How to Write an Exclusion the Customer Can Actually Use
Why this matters
Most exclusions in circulation do nothing. They were copied off another shop's form, they exclude a category rather than a thing, and they say what you are not responsible for without saying what happens instead. When the condition actually turns up, the customer reads the line, does not recognise their situation in it, and the argument proceeds as though the line were not there. You did not buy protection. You bought a sentence.
An exclusion that works is a small machine with four moving parts. This piece is about building one, by writing the same clause four times and watching each version fail.
Before the commercial question: if the condition you are excluding is a life-safety hazard you found, the duty to say so plainly and in writing comes first and does not depend on whether anyone is paying you to fix it. You can decline to do the work. You cannot quietly exclude a hazard and walk away without telling the customer what you saw. See related: Documenting the Unsafe Condition You Found.
The four tests
Every exclusion has to pass all four. Missing one is enough to make the line decorative.
- Observability. The excluded condition is described so a named person could look for it, at a stated time, and say yes or no. "Pre-existing conditions" fails. "Corrosion at the joints visible where the covering has been removed" passes.
- Consequence. The clause says what happens when the condition is found, not only that you are not responsible. A disclaimer with no path leaves both parties standing in the crawlspace with no agreed next move.
- Specificity of the thing. It names the component, the location, or the material. Category-level exclusions read as boilerplate, and boilerplate is what a customer skips.
- Timing. It was in front of the customer before they accepted, and it was legible. An exclusion produced after the discovery is not an exclusion, it is an excuse.
State law decides how far an exclusion or a limitation of liability can go, and the answer differs by state and differs again depending on whether your customer is a consumer or a business. Several states also impose specific format, disclosure and signature requirements on residential home improvement contracts. Have your own attorney review the language you intend to use repeatedly; this piece is about making the clause do its operational job, not about how far it reaches legally.
Draft one: the clause almost everybody has
Contractor is not responsible for pre-existing conditions.
Run the tests. Observability: nothing to look for. Consequence: none stated. Specificity: none. Timing: fine, it is on the form. One of four.
What actually happens when the condition appears: the customer asks what pre-existing means, you explain, and they say reasonably that the whole house is pre-existing and they hired you to deal with what is there. The clause has excluded everything and therefore excluded nothing. This is the most common exclusion in the trades and it is worth exactly nothing in a conversation.
Draft two: name the thing
Contractor is not responsible for the condition of concealed piping within existing walls or floors.
Observability: better, though still nothing that says who looks and when. Consequence: still missing. Specificity: good, this is a thing. Timing: fine. Two of four.
What happens now: you open the floor, the concealed run is corroded through, and the customer accepts that you did not cause it. Then they ask what you are going to do about it. You have no agreed answer, the floor is open, and the job is stopped while you improvise a price under pressure with the customer watching. You won the responsibility argument and lost the day.
This is the draft most careful shops stop at, and it is why so many shops believe exclusions do not help much. It settles blame and leaves the process undefined.
Draft three: add the consequence
Contractor is not responsible for the condition of concealed piping within existing walls or floors. If unserviceable piping is found, work stops and the additional work is quoted separately.
Observability: still weak on who and when. Consequence: now present and specific. Specificity: good. Timing: fine. Three of four.
This version does real work. But watch it fail on observability: the technician opens the floor, sees pitting, and calls it unserviceable. The customer's brother-in-law says it has been like that for twenty years and is fine. Nothing in the clause says who decides or what evidence settles it, so the argument moves to competence, which is a much worse argument to have than a scope argument.
Draft four: the working clause
Concealed piping. This quote covers only the piping shown on the attached sketch. Piping concealed within existing walls or floors has not been inspected and is not included. If, on opening, concealed piping shows through-wall corrosion, active seepage, or a joint that will not hold a pressure test at the system's normal operating pressure, work on that section stops, the condition is photographed with a scale reference and sent to you the same day, and the remedial work is quoted before it proceeds. You may accept the quote, have it done by others, or direct us to reinstate the opening and leave the section as found.
All four tests pass. Three named, observable findings anyone can check. A consequence with a defined process. A named component and location. It goes out with the quote, so it is in front of the customer before acceptance.
Note the last sentence. Giving the customer a genuine third option, including one where you do not get the extra work, is what stops the clause reading as a mechanism for generating change orders. A customer who believes the exclusion exists to sell them something will fight it. A customer who is offered an exit will usually not take it, and will trust the finding.
Note also what the clause does not claim: it does not say the piping is fine, and it does not say you inspected it. Do not write an exclusion that implies an inspection you did not perform, because it hands the customer the argument that you should have caught it.
Worked example: what the clause is worth over a year
A shop tracked one job type across a year: 30 jobs, all involving a tie-in through an existing floor or wall.
- Concealed condition found on 4 of the 30 jobs, which is 13 percent.
- Before the clause, each of those 4 absorbed an average of 3.5 hours of unsold work: stopping, documenting, improvising a fix or a partial reinstatement, and the schedule damage on that day. That is 14.0 hours a year absorbed, on 4 jobs.
The year after the clause went on every quote for that job type, the same 4 in 30 hit the condition. On all four, work stopped and the condition was documented, which took about 0.6 hours each. Three customers accepted the remedial quote, so on those three the 0.6 hours of documentation was inside quoted work and the remedial hours were sold. One customer declined and had the section done by their own plumber, so the shop absorbed 0.6 hours there.
Annual absorbed time went from 14.0 hours to 0.6 hours. Both figures are on the same basis: the 14.0 already included the same stop-and-document time, so this is not a corrected figure benchmarked against an uncorrected one.
Read what actually moved. The clause did not reduce the frequency of the condition, which stayed at 4 in 30. It did not make the remedial work cheaper. It converted 14.0 hours of unsold work into sold work by putting the decision in front of the customer at the only moment when they still had choices. That is the whole mechanism, and it is why the consequence sentence matters more than the disclaimer sentence.
The counterfactual worth naming: if the shop had written draft two instead of draft four, the finding still gets made, but the negotiation happens with the floor open. Under that pressure the realistic outcomes are splitting the cost or absorbing it, and this shop's own history says which one it picked, because that is what the 14.0 hours were.
The failure mode that survives a good clause
A clause this specific fails one way: nobody in the field knows it exists. The estimator writes it, the technician has never read the quote, opens the floor, sees the corrosion and just fixes it because fixing things is the job. Now you have performed unquoted work and, worse, altered the evidence.
Two controls. First, the exclusions ride on the work order the technician actually opens, not only on the customer's copy of the quote. Second, the stop condition is trained as a stop condition: photograph, do not remediate, call. A technician who improves the situation before anybody has seen it has destroyed the only proof that the condition pre-existed your work.
How to verify your exclusions are working
Pull the last ten jobs where a concealed or adjacent condition showed up. For each, answer: was there an exclusion covering it, did anyone in the field know, did work stop, and was the extra sold or absorbed? You are not looking for a perfect record. You are looking for which of the four tests your clauses keep failing, because it will be the same test every time. Shops that copied their form fail observability and consequence together. Shops that write their own usually pass those and fail timing, because the good language lives in a terms document nobody sends with the quote.
References
- The enforceability of exclusions and limitations of liability, and any format or disclosure requirements for residential contracts, are set by state law and differ for consumer and commercial customers; have your own attorney review recurring clause language
- See related: The Conditions Clause: Protect the Quote; Documenting the Unsafe Condition You Found; Documenting a Mid-Job Discovery to Protect the Re-Quote