How to Price Manufacturing Work When Costs Keep Moving
For manufacturers and distributors · Written for owners and finance leads alike
Winning the work is hard. Holding the margin on it as you grow is a separate problem, and it is the one that decides whether the growth was worth it. Fixed costs spread as you add volume, so the company total looks steady, sometimes steadier than it should, while individual parts quietly go underwater. Four ordinary cost increases took one part from 29.5 points of margin to 19.7 while the system kept valuing every unit at the old number, and the company total barely moved. Here is the arithmetic, how to find which of your parts it is happening to, and how to get the next quote out on numbers that are actually current.
The same company, two views
Swipe to see the full chart →
The company number moved 1.4 points. The spread underneath it runs twenty-one.
Why growth doesn't automatically fix your margin
Growth should help, and it does. Spread the plant, the equipment and the overhead across more volume and cost per unit falls. That is real, and it is why the company-level number usually holds up or improves in a growth year. The trouble is that it holds up whether or not the parts underneath it are still earning.
A standard cost is right the day it's set. Then a supplier goes up nine percent, a freight surcharge appears on inbound loads, a component starts carrying duty, and wages tick up. None of those is a crisis. Nobody calls a meeting about any of them.
But the system is still costing that unit at the old number, and everything downstream runs off it: whoever is quoting work, the monthly margin report, the decision about which line to run more of.
Here is what that looks like on one part.
If you run finance inside a business like this
The arithmetic below is written out longhand, and you will be ahead of it. It is set out that way because this is usually the version that has to go in front of an owner or a board before anything gets funded. The concept rarely moves people and the per-unit number usually does. The sections worth your time are the one on systems and the one on cost to serve.
| Cost component | Standard | Actual today |
|---|---|---|
| Material | $42.00 | $45.78 |
| Freight surcharge on inbound material | none | $1.80 |
| Duty on an imported component | none | $3.10 |
| Direct labor | $11.00 | $11.60 |
| Overhead (assumed unchanged) | $14.00 | $14.00 |
| Cost per unit | $67.00 | $76.28 |
| Selling price | $95.00 | $95.00 |
| Gross margin | 29.5% | 19.7% |
A 9% material increase, a freight surcharge, a duty line, and a small wage move take nearly ten points off a part that still shows a healthy margin on paper.
Nothing on that list is unusual. A nine percent supplier increase is a Tuesday. So is a freight surcharge. Together they took the part from 29.5 points to 19.7, while the system still values every unit at $67.00.
At 40,000 units a year, that is $371,200 of real cost that never got attributed to a part.
Why your P&L can't tell you which product lost money
That $9.28 doesn't vanish. It was spent, and it is in the financials, just not anywhere that answers the question.
The system says the unit cost $67. It actually cost $76.28. The gap lands in variance accounts: material in purchase price variance, the labor piece in a labor rate variance. Those get reported in total, for the whole company, as one or two lines. Everything mixed together, no way to tell which part it came from or which customer bought it.
The freight and duty may not even make it that far. If those are being expensed as period costs rather than capitalized into the part, they never touch a variance account at all. They just sit in operating expense, well below the margin line you are reading. That is check number four below, and it is the one that most often comes back a surprise.
The report isn't wrong. It is answering a question about the whole company, and your question is about one part and one customer.
Now run the arithmetic in the other direction, because this is the part that catches people. If that part is a fifth of your revenue, losing ten points on it moves total gross margin by about two. Two points is a rounding error in a board meeting. It gets explained as mix, or a soft month, or somebody's aggressive pricing, and every one of those explanations is believable. Ten points is a product you should probably reprice or stop making.
That is the whole problem in one sentence. The bigger and more diversified the line, the more effectively the average buries the individual part, and the more confident everyone is that nothing much happened.
It usually surfaces at the worst time: a year-end inventory adjustment nobody can account for, or a buyer's analyst in diligence who re-costs your line and comes back with different numbers than yours.
Five ways to check whether your product costs are current
No software, no consultant, no project. Somebody pulling reports you already have.
