The pattern is almost always the same: revenue and output get tracked because they're easy to count. The numbers that decide profitability (margin per product, machine efficiency, material usage) need structured data collection, and that infrastructure was never built. The owner senses something is off but can't identify the problem.
01Margin per product
The single most useful number a manufacturer can know, and the one most consistently missing. Revenue says which product sells. Margin per product says which product earns, after materials, labor, and a fair share of overhead.
The inputs usually already exist in the system: selling price, actual material usage, labor time. The calculation has simply never been done product by product.
What we typically find Compare the revenue ranking with the profit ranking; they can point to different priorities. The best earner can be a lower-volume product nobody pushes, because it "doesn't contribute much to revenue." A finding like this changes production scheduling, pricing, and sales incentives at once.
02Machine efficiency
Of the hours your machines were scheduled to run, how many actually produced sellable output? (Factories call this OEE.) Three questions explain the number: was the machine running, was it running at speed, and was the output good?
A floor running at 50% efficiency is sitting on hidden capacity. Whether the loss comes from downtime, slow cycles, or rejects matters: each has a different fix, and the split tells you which one to make.
03Reject and rework rate
Every rejected unit spent material, labor, and machine time to produce nothing. A reworked unit spends them again. Tracked by line, by shift, and by product, the pattern almost always points to a specific machine, team, or material batch: a fixable cause, not a vague "quality problem."
04Material usage against the BoM
The BoM says what each order should consume. Actual usage says what it did. The gap between the two is where material quietly disappears: over-issuing, waste, loss, or a BoM that no longer reflects reality.
For made-to-order manufacturers this number does double duty: it also drives purchasing. When every incoming order converts into a material list from the BoM, purchasing buys what production actually needs, before production starts, which helps keep stock from piling up or running short.
05On-time delivery
The share of orders delivered on or before the promised date. It reads like a customer-service number, but it's an operational X-ray: chronic lateness traces back to scheduling, material shortages, or overpromised capacity. Tracked by customer and by product, it shows exactly which promises the operation can't keep.
06Why five, not twenty
A KPI framework isn't a scoreboard of everything. It's the 5 to 8 numbers where a change demands a decision. More than that produces noise, and noise gets ignored.
- Margin per product tells you where to focus production
- Machine efficiency tells you whether you have the capacity to act on it
- Reject rate tells you whether quality is stable enough to scale
- Material vs. BoM tells you whether purchasing and the floor are in sync
- On-time delivery shows whether customers receive orders as promised
Review these five weekly and you spend less time chasing numbers on the floor and more time acting on them. That is the actual goal.
07The infrastructure question
None of this needs new software. Start by checking which of these your operation already captures: production logs, material receipts, inspection records, delivery records. The work is connecting those sources into one weekly view: precise definitions, agreed sources, and a refresh that runs in minutes.
In most businesses that's an Excel Power Query system. In some, a dashboard on the ERP export. The technology is secondary; the definitions are the product.

