Activity-based costing wired to actual production data. You stop arguing about averages and start managing real cost per part, per line, per shift — with finance and operations finally reading from the same numbers instead of two spreadsheets that disagree.
Per part
Costing tied to real production data
8-12 weeks
Scoped pilot on one line to usable real-time cost
One source
Finance and operations closing from the same numbers
It closes the books and it tells you almost nothing about which parts, lines, and customers earn their keep. If two or more of these sound like your last quarter-end, the number you are managing from is older than the decisions you are making with it.
01
Standard costing rolls up clean and hides where the money is actually leaking. The variance ledger reads like a horoscope: true enough to quote in a meeting, useless for deciding anything.
02
You are quoting against a cost model that was accurate two years ago and an SKU mix that has drifted three times since. Margin is the residual at month-end rather than a target you set.
03
One says the line is profitable. The other says it is bleeding. Neither has the data to settle it, so the meeting ends with a commitment to build another spreadsheet.
04
Or which one is quietly being subsidized by the rest. The mix decisions you should be making every quarter, you cannot make at all.
Every item below is something finance can close with or operations can act on the same day. Nothing on this list is a document.
Cost calculated against actual output, materials, labor, and downtime as the shift runs, rather than against last quarter's averages applied with an overhead rate.
Margin visible at the resolution where somebody can act on it. The conversation moves from “this plant” to this line, on this shift, running this part.
Costs tied to real activities, real volumes, and real downtime, sourced from the systems that already record them instead of retyped into a finance workbook.
Trace a margin surprise back to the line, the shift, and the event that caused it. The answer is a record, not a reconstruction.
Same-day signals when material yield, labor utilization, or downtime move the cost model meaningfully, instead of a surprise that surfaces at close.
Same numbers, same source, same definitions, agreed in writing. The standing argument about whether the line is profitable stops being a matter of opinion.
What the model is actually built from
Cost visibility fails when it is treated as only one of the three. We deliver across all four routes below, and most engagements blend strategy and integration with a rapid-impact entry point on a single line. Underneath all four is the same engagement model — assess, target, execute, sustain.
Strategy & Operations
When the question is which decisions the model has to support — pricing, mix, capex, make-or-buy — and what that implies for the data underneath. We turn a costing ambition into a sequenced, fundable plan.
Digital Integration
OT and IT integration, MES, historians, and the connective tissue that turns shop-floor signals into numbers finance can close with. A cost model is only as honest as the data feeding it.
Rapid Impact
One line, one honest cost number, before anyone funds a plant-wide program. This is the entry point most costing engagements actually start from.
Lifecycle
Adoption, training, and sustainment so the model keeps earning. Costing models decay faster than most systems, because the mix changes underneath them every quarter.
This page is about the method: finding your real cost drivers and wiring an activity-based model to what the floor actually did. The engineering argument for how machine, production, and quality data reaches the ledger is made in full on the finance page — we will not restate it here.
Read the cost genealogy argument for financeWhat has to be true underneath
When a model needs hands on the hardware or the business systems connected, Magic Engineering and Magic Integration are in the same P&L — same project, same accountability, no subcontracting.
We have not published a costing engagement yet, and we are not going to borrow somebody else's logo to imply one. What we can do is show you exactly what the work rests on, and let you judge whether that is enough to spend thirty minutes on.
The method
Not a philosophy. A model built from machine time, labor, scrap, and downtime as they are actually recorded, on the same assess-target-execute-sustain engagement model as every other path we run.
The data layer
The sub-second acquisition architecture we built for a global medical device manufacturer — roughly 10,000 candidate data points per line, throughput and yield each up about 15% — is the kind of foundation an honest cost model sits on.
The depth behind it
Magic Consulting sits inside a group with 15,000+ professionals and 6,000+ customers in 50+ countries — combined group figures — with OT engineering and enterprise integration in the same organization when the model needs them.
ERPs are very good at recording transactions and rolling up standard costs. They are not built to reflect what happened on the line in the last fifteen minutes. Activity-based costing needs the floor layer feeding it, at a resolution the ledger was never designed to carry. That gap is the work.
No. Most engagements leave the finance system exactly where it is and feed it cleaner, more granular data. The cost model lives between the floor and the ledger, which is where it belongs, and your close process does not get rebuilt to accommodate it.
A scoped pilot on one line typically delivers usable real-time cost inside 8 to 12 weeks. Plant-wide rollouts are sequenced after that, and we do not recommend attempting every line at once — the model has to be argued with and corrected on one line first.
They are the same data layer, sliced differently. The cost view sits on the same instrumented platform as throughput, yield, and downtime: same numbers, financial lens. If the production data is not yet trustworthy, that is the first phase, and we will tell you so.
By role. Operators see the production metrics they can act on. Supervisors see line-level cost behavior. Plant leadership and finance see the roll-up. The granularity is matched to the decision each role actually owns, which is also how you keep the model from being argued with for the wrong reasons.
That is the wrong first question, and we will not answer it on a web page. The model is specified against your cost drivers, your installed systems, and the decisions it has to support — and the recommendation names the right tool for that, which is not necessarily a Magic tool. Sometimes the right answer is the system you already own.
Most engagements touch two of these. If you are not sure which one you are looking at, the readiness questions on the Consulting page will point you.
Data Visibility
A cost model is only as honest as its inputs. If floor data is not trustworthy yet, visibility is phase one of this engagement.
Read the pageEliminate Waste
Once cost is visible per line and per shift, the losses sort themselves into an attack order by money rather than by volume.
Read the pageCapacity & Throughput
Knowing the real cost per part changes which line is worth more capacity, and which product mix deserves it.
Read the pageThirty minutes, working session, your numbers, with finance and operations both in the room if you can manage it. We will walk you through how we would build it — or tell you straight if we are the wrong people for the problem.
