the job should take, and it shows up in two very different flavors that get lumped together far too often.
One flavor is a training or staffing problem: a specific operator running consistently over standard on a specific operation, which is a coaching conversation, not a costing problem. The other is a quoting problem: every operator, regardless of experience, running over standard on the same operation, which means the standard itself is wrong and the next quote needs to reflect it. Labor efficiency variance, tracked at the operator-and-operation level rather than as a single shop-wide labor efficiency percentage, is what lets you tell these two problems apart. Collapse it into one shop-wide number and you'll spend a training budget solving a quoting problem, or requote a job that just needs a different operator on it.
What to Leave Off the Report
Some numbers belong on a wall poster for morale, not on a decision-making dashboard. Machine hours logged with no reference to what the job quoted for those hours. On-time delivery percentage with no breakdown of which stage of the routing caused the miss. A single blended shop-wide efficiency number that averages a great job and a terrible job into a forgettable "fine."
The test isn't whether a number sounds impressive in a meeting — it's whether it changes what you quote, staff, or fix next week.
If a KPI can't answer "so what do we do differently on Monday," it's decoration. Cut it, or push it down to the operation level where it earns a place.
Turning the List Into a Habit
None of this requires a KPI overhaul — it requires four numbers, tracked at the job-and-operation level instead of the shop-wide level: actual vs. quoted labor, work center utilization with reason codes, scrap and rework tied to the causing operation, and labor efficiency variance by operator and operation. That's the short list of machine shop KPIs to track if the goal is decisions, not decoration.
For shops rolling several open jobs together into a single monthly view, the Multi-Job Margin Rollup & Monthly Profitability Dashboard builds exactly this kind of operation-level summary across the jobs running in a given month, without waiting for the books to close to see where margin went. For a broader look at how routing, traveler, and clock-in data feed into this kind of costing picture from the ground up, the small job shop execution and costing guide walks through the full mechanism.
If you'd rather get this kind of breakdown in your inbox as new pieces come out — job costing mechanics, work center analysis, variance tracking — subscribe to the newsletter below.
