Why Contractors Go Out of Business

01 — The baseline

The failure rate, stated honestly

The numbers everyone repeats are unverified. The real ones are more interesting.

You have heard that 80% of contractors don't make it to year three and that 95% are gone by year five, usually attributed to the BLS or the SBA. Treat that carefully. No federal dataset publishes a survival curve for roofing, or decking, or bathroom remodeling. We checked directly: the Bureau of Labor Statistics publishes establishment survival at the two-digit sector level only. There is no federal table for roofing. Anyone showing you one built it themselves.

Here is what actually exists, and it's better than the folklore because it's specific.

Federal data — all construction

BLS Business Employment Dynamics, Table 7, NAICS 23 Construction; data through March 2025, table updated January 6, 2026. The 48.3% average is our own calculation — an unweighted mean of cohort survival rates. BLS does not publish it. Cohorts opening 2021–2024 are excluded because they have not yet reached year five.

So: somewhere between a third and six in ten survive five years, depending heavily on when you opened. That's a real range and the range itself is the finding — the year you started matters enormously, and it isn't a thing you control.

Trade-level data — and this is the part worth reading twice

Harvard's Joint Center for Housing Studies did what BLS doesn't: it broke five-year failure rates out by trade, using Census Bureau business tracking data.

Abbe Will, Joint Center for Housing Studies of Harvard University, "Five-Year Failure Rates for Remodeling Contractors Exceeded 50% During Downturn," February 10, 2016. Source data: U.S. Census Bureau Business Information Tracking Series, 2007–2012 cohort. Limitation stated by JCHS: BITS covers payroll businesses only, so a contractor who drops to self-employed status is counted as a failure.

Note the spread. A remodeler was substantially more likely to fail than a roofer over the same window — and that gap is not about installation skill. It's about how the work gets sold.

Then size, which is the sharpest cut in the whole dataset. Among remodelers with $100,000 or less in receipts, JCHS found seven in ten were gone within five years. Among those above $5 million, roughly one in four. And this isn't purely a recession artifact — a separate JCHS working paper found remodeling firms that started in 2003, well before the crash, still failed at over 50% by 2007.

Almost none of these companies died because they couldn't build the thing they sell.

02 — Cause of death

What actually kills them

The recurring causes across every trade post-mortem, and the one thing they share.

Post-mortems on failed trade contractors land on the same short list: cash ran out before profit showed up, jobs got priced without the overhead in them, nobody tracked the leads or followed up, the owner stopped running the business and went back to running jobs, the winter was slow, the crew left, the reviews went sideways.

Read that list again and notice what isn't on it. Craftsmanship isn't on it. Nobody goes under because they installed the flashing wrong. Every item sits downstream of one question:

So the real question isn't why contractors fail. It's why a contractor ends up buying every job he gets.

03 — The diagnosis

The price shopper is not a market condition

Homeowners consistently rank trust first. If yours rank price first, something in the interaction put it there.

This lands badly with anyone who got beaten on price this month. But the pattern shows up across separate studies, in separate trades, run by separate organizations.

The cleanest one covers renovation broadly — kitchens and bathrooms, not roofs — and it asked homeowners what was hard about the process rather than what they claim to value, which is a better question:

Houzz, "What Homeowners Want During Home Renovation Projects," July 2024, n=2,866 U.S. homeowners planning, undertaking or recently completing kitchen or bathroom renovations. platform-published — Houzz sells contractor software; the sample is Houzz users, not a general population. Also reported: 41% wanted a clear project timeline and 40% wished for better communication from the pro they hired.

Trust first. Estimate clarity second. Communication third. Price does not lead. And the roofing-specific studies land in the same place:

Weigh the vendor studies accordingly — three of the four above are published by companies selling something the finding endorses. But the Houzz result is large, independent of the roofing trade, and points the same direction as the rest. Four studies, four different sponsors, one answer.

Now hold that against a rep whose appointments keep collapsing into price comparisons. Both things can't be true about the same market. Either every one of those studies is wrong, or something specific is happening inside that rep's appointments that strips out every buying criterion except one.

When a homeowner can't evaluate anything else, price is the only thing left to evaluate.

04 — The mechanism

The trust equation

Why one variable cancels out twenty years of expertise.

In The Trusted Advisor, David Maister, Charles Green and Robert Galford set out a model of how buyers judge professional service providers. It's a conceptual framework, not a measured formula — but it's the most useful frame available for what happens on a homeowner's driveway.