Run these against your own numbers
- When was each standard cost last updated? If nobody can produce the date, that's your answer.
- Is unfavorable purchase price variance growing as a share of material purchases? Pull four quarters of each. If the share is climbing, the gap is widening.
- How big were last year's inventory and cost-of-goods adjustments? Big or repeating ones are a common sign the cost model isn't keeping pace.
- Is inbound freight and duty in the part's cost, or below the line? Sitting in operating expense, every part looks better than it is.
- Does the margin you quoted match the margin you got? Take three closed jobs. A gap of more than a couple of points points at the standard.
Two or more coming back fuzzy is enough to justify a real re-cost. None of it requires changing systems.
Two notes on those. Count accuracy, obsolescence and shrink produce the same year-end adjustments as stale standards, so rule those out before you blame the cost model. And the freight-and-duty check is the one most often missed. It takes a minute, and if the answer is wrong, every margin you've looked at this year is overstated.
Cost to serve: why two customers aren't worth the same
Parts get the attention. The other half of this is that two customers buying the same part at the same price are usually not worth the same to you, and nothing in a standard monthly package will tell you which is which.
What separates them has nothing to do with the part. It's how often they order, how small the orders are, who pays the freight out, whether it needs special labeling, how much comes back, and how long they sit on your invoice.
| Cost to serve, annual | Customer A 12 orders | Customer B 140 orders |
|---|---|---|
| Order handling and invoicing | $1,020 | $11,900 |
| Outbound freight | none | $43,400 |
| Custom labeling | none | $5,200 |
| Returns handling | none | $12,900 |
| Carrying cost of 36 extra days to pay | none | $20,100 |
| Total cost to serve | $1,020 | $93,500 |
| Units purchased | 24,000 | 23,800 |
| Contribution after cost to serve | $448,260 | $352,040 |
Using the $76.28 unit cost from above. Same part, same price, revenue within 1%, and a $96,000 difference in what the account actually contributes. Assumptions: order handling at roughly $85 per order fully loaded; Customer A buys FOB origin and pays its own freight, so the line is not zero cost, it is simply not yours; returns at a 3% rate on B; and carrying cost on the extra 36 days of receivable at 9%. Change those inputs and the gap moves. The point is that it is never zero and nothing in a standard package reports it.
On your P&L these two look identical. On the sales report, B probably looks better, since 140 orders reads like a stronger relationship than 12.
B isn't a bad customer. Nobody has priced the difference, that's all. Once you can see it, the fixes are ordinary: a minimum order, a freight policy, a small upcharge on the custom work, tighter terms. What you can't do is fix a gap you can't measure.
Both halves come back to the same thing. When cost data stops being current, it starts describing a business that was being run two years ago, and every decision taken off it inherits that lag.
How often should you update standard costs?
Most companies do it once a year. In a stable cost environment that's fine. In a volatile one it's how you end up eighteen months behind on a part you ship every week.
The practical version is not a calendar. It's a threshold and a short list.
Set a trigger, usually around five percent. When a material moves more than that and stays there for a couple of months, re-cost the parts it touches instead of waiting for the annual pass. A one-month spike you can let ride. A change that holds is a new reality.
Don't re-cost everything. Re-cost what carries the cost. Rank items by total annual material spend rather than by unit price or by revenue. The parts that matter are the ones where a percentage move is a large absolute number. In most shops that list is far shorter than the catalog, which is what makes this maintainable rather than a project.
Beyond the threshold, a handful of things should always trigger a look: a change in duty or trade classification on anything you import, a change in freight terms or carrier contract, any substituted component or new supplier for the same part, and any wage change that moves direct labor for real, including overtime that stopped being occasional.
The frequency matters less than the ownership. Cost standards tend to sit between purchasing, operations and finance, which is an efficient way for them to belong to nobody in particular. Naming who owns them is a smaller change than it sounds and it outlasts any single re-cost.
How tariffs and freight costs affect your margins
The mechanism above has always been true. What's changed is the speed. Costs are moving faster than the systems most companies built to track them.
