Articles

Is a longer warranty worth more than a better remedy?

A warranty has three terms: how long it runs, when it starts, and what it pays when a part fails. The negotiation is usually about the first, and on a part that costs a fraction of its failure the first is worth the least.

Larchmere Industrial signed a twelve-month warranty when the rejected bid offered twenty-four, and the book prices the difference at 196,996.34. That figure assumes every covered defect is repaid at the full 890 a field repair costs. If the clause pays only for the part, the same extra year is worth 23,019.80. And moving the remedy from the part to the repair, with the length left at twelve months, is worth 200,336.62: more than the extra year is worth even on the generous assumption.

Chapter 10 makes the point in words: that the trigger date is worth more than several percentage points of length, that a part-only remedy transfers about eleven per cent of the loss. It then prices the duration and neither of the other two. The third case in the companion files takes all three to a number on Larchmere’s own field curve, and the ranking it produces is the reverse of the order in which the terms are usually argued.

The facts, and the curve that prices every option

The term delivers 159,700 units at a defect rate of 0.42 per cent, which is 670.74 expected defects. A failure costs 890 to repair in the field; the part itself costs 104.00. Units take 5.2 months from delivery to commissioning, and Larchmere gives its own customers a 36-month warranty, so a unit delivered today can come back on it 41.2 months from now. The field curve says 38 per cent of defects have surfaced by month 12, 71 per cent by month 24 and 89 per cent by month 36. The case adds two points the book does not give, nothing at month 0 and everything by month 48, and draws the curve straight between points.

Every warranty option is then one reading off that curve: how many months from delivery the cover reaches, what share of defects surfaces inside it, and what each covered defect returns.

OptionMonths from deliveryShare of defects coveredWorth, repair remedyWorth, part only
As signed: 12 from delivery12.038.0%0.000.00
12 from commissioning17.252.3%85,365.089,975.25
As offered: 24 from delivery24.071.0%196,996.3423,019.80
24 from commissioning29.278.8%243,559.1128,460.84
Back to back: 36 from commissioning41.293.8%332,903.9138,901.13
Each option is valued against the signed twelve months from delivery, on 670.74 expected defects. The tail beyond 36 months is ADDED.

Two things are visible before any argument starts. The right-hand column is the left-hand column multiplied by 104 over 890, which is 11.7 per cent: every option on a part-only remedy is worth about an eighth of what it is worth on a repair remedy. And the trigger, which the chapter says suppliers concede because it sounds administrative, is worth 85,365.08 on its own, 43 per cent of the 24-month concession, because starting the clock at commissioning rather than delivery adds 5.2 months of cover at the steep part of the curve.

What the book’s figure quietly assumes

The 196,996.34 in the clause package is the uncovered cost at twelve months, 370,114.33, less the uncovered cost at twenty-four, 173,117.99. Both are defects times one minus the share surfaced times 890. The arithmetic is right. The assumption inside it is that a covered defect is a defect the supplier pays for in full.

On a part-only remedy that is not what cover means. The supplier ships a replacement part at 104 and the buyer still sends the engineer, pays for the downtime and absorbs the rest of the 890. The 33 extra percentage points of defects the second year catches then return 104 each rather than 890, and 196,996.34 becomes 23,019.80.

Put the whole expected failure cost beside it. 670.74 defects at 890 is 596,958.60 over the term. A twelve-month warranty that pays the part returns 26,507.64 of that, 4.44 per cent. The clause reads as a warranty and behaves as a small discount.

The same three terms, ranked from a part-only clause

So start from the weaker clause, twelve months from delivery paying the part, and price each single ask a buyer could make against it.

Single ask, from 12 months paying the partWorth
Remedy: pay the repair, length unchanged200,336.62
Back to back: 36 from commissioning, part only38,901.13
Length: 24 from commissioning, part only28,460.84
Length: 24 from delivery, part only23,019.80
Trigger: 12 from commissioning, part only9,975.25
For comparison: 24 from delivery, on a repair remedy196,996.34
Case three, rows 25 to 28 and row 32. The last line is the book’s own clause cost, which already assumes the remedy.

The remedy is worth 254.88 covered defects, which is 670.74 times 38 per cent, times the 786 of each failure the part does not pay for. It needs no extra month of cover and no change to the trigger. Against it, the longest cover in the table, back to back with Larchmere’s own obligation, is worth 38,901.13 while the remedy stays at the part: less than a fifth of the remedy on its own.