• Credibility is your words. Can you be believed?
• Reliability is your actions. Do you do what you said, when you said?
• Intimacy is emotional safety. Does this person feel safe being honest with you?
• Self-orientation is your focus. Are you centered on your outcome, or theirs?

The structural point is where that last variable sits. Self-orientation is a divisor, not a subtraction. It doesn't cost you a few points off the top — it scales down everything above it at once. The first three build trust additively when you strengthen them. Self-orientation works against all three simultaneously.

• Same rep · Low self-orientation

  22

  Expertise, follow-through and rapport, divided by a genuine focus on the homeowner's outcome.

• Same rep · High self-orientation

  5.5

  Identical expertise. Identical follow-through. Divided by a visible agenda. The credentials didn't change. The number collapsed anyway.

Illustrative only. The trust equation is a conceptual model, not a measurement instrument, and no study attaches values to it. The point is the shape of the math.

How fast this gets decided

Faster than anyone is comfortable with. In a Princeton study, participants shown a photograph of a face for one tenth of one second formed trait judgments that correlated closely with judgments made by people given unlimited time to look. Of the five traits tested, trustworthiness showed the strongest correlation — r = .73 at 100 milliseconds. More looking time increased people's confidence in the judgment. It didn't much change the judgment.

Willis, J. & Todorov, A., "First impressions: Making up your mind after a 100-ms exposure to a face," Psychological Science 17(7), 2006; trustworthiness experiment n=24. Limitation: static photographs in a laboratory. It does not measure trust formation across a live conversation on a driveway.

Homeowners don't run the equation. They feel the output of it, early, and afterward they describe it in the vaguest language available — "something about that guy," "he seemed like he was in a hurry," "I just felt better about the other one." That vagueness is exactly why the feedback loop never closes. Nobody tells you that you lost on self-orientation. They tell you the other guy was cheaper.

05 — The product

In all three trades, the part that fails is the part nobody can see

That isn't a coincidence. It's why these exist as trades.

A homeowner can evaluate a paint job from the doorway. They can evaluate a countertop with their hand. They cannot evaluate any of the following, and they never will:

• Roofing

  Nobody climbs their own roof. Every claim you make — hail bruising, granule loss, flashing separation, soft decking — is taken entirely on your word.

  The failure lives in the attic and under the shingle. It shows up as a stain on a ceiling three years later.

  They never go up
• Decking

  Nobody crawls under their own deck. The boards look fine from above right up until the day they don't.

  The failure lives in the ledger connection, the joist hangers, the footings — the structural attachment to the house nobody has ever looked at.

  They never go under
• Bathrooms

  Nobody sees behind their own tile. The surface can be immaculate while the assembly behind it has been failing for a decade.

  The failure lives in the pan, the waterproofing, the subfloor and the joists — usually not discovered until demolition day.

  They never see behind

This isn't a rhetorical device. It's written into law.

Florida Admin. Code R. 61-30.805 (Standards of Practice, Roof Covering); Massachusetts 266 CMR 6.05 (General Limitations and Exclusions of the Home Inspection); InterNACHI Home Inspection Standards of Practice.

What that concealment costs, by trade

On the deck side, the most-quoted number in the industry — that 90% of deck collapses trace to the ledger connection — deserves an honest label. It comes from Frank Woeste, professor emeritus at Virginia Tech, and his actual published wording is "I believe 90% of all deck collapses result from the failure of the deck ledger-to-house connection." That's a credentialed expert's estimate, not a measurement, and no sample size was ever disclosed. It gets restated as a hard statistic constantly, including by the industry's own trade association, with no source attached. We're telling you where it came from so you can decide what it's worth.

Which brings us to the consequence.

They don't evaluate the work. They can't. They evaluate the inspection — because it's the only sample of your work they get before they have to decide.

The inspection is not a step before the sale. In a trade where the product is concealed, the inspection is the only visible product. It's where all four trust variables get measured at once, in one visit, by someone every previous contractor has trained to look for the tell.

Read the RAIN finding carefully, because it's the whole argument in one line. The winner is the person who taught the buyer something about their own house that they didn't already know. There is exactly one place in any of these three sales where that can happen — on the roof, under the deck, behind the bathroom wall.

You cannot teach a homeowner something you didn't go and find out.

