What do you do when a client refuses to pay the final invoice?

Do the procedural things first: ask for the complaint in writing with photographs, send a plain demand letter, find your jurisdiction's lien deadline before it passes. Then answer the harder question, which is whether you're willing to lose this client to collect. If you aren't, the term was never enforceable, and a better contract wouldn't have changed that.

Why does the advice stop working right at the moment you need it?

Search any version of this and you'll get the same page seven times. Get it in writing. Send a demand letter. Register the lien before the deadline. Escalate to small claims. Every step is correct. I have no quarrel with any of it, and if you're standing in the middle of this right now, do those things.

The trouble is that the whole ladder assumes a contractor who will climb it.

This week a contractor put a version of this question to a room of other contractors online. Four thousand dollars signed, nineteen hundred taken as a deposit, eleven hundred still owing. The client came back with a list of alleged deficiencies from a realtor and offered a thousand to make it go away. The room gave him the correct answer, in detail, more than once. Demand the list in writing. Photograph the work. Register the lien, because a condo can't sell over top of one.

He turned all of it down. In his own words, "the idea of fallout always concerns me. Especially in todays age of FB posts and Google reviews."

He's being accurate about his position, which is a different thing from being weak. And the entire published answer set for this problem was written for a business he doesn't have.

What is the term actually worth if you won't test it?

Some numbers, because this behaviour is not rare and it is not personal.

In the industry-wide payment data, only 12% of construction businesses report always getting paid on time, and only 15% report always being paid in full (Levelset/Procore, 2022 Construction Cash Flow & Payment Report). 82% of contractors now wait more than 30 days for payment, up from 49% two years earlier (Rabbet, 2024 Construction Payments Report). Slow payment is estimated to have cost the U.S. construction industry $280 billion in 2024, roughly 14% on top of total construction spending.

And here is the part that names the mechanism. Contractors inflate their bids by an average of 8% to protect themselves against slow payment, and 97% of general contractors raised bid prices in 2024 for the same reason (Built, 2025 Construction Cash Crunch Survey; Rabbet 2024). Nobody collects the unenforceable term. Everybody pays for it, forever, as a permanent 8% surcharge on the whole market.

So the cost is real, it is measurable, and it has already been passed to your customers. What it has not done is buy you the ability to hold a single line.

A working definition, since the word "term" is doing a lot of quiet work in these conversations: a term is enforceable when the cost of the client walking away is lower to you than the cost of letting the term go. That's the whole test. It has nothing to do with the wording, the font, or whether anyone initialled the page. It's an arithmetic problem about replaceability, and you can run it in about ten seconds on any clause in your agreement.

Run it on your deposit policy. Run it on your change-order clause. Most owners find, honestly, that they're carrying four or five terms they've never once tested, and one they'd fight for.

Why won't he file the lien he's entitled to file?

Because he believes the review is worth more than the money.

The belief has something behind it. 71% of consumers read reviews regularly when they're browsing local businesses (BrightLocal, Local Consumer Review Survey 2025), and for a local trade with no distinctive position that profile is most of the top of the funnel. So look at the arithmetic he's solving, because it isn't the one the escalation ladder solves.

Contract: $4,000. Owing: $1,100. Legal remedy: available. Amount he will collect: $1,000, and only if the client keeps their word. The $1,100 he walks away from is 27% of the contract value, paid as a reputation tax.

Now the part I want to push on, because I think the fear is bigger than the thing it's afraid of.

You should never be this frightened of one bad Google review. I'm not sure I trust a company with nothing but five stars. Reviews can be bought, profiles can be curated, and a wall of unbroken praise reads to me as a company that has either never been tested or has quietly managed the evidence. An honest profile has a bad one in it. What tells you about a company's character is how they answered it, and a calm, specific, non-defensive reply to an unfair one-star does more for a buyer than the four-point-nine average ever did.

So the man is paying twenty-seven percent of a contract to protect an asset he has misjudged in two directions at once. He's overrating the damage of the review, and he's underrating the response as the place where the trust actually gets built.

That's the first correction. The second one is bigger.

Where was this actually decided?

Long before the invoice. Usually before the quote.

The reason he can't afford the review is that he has no position. He is a good carpenter who does good work for whoever calls, which means every future job comes from the same undifferentiated pool of people comparing three names on a screen, which means the profile is load-bearing, which means any client with a grievance holds a hostage. What he had written down were preferences he'd had printed, and the first client to test one found that out in about four seconds.

The uncomfortable version is that this is what "we serve everybody" costs, and the bill arrives at the worst possible moment, at the end of a job, in a fight with somebody unpleasant.

