What a Guarantee Costs You That a Warranty Does Not

Why this matters

A warranty and a satisfaction guarantee look like the same kind of promise on a truck wrap. On the books they behave nothing alike. A warranty triggers on a defect you can inspect, inside a term you set, which means last year's failure rate predicts next year's cost. A guarantee triggers on a customer's judgment, which no inspection resolves and no failure history predicts. Shops add the second one to win jobs against a competitor and price it as though it were the first. This is what that actually costs, and what to cap instead of arguing.

The distinction that matters is the trigger, not the word

A sibling article covers the definitional split between the two promises. This one is about the consequence of that split for money, and the consequence turns on one property: is the trigger observable by someone other than the claimant?

• A workmanship warranty triggers on a defect in your installation. Two competent techs looking at the same joint will usually agree it leaks. Observable.
• A parts warranty triggers on a component failing inside its term. Observable, and usually adjudicated by somebody else entirely.
• A satisfaction guarantee triggers on the customer not being satisfied. There is no second opinion. The claimant is also the judge.

Everything downstream follows from that. An observable trigger has a rate. A rate has a distribution. A distribution can be reserved against and priced into every job. An unobservable trigger has none of those things, because the thing generating claims is not the equipment, it is expectation, and expectation is set by your salesperson rather than by physics.

Note the other end of that range, because it is the useful part: a guarantee whose trigger is written to be observable behaves like a warranty. "We arrive inside the window we promised or the trip charge comes off" is a guarantee by name and a warranty by mechanism, because dispatch timestamps settle it and nobody's feelings are evidence. If you want a guarantee you can reserve for, that is the shape to write.

The case

A ten-person residential mechanical shop, roughly 480 completed jobs a year, carried a 12-month workmanship warranty and nothing else. Their claim history was boring in the way good data is boring: warranty return rates of 3.1%, 3.6% and 3.4% of completed jobs over three years, averaging about 3.4%, with corrections averaging 2.2 hours on site.

That is a reservable number. 3.4% of 480 jobs is about 16 claims a year, and 16 claims at 2.2 hours is roughly 36 hours of unbilled correction. Spread across 480 jobs that is 0.075 hours per job sold, so they carried 0.1 hours per job in their pricing and slept fine. Three years of history sitting inside a half-point spread is what let them do that.

Then a competitor started advertising a satisfaction guarantee and they matched it, in five words on the estimate: if you are not happy, we make it right. No scope limit, no remedy limit, no window.

What arrived

Year one under the guarantee, on the same 480 completed jobs, return visits went from 16 to 41. The split:

Read the first row before the second. The warranty side did exactly what three years of history said it would: 15 claims against a predicted 16, at the same 2.2 hour average, for 33.0 hours against a reserve of 36. The promise they could inspect was, if anything, slightly better than reserved.

The entire overrun is the second row. 88.4 unreserved hours, which is 2.7 times the 33.0 hours the warranty cost them in the same year. Both figures are unbilled technician hours on site, so that ratio compares like with like. It does not say the shop lost 88.4 hours of profit, it says 88.4 hours of field capacity went to work nobody had priced.

Breaking the 26 satisfaction claims down is where the lesson is:

• 9 were expectation gaps set at the sale. The system did what it was designed to do and not what the customer had been led to hope. A back bedroom still ran warmer than the front room after a replacement, because the duct that serves it was never the problem and nobody said so out loud during the sale.
• 11 were finish, appearance and noise items. Line sets routed where the customer did not expect, a condensate line crossing a sightline, a unit audible from a bedroom.
• 6 were genuinely out of scope. The customer wanted something else than what they bought.

Nine plus eleven plus six is 26. Not one of those 26 is a defect. Every one of them is real to the customer, which is why arguing cause on a satisfaction claim is unwinnable: you are right and it does not matter, because the promise you made did not mention cause.

Two of the 26 were noise complaints on combustion appliances. On those returns the tech carries a personal CO monitor, switched on before entering the equipment room and worn for the whole visit, because verifying a noise complaint means running the appliance and a running appliance is the condition under which a flue or heat exchanger defect actually shows itself. If the monitor alarms, everyone leaves the building, the appliance is shut off at its service switch on the way out only if the switch is on your path out, and nobody re-enters until the space has been ventilated and the appliance has been proved sound.