What this looks like when it's real

A homeowner our team was working with had been stalling for weeks on a roof he'd already decided to buy. He was picking colors. He kept deferring. The assumption on our side was price, or the wife, or financing — the three things everyone assumes.

Our agent called back and dug for the actual objection. It wasn't any of those.

He'd have liked the rep to get on his roof.

That was it. Weeks of stall, and the reason was that the man selling him a roof had never been on it. He couldn't articulate it as a trust problem. He articulated it as a delay.

Here's the counterpart, from the other side of the same trade. This is our founder describing the single most effective thing he has found to say at a kitchen table:

Notice what that question does. It doesn't attack the competitor and it doesn't defend the price. It establishes that the other estimates were written by people who never looked at the part that determines the number. The homeowner reaches the conclusion themselves, which is the only way anyone ever actually reaches one.

06 — The trap

Conditional effort

Scaling inspection quality to perceived close probability is the most self-defeating habit available in this business.

The logic feels rational. The appointment looks weak, the homeowner seems non-committal, there are two other trucks on the street — so you do the lighter version, save the real effort for the one that looks closable, and get to the next stop. Every field rep in America has done a version of this, in every trade.

We see the industrial version of this across the companies we work with, and it doesn't look like laziness at the appointment. It looks like process design. Our founder, describing the pattern on an internal call this month:

Read the conversion rate again. 75%. That is what a book looks like when you only run the appointments that were already sold. It is a beautiful number and it is measuring almost nothing, because the denominator has been filtered down to people who had already decided. The company reporting it has not found a better sales process. It has found a way to stop having most of the conversations.

The habit is inverted, for three reasons that compound.

1. It removes the only evidence that would have justified your price.

The homeowner can't judge the assembly. They judge the inspection. A thin inspection produces no photos worth showing, no findings worth repeating, no diagnosis they couldn't have gotten from anyone with a ladder. You've handed them a decision with exactly one comparable variable left in it. They'll compare that variable. It'll be your number.

2. The judgment gets made before you've earned anything — so it comes true.

Perceived close probability gets assessed at the door, before the trust equation has had a single input. Withholding effort at that exact moment guarantees the read stays accurate. Every "I could tell they were just price shopping" gets confirmed by the behavior that followed the thought. The book of business becomes a record of the prediction, not the market.

3. Variable effort is the most visible form of self-orientation there is.

Homeowners are trained to watch for this exact signal. Consumer protection agencies actively warn about the contractor who "happens to be in the area," offers a quick free inspection, and applies pressure — that guidance exists because the pattern is a recognizable scam profile. A rep who inspects hard when the deal looks good and light when it doesn't produces, from the homeowner's side of the driveway, an indistinguishable signal. Both say the same thing: the level of care here is a function of what he's getting, not what I need.

The Gong consistency finding is the clean version of this. What separates high from low performers isn't how good they are on their best call. It's that high performers run the identical process whether the deal looks won or lost, and low performers don't. Consistency is the skill.

The fair version of the objection

None of this is worth reading if it doesn't survive contact with a bad Friday. So here is the counter-argument, in the actual words of a roofing owner who was arguing with research very much like this one:

"If somebody gets enough estimates, they're eventually going to find a roofer that's desperate and basically does the job for free. So how do you compete with that?"

And the sharper one, aimed straight at survey data:

"Of course somebody's going to say they want the best quality thing. Of course that's going to be everybody's answer. But is that what happens in reality? I don't know. Maybe yes, maybe no."

He is right about the desperate roofer. There is always one, and you will lose some jobs to him no matter what you do. The argument here isn't that trust beats a free roof. It's about which homeowners end up in front of you in the first place, and what they have to compare when they get there.

You cannot stop the market from producing a cheaper number. You can stop being the person whose appointment gave the homeowner nothing else to weigh it against.

07 — The economics

What a thin inspection actually costs

The lead is already paid for. The only variable left is whether it produces anything.

We went looking for hard cost-per-lead and close-rate benchmarks for roofing, decking and bathroom remodeling. Most of what circulates does not survive checking — see the list at the end of this report. What did survive:

SearchLight Digital, "Roofing Google Ads Cost Per Lead (2026 Benchmarks)," April 2026, based on 15 accounts. vendor — a marketing agency reporting its own client accounts. Small sample. No independent cost-per-lead benchmark exists for deck building or bathroom remodeling; every figure we found was published by a company selling leads or marketing services, with no disclosed methodology.