The Institute for Supply Management's Spring 2026 Supply Chain Planning Forecast, published in June, projects manufacturing raw material prices rising 14.1% year over year. In the same survey, 26% of the manufacturers ISM polled said they would pass all tariff-related increases through to customers, 46% expected to split it between price increases and their own margin, and 17% planned to absorb it entirely.
The pass-through usually happens. What costs you is the lag between the cost landing and the price moving, and that lag is longer when nobody can say what the increase was worth per part.
To pass a cost increase through, you have to know the size of it per part. That's exactly the number a stale standard can't give you.
Which is the real problem. Nearly three quarters of the manufacturers in that survey intend to reprice. The ones working from current numbers will protect their margin. The ones working from an eighteen-month-old standard will either not raise enough and quietly lose ground, or raise too much on the wrong parts and lose volume they wanted to keep.
Have you outgrown your accounting system?
Sometimes the numbers are stale because nobody updated them. Sometimes they're stale because your system can't hold the information in the first place, and no amount of diligence fixes that.
The tells are specific. There's no way to put freight and duty into the part's cost, so they end up as period expenses. Units of measure don't convert cleanly. You buy in kilos, stock in pounds, consume in pieces, and the math quietly goes wrong. There's no real bill of materials or work order, so production lives in spreadsheets next to the accounting system instead of inside it. Inventory accuracy fell apart after the second location opened. Multiple entities get consolidated by hand every month.
Those are ceilings, not effort problems. If several are true at once, re-costing by hand gets you a better number this quarter and the same conversation next year.
The order that works: get the numbers trustworthy in the system you have, find out exactly where it runs out of road, then change deliberately, rather than during a year-end audit or a diligence request, which is when most companies get forced into it. Evaluating that, and running the move when it's time, is what we do under system design.
How to re-cost your products: a 30-day plan
- Date every standard.One column, one afternoon. Which parts haven't been re-costed since your last supplier increase? That list is where you start.
- Re-cost your top ten by revenue.Not the whole catalog. Ten parts will tell you how big the problem is and whether it's concentrated in a few places or spread across everything.
- Find the freight and the duty.Confirm whether inbound freight and duty are in your inventory cost or in operating expense. If they're below the line, every margin you've looked at this year is overstated.
- Rank your top twenty customers by order count, not revenue.Then look at freight terms and days-to-pay for the ones at the top. That is where cost to serve concentrates, and the ranking alone usually surprises people.
- Give the standards to somebody.The highest-return change on this list is naming one person responsible for cost data being current, with a trigger that doesn't wait for year end.
None of that needs a new system, a consultant, or a project plan. It needs a few hours from whoever knows where the reports live, and it'll tell you whether you have a maintenance problem or a structural one.
Turning the numbers into price moves
Re-costing tells you the size of the gap. It does not close it. The part most companies skip is the one that actually recovers the money, and it is worth being deliberate about, because a badly handled price increase costs more than the margin it was meant to fix.
Move the parts that moved, not the price list. The across-the-board increase is the move companies make when they don't have per-part numbers. Now you do. If one part lost nine points and another lost none, a uniform three percent overcharges the second customer and still leaves you short on the first. Take the ten or fifteen parts where the gap is real and price those.
Size the move to the gap. If a part is down 9.8 points, three percent does not get it back. Work from the new landed cost and the margin you actually want, then decide how much of the gap you are recovering now and how much waits. Partial is fine. Arbitrary is not, because you will not know later whether it worked.
Sort out what you can and cannot move today. Some of the line is on annual pricing, some is under contract, and some is neither. The uncommitted work can move on the next quote. For anything under agreement, put the renewal date on a list now, and make sure the next version carries a material adjustment clause tied to an index rather than to your goodwill. That clause is worth more over three years than the increase you cannot take today.