Contrast that with a business that has a position. A restoration contractor who does one thing for one kind of client, whose next three months are booked from a source that isn't a search result, can afford to say: send me the written deficiency list, I'll make the corrections that are real, the balance is due on the fifteenth, and if we can't agree I'll register the lien on Thursday. Same clause. Same law. Completely different sentence, because the man saying it can afford to be told no.

This is the same structure I've written about in why you keep having to defend your price on every job, and in why the flinch before billing for extra work is a positioning failure. Pricing power and enforcement power are the same power. They come from the same place, and they arrive together or not at all. In our four frameworks this looks like a Sales problem and lives in Positioning.

Which is where I'd normally stop, and where I'd be about half right.

The clause I've never once used

There is a term in my own contract. If a client misses an invoice, the next one carries a surcharge. It's written down, it's clear, and the client agreed to it.

In six years I have never enforced it. Not once.

I don't have a demand problem. I'm not easy to replace, I'm not scrambling for the next client, and by the argument I just made I should have been able to use that clause without breaking a sweat. I didn't use it, and the reason had nothing to do with my position in the market. I was being lenient because being lenient was more comfortable than the conversation, and I told myself a story about grace that was mostly a story about not wanting to be disliked.

So take the positioning argument as necessary and not sufficient. It buys you the ability to hold a term. It does not make you hold it. What makes you hold it is closer to leadership than to strategy, and a fair bit of it is just knowing what you're worth.

I still don't think I've ever quoted a price without my heart fluttering. Some part of you second-guesses whether you're worth what you're asking, and that's a human thing rather than a marketing failure, but it's the same part that decides not to send the surcharge, not to make the call, not to mention the balance. Business is essentially leadership, and a leader's job is fixing problems. When you stop being willing to name them, you've stopped doing the job.

Fear-based people-pleasing is not worth it. If what you're doing cheats you, or cheats your customer, or exists because you're afraid, that's a sign, and the amount of money involved doesn't change the reading. Is a hundred dollars worth the conversation? That's a real question with a hard answer, and you still have to answer it.

I'm fixing mine. It has taken me longer to grow up about this than I'd like to admit.

Does the power ever run the other way?

It does, and the same data set shows it plainly.

67% of subcontractors have declined to bid on a project because of the general contractor's slow-pay reputation, and 100% of subcontractors say they consider a GC's payment history before bidding (Rabbet, 2021 and 2024 Construction Payments Reports). That's an enforceable term. Nobody wrote it into a contract. Nobody sent a demand letter. A group of trades simply became willing to lose the work, and the market repriced.

The public example worth studying is Carillion, the UK construction and services group that collapsed in January 2018. Carillion was a signatory to the government's Prompt Payment Code, which asks for payment inside 60 days. Its actual standard terms for suppliers were 120 days, and it ran a supply chain finance scheme under which a supplier could get paid at 45 days by selling the invoice to Carillion's bank at a discount, while Carillion itself didn't have to settle until the 120 days ran out (UK Parliament research briefing, The collapse of Carillion).

Read that again as a positioning story rather than a scandal. Carillion had a written commitment to 60 days and ignored it for years, because the code had no teeth and the suppliers needed the work. The suppliers had contracts too. Thousands of them, presumably well drafted, and the terms in them were worth roughly nothing, because the businesses holding them could not afford to lose Carillion. When Carillion went under it took a large part of that supply chain with it, which is the final bill for a term nobody could enforce.

Same mechanic, at both ends of the market, in two countries, with a decade between them. Whoever can most easily walk away sets the terms. Everything else is stationery.

So what makes a term real?

Five things, in this order, and only the last one is paperwork.

1. Demand you didn't buy. Work arriving from referral, reputation and repeat clients rather than from an ad account. Bought demand stops the day you stop paying, so it can never underwrite a refusal.
2. A specific enough position that you're hard to swap. Not better. Harder to replace for a particular kind of client with a particular kind of problem. Being the only sensible option for someone is worth more than being an option for everyone.
3. A written refusal you've actually used. One client type you turn down, one job type you decline, one term you hold without exception. If you can't name all three off the top of your head, you don't have terms yet, and the honest first step is choosing one and holding it once.
4. A trigger that takes the decision away from your mood. Mine is a pay period. If an invoice is still unpaid when the next one comes around, I'm having the conversation. Not a threat, not a letter from anybody's lawyer, a conversation, and the point of it is transparency rather than pressure. If a client is genuinely having a rough stretch, seasonal work or a slow quarter, that gets taken into account, and it gets taken into account out loud rather than by me quietly absorbing it and resenting it. What the trigger removes is the daily judgment call about whether today is the day I feel brave.
5. The clause. Which now has something to catch on.