What they capped, and why those three

They did not withdraw the guarantee. They bounded it in three places, and the three are chosen because they are the only three attributes of a claim you control once the trigger is unobservable:

Scope cap. The guarantee applies to the work described on the agreement. It does not extend to the performance of systems, components or conditions not included in that scope. This is the cap that answers the 9 expectation gaps, and it only works if the agreement actually names what is excluded. A guarantee is only ever as bounded as the scope it sits on top of.

Remedy cap. The remedy is correction of the work, or a credit toward correction, at the shop's election. Not removal and refund, not replacement with different equipment, not consequential losses. This is the cap that keeps a single claim from being unbounded, and it matters most on the 6 out-of-scope claims where the customer wants a different outcome rather than a corrected one.

Window cap, and this is the one shops leave out. The guarantee runs 30 days from completion. The workmanship warranty still runs 12 months. Those two numbers differ on purpose: a satisfaction judgment is formed early, while a workmanship defect surfaces late, often on the first heavy load of the opposite season. A satisfaction window as long as the warranty term is not generous, it is undefined.

The unit of analysis on all three is the individual claim, and they combine as AND: a claim must be inside the scope, inside the window, and satisfiable by the capped remedy. Fail any one and it is handled as a commercial decision rather than as an obligation, which is a different conversation and a fine one to have.

What happened next, honestly narrated

Year two, same 480 completed jobs, so the two years compare without adjustment: 15 warranty claims at 2.2 hours (33.0) and 12 satisfaction claims at 1.8 hours (21.6), for 54.6 total hours against the prior year's 121.4. That is a reduction of 66.8 hours, or 55%.

Do not credit that entirely to the three caps. The shop changed something else at the same time: they started naming the thing the new equipment would not fix during the sale, out loud, and writing it on the agreement. Nine of the prior year's 26 claims were exactly that failure. Two changes ran together and only a shop that had staged them a year apart could say which did more. Say that plainly rather than attributing the whole 55% to the clause, because the temptation to credit the document is how a shop ends up with excellent paperwork and the same conversation.

What can be said cleanly is narrower and still useful: the satisfaction average fell from 3.4 to 1.8 hours per claim, and the remedy cap is the only change that could move average hours per claim rather than claim count.

The legal shape, which is not yours to resolve

Written promises to consumers carry rules you do not get to write. The federal Magnuson-Moss Warranty Act governs written warranties on consumer products, requiring among other things that such a warranty be designated as full or limited and be available to the buyer before sale, and restricting your ability to disclaim implied warranties on a product for which you have given a written warranty. Implied warranties themselves arise under Article 2 of the Uniform Commercial Code as enacted by your particular state, generally reaching goods rather than services, with services usually governed by separate state doctrines. Several states further restrict disclaimers in consumer transactions and some regulate home improvement contract terms directly.

Which of those reaches you depends on whether you are selling goods, services or both, and on whether your customer is a consumer or a business. That is a question for your own attorney reviewing your actual agreement, not something to settle from a reference card. What this card can tell you is the engineering: the caps above are commercial bounds on a promise, and a commercial bound does not survive a legal rule that says otherwise.

How to verify you got this right

• Can you state your warranty claim rate as a percentage of completed jobs, with three years behind it? If not, you do not have a reservable warranty either, and the guarantee is not your first problem.
• Pull your last ten guarantee claims and mark each against scope, window and remedy. If more than a couple fail one of the three and were honoured anyway, the caps exist on paper only.
• Compare average hours per satisfaction claim against average hours per warranty claim. If satisfaction claims are running longer, your remedy is uncapped in practice regardless of what the clause says.
• Read the last five sold agreements and find the sentence naming what the work will not fix. If it is not there, your scope cap has nothing to attach to.

References

• Federal Trade Commission, Magnuson-Moss Warranty Act (15 U.S.C. ch. 50), businessperson's guide to warranty law
• Uniform Commercial Code Article 2, as enacted by the state whose law your agreement says governs
• See related: The Difference Between a Warranty and a Guarantee
• See related: Setting Warranty Terms You Can Actually Afford to Honor
• See related: The Warranty Reserve Most Shops Never Set Aside