The order to negotiate in is the remedy first, the trigger second and the duration last. It is the opposite of the order in which warranty clauses are usually argued, where length is the term everybody fights over.

How wide the margin is, and when it flips

The finding is honest only with its margin attached. On Larchmere’s figures the remedy beats the 24-month extension, even with the extension assumed to pay the repair, by 3,340.29: 200,336.62 against 196,996.34. That is not a large cushion, and it is worth knowing what moves it.

The remedy is worth the defects covered by month 12 times the gap between repair and part. The extension is worth the extra defects covered between months 12 and 24 times the full repair. Set the two equal and the part price cancels out into a ratio: on this curve the remedy wins whenever a repair costs more than 0.38 ÷ (0.38 − 0.33), which is 7.6 times the part. Larchmere’s repair is 8.56 times its part.

Repair costRepair ÷ partRemedy at 12 months24 months, repair remedyRemedy less length
500.004.81100,932.96110,672.10−9,739.14
700.006.73151,909.20154,940.94−3,031.74
790.407.60174,950.46174,950.460.00
890.008.56200,336.62196,996.343,340.29
1,200.0011.54279,349.80265,613.0413,736.76
Part at 104.00 throughout, the book’s field curve, 670.74 expected defects. The 790.40 and 890.00 rows were recomputed in the case workbook itself.

Held the other way, with the repair at 890, the part would have to cost more than 117.10, about 12.6 per cent above its price, before the extra year overtook the remedy. And the 7.6 is a property of this curve, not a constant: it is the share surfaced by month 12 divided by how much more surfaces in the first year than in the second. A part whose failures arrive later raises the ratio the remedy needs, and once the second year surfaces as many defects as the first, length wins at any ratio.

Two conclusions survive any curve. The first is that no length of cover is worth what the clause cost table says it is unless the remedy pays the repair, because every figure in the left-hand column of the first table assumes it does. The second is that the remedy’s value grows with the gap between the price of the part and the cost of its failure, so it is largest on a cheap part whose failure is expensive.

What it means in the room

Read the definitions before the duration. A clause headed “twenty-four months’ warranty” that pays only for replacement parts is, on this curve, worth 23,019.80 against twelve months, and a buyer who wins the extra year on it has spent negotiating credit on the smallest line in the table. If the supplier will move on only one term, the one to ask for is the repair: on Larchmere’s figures it is worth marginally more than the extra year and it does not depend on the extra year being paid in full.

Then the trigger. Twelve months from commissioning is worth 85,365.08 on a repair remedy, and it is the ask the chapter says suppliers give away because it reads as drafting. The duration comes last, and should be measured against the buyer’s own downstream obligation rather than against market practice: the signed clause leaves a window of 29.2 months in which Larchmere’s customers are covered and Larchmere is not.

Reproducing it in the workbook

Everything above is in Three_Cases.zip, in the file for case three, The trigger, the remedy and the duration. On the sheet 1. The warranty the inputs sit in C5 to C12 and the field curve in rows 15 to 19. Row 21 holds the 670.74 expected defects. Rows 24 to 28 carry each option: months from delivery in column C, the interpolated share covered in D, the uncovered cost on a repair remedy in E, and the worth against the signed clause in F (repair) and G (part only). C31 is the part’s share of a failure, 11.7 per cent; C32 is the remedy on its own, 200,336.62; C34 is the trigger as a share of the 24-month concession, 43 per cent; C36 rebuilds the book’s 196,996.34 as the book rounds it. The Checks sheet ends ALL OK.

To find your own break-even, overwrite C7 with the repair cost or C8 with the part price and compare C32 with F25. The same clause priced without the case around it is the Warranty sheet of The_Clause_Calculators.xlsx, where the part price is now an input in C16 and C24 prints its share of a failure. The sensitivity table and the repair-cost break-even in this article are straight-line arithmetic on those cells.

The workbooks behind this article

Every figure above is a live formula in the free companion files for Contract Management, which also price the other seven clauses, the month-nineteen crossover and the negotiation ledger. Each workbook ends with a Checks sheet setting the printed figure beside the computed one. No account and no email address.

Open the companion files →

Also on this site

This note is drawn from Contract Management. The book is on Amazon.

If this book helped — or didn’t — a few lines on Amazon are worth more than they look: they are what the next reader goes on. Write a review. The workbook stays free either way.