Founders reverse this order constantly. They rewrite the contract, add three clauses, buy a template, and then discover the following spring that the new document performs exactly like the old one, because the constraint was never in the document.

I had a call yesterday with a company I barely know, the founder and their CRO. At the end of it they thanked me for asking questions that people who like them don't usually ask. That was harder for me than it sounds. I'm Canadian and I'm polite, and pushing back is not my natural setting. But accepting the status quo is not the job, and that holds for a client's strategy and for a client's unpaid invoice in exactly the same way.

Putting it to work

Before you touch your contract, do this.

Run the test on five clauses. Take your deposit terms, your change-order clause, your payment schedule, your scope boundary and your cancellation policy. For each one, answer honestly: if a client refuses this, will I lose the client over it? Write yes or no. No explanations.

Count the yeses. That's how many terms you have. The rest are preferences, and knowing which is which is worth more than another clause.

Name your three refusals. One client type you turn away. One job type you decline. One term you hold without exception. If any of the three is blank, that's the gap, and it's a positioning gap, not a legal one.

Price the reputation tax. Go back through the last two years and add up what you've written off, discounted under pressure, or absorbed to keep a profile clean. Put a real number on it. Most owners have never totalled it, and the total is usually larger than the marketing spend it was protecting.

Set a trigger, not an intention. Pick a moment that isn't a mood. An unpaid invoice still outstanding at the next pay period, an overdue balance at thirty days, whatever fits how you bill. When it hits, you have the conversation. Adjust for a client genuinely in a hard stretch, out loud, and never by silently eating it.

Then fix the immediate one. For the invoice sitting on your desk right now: get the complaint in writing with photographs, send one plain demand with a date on it, find out your lien deadline today, and get advice from a lawyer in your jurisdiction before the deadline rather than after. Deadlines are short and unforgiving, and they vary enormously by province and state.

Adapt all of it. A remodeller with a four-thousand-dollar balance and a consultancy with a net-90 enterprise client are running the same arithmetic at different scales.

Sources

• Construction Payment Statistics 2026: Slow Pay, Lien Filings, and Cash Flow, DocJoist, March 2026 — compiling Rabbet 2021/2024 Construction Payments Reports, Levelset/Procore 2022 Construction Cash Flow & Payment Report, and the Built 2025 Construction Cash Crunch Survey
• The collapse of Carillion, UK Parliament research briefing CBP-8206
• Local Consumer Review Survey 2025, BrightLocal

Frequently asked

Should I file a lien when a client refuses to pay?

If you're going to collect, usually yes, and the deadline is the thing that catches people out. Lien rights are time-limited, the windows are short, and they differ by province and state. Find out your deadline the week the dispute starts and talk to a lawyer where you operate. Treat this article as context, not advice.

What if the client claims the work is deficient?

Ask for the list in writing, with photographs, itemised. A real deficiency list is specific and can be corrected. A negotiating tactic tends to arrive vague, second-hand, and attached to a number. Requesting it in writing costs you nothing and tells you which one you're holding within about a day.

Is it ever right to take the discount and move on?

Yes, and I'd rather you take it deliberately than pretend it was forced on you. Sometimes the client is genuinely dangerous to your reputation, sometimes the balance is small enough that your time is worth more, and settling under a full and final release is a legitimate business decision. What isn't legitimate is calling that a policy and then writing terms you have no intention of using.

Won't a stronger contract prevent this next time?

It helps at the margins, mostly by making the conversation clearer earlier. It won't produce the ability to enforce. If a client works out that you can't afford to lose them, the contract becomes a description of what should have happened.

What about the bad review, though?

Answer it. Calmly, specifically, without defending yourself, in public. I'm wary of a company with nothing but five-star reviews, because reviews can be bought and profiles can be curated, and an unbroken wall of praise tells me either nothing went wrong or nothing was left visible. A fair reply to an unfair review does more for the next buyer than a perfect average does.

How do I get to the point where I can hold a term?

Build demand that doesn't come from an ad account, get specific enough that a particular kind of client can't easily replace you, and then practise refusing something small before you have to refuse something large. The position gives you the room. Using it is a separate skill and it takes reps.

Ready to look at the architecture honestly?

If the last three disputes all ended with you absorbing the cost, the contract isn't the constraint. Book the conversation and we'll tell you what we see, including whether this is a positioning problem or a genuinely bad client. Have a look at how we work first if you'd rather.

Book the conversation →