Be specific when you tell them. "Costs are up" invites a negotiation. "The alloy on this part is up nine percent since we set this price, and there is a freight surcharge of $1.80 a unit that wasn't there before" is a fact, and it is much harder to argue with. You do not have to show your cost sheet. You do have to sound like someone who has read it.
Decide where you are choosing not to move. There will be accounts you hold flat on purpose: the one filling second shift, the one that opens a market, the one whose volume covers real overhead. Holding those is a legitimate business decision. Holding them without knowing what they cost you is not. Write down what each exception is worth so it gets revisited rather than inherited.
Check it in ninety days. Did volume hold on the parts you moved? Did the gap close by roughly what you expected? If margin did not improve as much as the arithmetic said it should, the standard was still wrong somewhere, and that is worth knowing before the next round.
Common questions about product costing
What's the difference between standard cost and actual cost?
Standard cost is what you decided a unit costs, a fixed figure your system uses to value inventory and cost of goods sold. Actual cost is what you really spent. The difference between them piles up in a variance account, and the further apart they drift, the less your reporting means.
Standard costing isn't a flaw. It's what makes closing the books fast and quoting possible. It just assumes costs hold roughly steady between updates, and that assumption is the part currently breaking.
How do I know which products are actually profitable?
Re-cost them at today's prices, including material, inbound freight, duty, labor, and overhead, then compare that to what you're charging. Do your top ten by revenue first. If the answer differs from what your system shows by more than a couple of points, you have your answer and you know roughly how far it extends.
The trap is comparing against a standard that's already out of date, which just returns the number you started with.
Should I raise prices to cover tariffs, or absorb them?
That's a commercial decision and it varies by product and by customer. What matters is that you can't make it well without knowing the size of the increase on each part. Most companies raise across the board because it's the only move available when the per-part number isn't there, which protects the average and quietly damages both ends of the line.
Get the per-part number first. Then the pricing conversation is a strategy question rather than a guess.
How do I know if a customer is worth having?
Start with order count rather than revenue. Then add up what that account costs you outside the product: order handling, outbound freight, special packaging or labeling, returns, and the carrying cost of however long they take to pay. Compare that to the margin the account generates at current costs.
Most companies find one or two accounts near the top of the revenue list that are considerably further down the profit list. The point isn't to fire them. It's to price them correctly.
What is landed cost and why does it matter?
Landed cost is the total cost to get a material into your building: the price you paid plus freight, duty, brokerage, and insurance. If those extras sit in operating expenses instead of in the product's cost, every part looks more profitable than it is, and the error grows with every freight surcharge and tariff line.
Do I need new software to fix this?
Usually not, at least not first. Most of what's described here can be corrected inside the system you already have. New software becomes the right answer when the system genuinely cannot hold the information: no landed cost, no bill of materials, no clean unit-of-measure conversion, no real multi-entity consolidation.
Getting the numbers right in the current system also makes an eventual migration far less painful, because you'd have to do that work anyway before moving.
How long does it take to re-cost a product line?
Ten parts is usually a few days once someone has current supplier pricing and freight in hand. A full line depends on how many active items you carry and how good your bill of materials is. The bottleneck is almost never the arithmetic. It's assembling accurate current input costs, which is also the part that pays off permanently.
Where this fits in the bigger question
Cost data going stale is usually a symptom, not the disease. It happens when the finance side is fully occupied keeping the business closed and current, and nobody owns the work that sits above the close: keeping standards current, watching variance, looking at what each customer actually costs to serve.
That's a capacity question, and the answer depends on where the gap really sits.
If you want a second set of eyes
APN Group works with manufacturers and distributors across Georgia, Florida, and Tennessee on exactly this: inventory and cost accounting, margin by product and by customer, the pricing decisions that follow from it, and the systems underneath all of it.
Run the five checks first. If they come back clean, you have your answer and it cost you an afternoon. If they don't, send us what you found and we'll tell you what we think it means.
Survey figures: Institute for Supply Management Spring 2026 Supply Chain Planning Forecast, published June 2026. Figures in both tables are illustrative examples built to show the mechanism. They are not client data and not industry benchmarks.