SORT, REWORK OR SCRAP
Three ways out of the same suspect lot. This prices all three per good part.
You made the call at 7am. A week later you have to defend it to finance, or to an MRB, with numbers. This sort, rework or scrap decision calculator prices the cost of sorting vs scrapping parts on a cost-per-good-part basis, runs the rework vs scrap cost analysis alongside it, and returns the defect rate at which sorting stops paying for itself.
Free, no sign-up, nothing gated. It calculates the moment the page loads, and every line item is on screen so you can check the arithmetic yourself.
PRICE ALL THREE.
PER GOOD PART.
Total cost is the wrong yardstick, because the three dispositions do not deliver the same number of shippable parts. Cost per good part is the comparison that survives a finance review. Change any figure below and every number updates as you type.
Calculating with the default assumptions…
Sort
Cost of poor quality per good part — —Sort + rework
Cost of poor quality per good part — —Scrap + replace
Cost of poor quality per good part — ——
Sort
C = (Q÷R)×L + escapes×k₂ + freight + cover
| Line item | Low | High |
|---|---|---|
| Defectives in lot | — | — |
| Caught by the sort | — | — |
| Escapes to customer | — | — |
| Sort labour | — | — |
| Escape cost | — | — |
| Expedite freight | — | — |
| Cover for parts pulled | — | — |
| Total cost | — | — |
| Good parts shipped | — | — |
| Cost per good part | — | — |
Escapes are not credited as good parts. They shipped, but they were defective, so the denominator is the genuinely conforming population only.
Sort + rework
C = (Q÷R)×L + caught×(m÷60)×Lrw + fallout + escapes×k₂
| Line item | Low | High |
|---|---|---|
| Defectives in lot | — | — |
| Caught by the sort | — | — |
| Escapes to customer | — | — |
| Recovered by rework | — | — |
| Sort labour | — | — |
| Rework labour | — | — |
| Rework fallout scrapped | — | — |
| Escape cost | — | — |
| Total cost | — | — |
| Good parts shipped | — | — |
| Cost per good part | — | — |
This column carries no expedite freight and no replacement cover: recovered parts fill the hole the sort made. If you would still expedite on top, add that freight to this total by hand before you compare.
Scrap + replace
C = Q×(crepl − s) + freight + exposed days×downtime
| Line item | Low | High |
|---|---|---|
| Pieces replaced | — | — |
| Exposed days | — | — |
| Replacement buy | — | — |
| Less scrap recovery | — | — |
| Expedite freight | — | — |
| Downtime | — | — |
| Total cost | — | — |
| Good parts shipped | — | — |
| Cost per good part | — | — |
Scrap-and-replace does not move with the defect rate: you throw the lot away either way. That flat line is exactly what the sort plan has to beat.
Where sorting stops paying
—
—
What gets past the sort
—
Disclosure: PLI provides sorting and rework services and therefore has a commercial interest in the outcome of this calculation. That is precisely why every line item is on screen and every assumption is yours to change — and why the worked example below is one where scrapping wins and we lose the job.
WHEN SCRAP WINS
AND WE LOSE THE JOB.
A tool published by a sorting company is worth nothing if it never says “do not sort.” Here is a real shape of job we turn down, with the arithmetic in full.
20,000 stamped clips at $1.20 standard, found 40 % defective. Replacements are $1.45 and available. Sort effectiveness is a realistic 80 %, the sort runs at 90 pieces per person-hour on $38 loaded labour, an escape costs $60, and premium freight is $2,500 either way.
Sorting means touching all 20,000 pieces to recover 12,000 good ones, then buying 6,400 replacements for everything the sort pulls out, because the full quantity still has to ship. Rework is worse again: 6,400 caught defectives at four minutes each is 427 person-hours of skilled labour on a part worth $1.20.
Scrap-and-replace costs $1.495 per good part. Sorting costs $9.685. Rework costs $7.078. There is no version of this where we should be on that line, and the honest answer at 7am is quarantine the lot, order replacements, and put the sort crew somewhere they earn their keep.
The tell is the ratio. When the defect rate is high enough that most of what you handle is scrap, you are paying full sort labour on the bad parts as well as the good, and the replacement bill arrives regardless. Sorting protects a mostly-good lot. It cannot rescue a mostly-bad one.