You don't actually need a precise benchmark for the argument to hold, because the logic doesn't depend on the size of the number:

• The acquisition cost is already spent. It was spent when the appointment got booked. A light inspection doesn't save it. It converts it to zero.
• The jobs you win on price are your worst jobs. Lowest margin, highest scrutiny, most likely to produce a dispute, least likely to refer anyone.
• A thin inspection produces no reputational asset. No review, no photos worth sending, no story the homeowner tells a neighbor over the fence. Nothing compounds.
• Referral flow is the only channel you cannot buy at any price — and it is generated almost entirely by inspections you weren't paid for at the time.

What it looks like when reputation is the whole company

In March 2026, Roofing Contractor reported that Amstill Roofing in Houston was hit with roughly 300 fraudulent one-star reviews. Its Google rating fell from 4.9 to 3.2 within hours. Owner Sam Stilley's explanation of why that was existential is the sentence that matters here:

"Ninety-five percent of my business is from leads calling us, warm leads."

Chris Gray, "Negative Review Scams: How Roofing Contractors are Fighting Back," Roofing Contractor, March 9, 2026. The 95% figure is Stilley's own characterization of his business as quoted, not an audited metric. The same article documents the same extortion pattern against named contractors in Ohio, Florida and Tennessee, and the NRCA published a separate warning in October 2025.

Reputation wasn't one of his marketing channels. It was the company. That asset gets built one thorough inspection at a time — including, especially, the ones that didn't close.

08 — The standard

What a trust-building inspection contains

One protocol, three trades. Run it on every job, regardless of how the appointment looks.

Every trade has one place a real inspector goes and a lazy one doesn't. It's the clearest available signal that you actually inspected rather than eyeballed, and it costs about eight minutes.

The rest is identical in all three:

• The same protocol every time. Not scaled to the homeowner, the neighborhood, or the read you got at the door.
• Go to the hidden place. Every time it's accessible.
• Photograph everything. Overview of each area, then labeled close-ups of every defect.
• Show the photos at the house, in real time, on the device. Not emailed later that night.
• Report findings the homeowner could accurately repeat to someone who wasn't there. If they can't retell it, you didn't explain it — you performed it.
• Name what's fine, and what doesn't need replacing. Voluntarily disclosing something you can't bill for is the fastest available reduction in perceived self-orientation.
• Walk them through what happens next — including what an adjuster looks for, if there's a claim.
• Present the proposal in person, at the appointment. There is no credible statistic on in-person versus emailed proposals, and we cut the one everybody quotes. What there is: a peer-reviewed study finding face-to-face requests roughly 34 times more effective than emailed ones. It measured survey requests to strangers, not proposals — so treat it as direction, not as a close-rate multiplier.

  Roghanizad, M.M. & Bohns, V.K., "Ask in person: You're less persuasive than you think over email," Journal of Experimental Social Psychology 69, 2017; 45 participants, 450 requests.

• Talk less than you think you should. If you've been talking ninety seconds without asking a question, stop and ask one.
• Leave with a signature or a specific return appointment. Never an open-ended "think it over."

Self-audit — answer honestly on your last ten appointments

Count the ones you can answer yes to for all ten. Not your best one. All ten.

• I ran the same inspection protocol on all ten, regardless of how the appointment looked when I pulled up.
• I went to the hidden place — attic, underside, behind the panel — on every one where it was accessible.
• Every homeowner left with labeled photos of their own house.
• I told at least one homeowner about something that did not need to be replaced.
• Each homeowner could have accurately explained my findings to their spouse that night.
• I listened more than I talked.
• I asked about their situation before I described my company.
• I presented the proposal in person, at the appointment.
• I spent the same effort on the ones I didn't think I'd close.
• I'd be comfortable if the homeowner watched a recording of my inspection next to my best competitor's.

09 — Conclusion

The one-sentence version

Homeowners can't grade your roof, your deck frame, or what's behind your tile. They grade your inspection. If the inspection is conditional, the only thing left to compare is your number — and the market will always produce someone with a lower one.

Roughly half of construction businesses are gone inside five years, and among the smallest remodelers it's seven in ten. Almost none of them lost the ability to build the thing they sell. They lost the ability to get work without buying it, and eventually the price of buying it exceeded what the work returned.

The way out isn't a better pitch, a better price, or a better close. It's the inspection — run identically every time, on every house, for every homeowner, including the ones you're certain will never sign. Those are the ones that build the only lead source that gets cheaper over time.