Invisible Variance: Why Standard Cost Fails in Demand-Driven Production

A March closing meeting at an industrial parts manufacturer. The CFO presents the numbers: “Product A planned unit cost was 130 TL, actual came in at 142 TL. That’s a 9.2% variance. Why?” The production manager shrugs: “One customer wanted different materials, another changed the operation sequence, and a third had higher scrap rates.” CFO: “But those weren’t in the plan.” Production: “We decided when the order came in. On the spot.”

This dialogue repeats itself every month-end in every factory doing make-to-order. The standard cost formula starts with “Product A = 50 TL materials + 75 TL labor + 5 overhead.” By March-end, the actual is “Product A = 55 TL materials + 82 TL labor + 5 overhead.” The 12 TL difference doesn’t get detailed anywhere. It just shows up as “variance.”

Over the first five days of this series, we broke the system into pieces: ERP’s record system, MES’s execution layer, MOM’s orchestration role, and how these three connect. Today’s question is more concrete: Without these layers, why does calculating cost correctly in demand-driven production become impossible?

Standard Cost’s Core Assumption

Standard cost calculates what a product “should” cost. Fixed recipe, fixed routing, fixed hourly rate. If you’re producing a thousand standard units of A per month, this formula works correctly hundreds of times over.

But in make-to-order, every order is different. First customer: “50 units, standard.” Second customer: “50 units, no washer.” Third customer: “50 units, different material, and I need it by Friday.”

Let’s look at the no-washer order. The standard recipe includes a washer. Material cost is 50 TL. Without the washer, it’s 45 TL. ERP continues with 50 TL. Because in ERP, Product A’s material cost is 50 TL. “Product A with washer” and “Product A without washer” aren’t separate product codes—they’re variants of the same product.

The operator skips the washer. Actual material cost becomes 45 TL. But the cost system still deducts 50 TL. A “positive variance” of 5 TL appears. Then the third customer requests a different material costing 60 TL. The system still runs at 50 TL. A “negative variance” of 10 TL materializes.

These two different variances appear to offset each other at month-end. Net variance might zero out. But none of reality reflected itself: One order came in 5 TL cheaper, another 10 TL more expensive. Net variance: 5 TL. But which customer generated how much profit, how much loss—that data is gone.

The Hidden Cost of Routing Changes

The second problem is operation sequence. Standard routing says: CNC, then welding, then inspection. Total time: 45 minutes. Labor cost: 0.75 hours × 100 TL/hour = 75 TL.

But if a customer asks, “Can we use a lathe instead of welding? The welder is busy,” the routing changes: CNC, then lathe, then inspection. Time becomes 55 minutes. Labor: 0.92 hours × 100 TL = 92 TL.

Where does this 17 TL difference go? It goes nowhere. ERP says “Product A labor is 75 TL.” Production pays 92 TL. Accounting records 75 TL. The 17 TL gap between actual and recorded cost gets buried under “overhead” or “efficiency variance.”

Worse: This routing change might also affect the quality process. Inspection after lathe produces different results than inspection after welding. Scrap rate might jump from 3% to 6%. That additional scrap cost also disappears.

Scrap: The Unrecorded Accumulated Loss

The third and largest problem is scrap. Standard recipe assumes a 2% scrap rate. If 100 units are to be produced, 102 are started, 2 become scrap, 100 complete.

But in make-to-order, every variant produces different scrap. When a new material is tried, scrap hits 8%. When operation sequence changes, scrap reaches 10% with an unfamiliar sequence.

Without MES, scrap data gets written on paper or lives in the operator’s memory. When the work order closes, “98 completed, 2 scrap” is recorded. But why did those 2 scrap? Material? Operation? Machine setup? That information is lost.

ERP says “100 planned, 98 complete, 2 scrap.” Standard cost calculation assumes 2% scrap. Even if real scrap is 8%, the system calculates at 2%. The 6% scrap gap doesn’t flow into product cost. At month-end, someone calls it “efficiency variance” and nobody asks why.

Real-Time Cost: What Changes with MES+MOM?

Calculating cost correctly in make-to-order isn’t a “correct formula” question—it’s a “correct data” question.

MES records, from the moment a work order starts, which recipe was used, which routing was followed, how long each operation took, and how much scrap occurred. MOM takes this data and makes it meaningful.

A real-time cost engine works like this:

1. Order entry: MOM determines the recipe based on customer requirements. “No washer” gets entered into the system. Material cost is calculated at 45 TL.

2. Routing determination: MOM dynamically determines routing based on current machine and operator availability. “CNC, then lathe, then inspection. 55 min, labor cost 92 TL.”

3. During production: MES records start and end times for each operation. If actual time is 58 minutes, cost updates to 97 TL.

4. Scrap recording: MES records scrap quantity at each operation. If 4 units were scrapped at the lathe, that gets recorded and cost is added.

5. Work order closure: Actual cost returns to ERP. “For this order: planned 130 TL, actual 145 TL, difference 15 TL. Cause: Material change +5 TL, routing change +12 TL, additional scrap +3 TL.”

When this flow repeats for every work order, month-end reports build themselves. The CFO stops asking “why is there a 12 TL variance?” The system already shows why.

Organizational Impact: Right Pricing, Right Decisions

Knowing cost correctly isn’t just an accounting issue. It affects pricing, capacity planning, and customer selection.

One customer’s actual order cost might be 15% higher than standard cost. Are you selling to this customer at a profit or a loss? If you’re looking at standard cost, you might appear profitable. Looking at actual cost, you might be losing money.

Or the reverse: You’re wondering, “Can we do more business with this customer?” Standard cost says “yes, 20% margin.” But with actual cost, you see that every order varies and variance constantly shifts direction, bringing average margin down to 8%. This doesn’t mean “this customer is killing us,” but it does mean “standard pricing doesn’t work with this customer.”

Conclusion

In demand-driven production, cost isn’t a static number—it’s a dynamic flow. The standard cost formula might be a good guess for “average,” but it’s wrong for “actual.”

Without correct data, you can’t make correct decisions. Without correct data, you can’t price correctly. Without correct data, you can’t know which customers are profitable and which aren’t.

MES and MOM layers provide this correct data. They write each order’s own cost, own variance, own story.