The arithmetic
| Line item | Sort | Scrap |
|---|---|---|
| Defectives (40 %) | 8,000 | 8,000 |
| Caught / escaped | 6,400 / 1,600 | — |
| Sort labour (20,000 ÷ 90 hr × $38) | $8,444.44 | — |
| Escape cost (1,600 × $60) | $96,000.00 | — |
| Expedite freight | $2,500.00 | $2,500.00 |
| Cover for 6,400 pulled × $1.45 | $9,280.00 | — |
| Replacement buy (20,000 × $1.45) | — | $29,000.00 |
| Less scrap recovery (20,000 × $0.08) | — | −$1,600.00 |
| Total cost | $116,224.44 | $29,900.00 |
| Good parts shipped | 12,000 | 20,000 |
| Cost per good part | $9.685 | $1.495 |
Sort + rework on the same lot lands at $7.078 per good part: $123,439.64 across 17,440 good parts, after 5,440 of the 6,400 caught defectives are recovered at 85 % first-pass yield. Break-even for this job is a defect rate of 6.47 %. It is at 40 %.
Press Load the 40 % clip case in the calculator above to run these numbers yourself.
HOW THIS IS CALCULATED.
No black box. Seven formulas, every one of them checkable against the line items above, plus a note on where the defaults came from.
Splitting the lot D = Q × p · G = Q − D
The lot divides into defectives and good parts at the defect rate you entered. Everything downstream is driven off that split, computed twice: once at the low end of your range and once at the high end.
What a sort actually catches caught = D × e · escapes = D × (1 − e)
Sort effectiveness e is never 1.0. A single visual pass at 80 % leaves one defective in five on the truck, and those escapes are charged at your escape cost. This is the single most common way a sort business case is overstated.
Cost of sorting C = (Q÷R)×L + escapes×k₂ + freight + caught×crepl
Labour to touch every piece, plus the cost of what still gets through, plus premium freight, plus replacements for the parts you pulled if the full quantity has to ship. Divided by G, the good parts only — escapes shipped, but they were never good.
Cost of sorting plus rework C = (Q÷R)×L + caught×(m÷60)×Lrw + caught×(1−y)×(c−s) + escapes×k₂
You still pay to sort. On top of that comes rework labour on every caught defective, and the standard cost of the ones rework fails to save, net of scrap recovery. The denominator grows by the parts rework recovers: G + caught × y.
Cost of scrapping and replacing C = Q×(crepl − s) + freight + max(0, lead − days)×downtime
Buy the whole lot again at replacement cost, take the scrap credit back, add freight, and charge downtime for any lead-time days beyond your ship date. It divides by Q, because every replaced part is good. It does not move with the defect rate at all.
The break-even defect rate solve Csort(p)÷G(p) − Cscrap÷Q = 0
Because the sorting cost curve is not a straight line, this is solved numerically by bisection on p across 0 % to 100 %, 50 iterations to a tolerance of 1e−9. Below that rate sorting is the cheaper way to a good part; above it, scrap-and-replace is.
The Deming benchmark k₁ = L ÷ R · p* = k₁ ÷ k₂
Deming’s all-or-none rule compares inspecting every piece against inspecting none. k₁ is the inspection cost of one piece, k₂ the cost of one escape. It is a different question from the one above — it never asks about replacing the lot — so it usually gives a much lower number. It is shown as a sanity check, not as the answer.
Where the defaults come from every field marked ESTIMATE is a placeholder
Sort rate, sort effectiveness, escape cost, rework minutes and rework yield are the five numbers that move the answer most, and all five are starting points rather than facts about your part. Replace them with PLI’s own job data from a previous containment on this part number, or with your plant’s. The tool is only as good as the numbers you put in it.
WHAT THIS TOOL IS NOT.
Not a release or disposition authority
This is a planning and cost-comparison tool. It does not release material, approve a disposition, or substitute for your MRB, your control plan, or your customer’s requirements.
Rework and use-as-is need customer approval
Reworking or using-as-is on customer product typically requires a written customer deviation or concession before anything ships. Price it here; get it approved there.
Not a safety or regulatory judgement
Safety-related characteristics, regulated product and anything already in the field are not cost decisions. Cheapest per good part is irrelevant when the answer has to be containment.
Not neutral, and it says so
PLI provides sorting and rework services and has a commercial interest in this calculation. Every assumption is editable and every line item is shown so you can check us.
Not a single-pass model
The maths assumes one sort pass at the effectiveness you entered. Multi-pass sorting, sample-then-100 % schemes and gauge-based screening all behave differently. Talk to us before modelling those.
Not the whole cost of poor quality
Customer scorecard damage, a new-business hold, PPM impact and engineering time are real and are not in these numbers. They almost always push the answer further away from “ship it and hope.”
GET THE REAL
NUMBERS ON YOUR PART.
The defaults above are placeholders. We can replace the sort rate, the effectiveness and the escape exposure with figures from an actual containment on a part like yours, usually inside a day — and if the answer is that you should scrap it, we will tell you that too.
sales@phillipslightindustrial.com
Lines answered 24 hours a day, 7 days a week.