Not Knowing Your Numbers

01 — Where part one left off

Half of construction companies don't make it five years

Part One of this series — Why Contractors Go Out of Business — was about the front of the business. This one is about the number underneath it.

The federal government tracks how long new businesses actually last. The Bureau of Labor Statistics follows every group of new establishments from the year they open and records how many are still running each year after.

So roughly half make it, and half don't. That's not a first-year cliff. It's a slow erosion, and most of it happens quietly.

Part One covered the front end — why the inspection is the actual product, and why a contractor who only works hard on the deals he thinks he'll win ends up with nothing but price shoppers.

This one covers the other half. Not how the job is sold. What the job is priced at, and whether anybody in the company actually knows.

02 — How it actually happens

Most construction companies don't die slowly. One job kills them.

The insurance companies who pay out when a contractor fails have studied this. Their answer is specific.

Travelers runs one of the largest construction surety operations in the country. When a contractor defaults, they're the ones covering it — which means they've built a detailed record of what actually went wrong.

The rest of the industry research lines up with it. The Surety Information Office, working from the accountants and underwriters of failed construction companies, publishes the same short list of causes over and over:

Read those four again. Every one of them is the same failure wearing a different hat: the company does not have a reliable number, and has no system that would tell them.

The job that ends a construction company is almost never the job anybody was worried about. It's the one that looked fine on paper.

03 — The math

The mistake that lets a company be busy all year and still lose money

This is arithmetic, not opinion. Check it on a calculator. Most contractors in the industry get it wrong.

There are two different words for the money above your costs, and they are not the same number.

Because the price is always bigger than the cost, margin is always a smaller number than markup for the same job. Multiple industry sources — Procore, ServiceTitan, and construction accounting firms — independently name confusing these two as the single most common pricing error contractors make.

Here's what it costs.

That contractor didn't get beaten on price. He didn't lose a bid. He won the job and handed back $12,857 he was never going to see again — and nothing in his week told him it happened.

Run $1,000,000 of job cost through that same error in a year and it's roughly $128,000 of profit that never existed.

The conversion nobody memorizes

If you want a specific margin, this is the markup you actually have to apply:

Formula: markup = margin ÷ (1 − margin).

And then there's overhead

Markup has to cover two things, not one: overhead and profit. Overhead is everything the business pays for that isn't tied to a specific job — the truck payments, the phone, the insurance, the office, the software, the owner's own time when he isn't on a roof.

If overhead isn't recovered inside the price, it comes out of profit. Which is the technical explanation for a very common feeling:

Booked solid, every week. Crews running every day. Nothing in the bank at the end of the year.

04 — Why nobody catches it

The estimate is a guess that never gets graded

The reason this error survives for years isn't ignorance. It's a missing feedback loop.

In most residential construction companies, the sequence goes like this:

1. 1Somebody prices the jobUsually from experience, a rough measurement, and what the last similar job went for.
2. 2The job gets builtMaterial gets ordered — often more than once. Labor takes however long it takes.
3. 3The invoice gets paidMoney comes in. It feels like a win, because it is a payment.
4. 4Nobody ever compares step 3 to step 1This is the whole problem. The estimate is never scored against what the job actually cost. So the estimate never improves, and the error repeats on every job that follows.

Travelers names this directly in their failure research: contractors struggle with accurate and timely reporting of job cost information, particularly between the field and the office. The information exists. It just never makes the trip back.

Two things make it worse in residential work specifically.

The money arrives out of order

Payment timing hides the damage. Levelset's survey of 519 U.S. construction companies found that less than one in ten always get paid on time. Residential contractors do better than most — they report getting paid within 30 days about 48% of the time — but that still means half the time the cash on hand has nothing to do with whether the job made money. A healthy bank balance this month can be last month's deposits covering this month's losses.

Nobody complains about a bad estimate

If a crew installs something wrong, the phone rings. If a job is priced $12,000 too low, the homeowner is delighted, the crew gets paid, the job closes, and the reviews are good. The single most expensive mistake in the company produces the most positive feedback of any event in the week.

Every other mistake in a construction company generates a complaint. This one generates a thank-you note.

05 — The other side of the same number

The homeowner is auditing the number you aren't

Homeowner research points at one thing above every other complaint: a price that didn't hold.

Houzz runs the largest annual renovation study in the U.S. Their 2026 edition surveyed 20,358 people, including 10,176 homeowners who renovated in 2025.

A separate 2026 survey of 1,000 property owners by Clever Offers found the same pattern harder:

And from the trust research in Part One of this series: a 2025 national survey by Leaf Home and Morning Consult found nearly 70% of homeowners worry about contractors being unreliable, and 41% said they had already been deceived by a service provider on a home project.

Put those together and the homeowner's experience of this industry is simple. The number they were told is not the number they paid. It has happened to them, or to somebody they know, and they are now sitting in their kitchen deciding whether it is about to happen again.

They are not evaluating whether your price is high. They are evaluating whether your price is real.

06 — Why a bottom-line price reads as a risk

One big number tells a homeowner nothing — so they judge the only thing they can see

A quote with no scope doesn't look confident. It looks like something is being withheld.

Homeowners collect bids. A Modernize survey found close to half of homeowners — 44% — compare three or more contractor estimates before a home improvement project. That's not distrust of you specifically. It's just what people do with a five-figure decision.

Now think about what happens when three contractors each hand over a single number and nothing else. Take an illustrative example — no real job, just the shape of the problem:

This is the part most owners miss. A bottom-line price doesn't just fail to build trust. It actively manufactures the price shopper you're complaining about.

You removed every basis for comparison except cost, and then you're surprised the homeowner compared on cost.

What detail actually signals

Homeowners today have material prices on their phone. When a number arrives with no explanation attached, the natural question isn't “is this fair?” It's “what's in here that they don't want me to see?”

A written scope reverses that, and it does it through three specific mechanisms:

Part One of this series ended on the Trust Equation — the idea that trust divides by self-orientation, how much you appear to be focused on yourself instead of the customer.

An undocumented price is self-orientation in written form. You know exactly what's in that number. They don't. And you decided that was fine.

07 — One number, two failures

It's the same missing information both times

The margin problem and the trust problem are not two problems. They're one problem seen from two sides.

Every row below is a single number a contractor either knows or doesn't. Look at what one missing input does in both directions at once.

08 — The same failure, industry-wide

This isn't a construction problem. Construction just has nowhere to hide it.

The same missing-system failure that eats a margin also eats a pipeline. The sales research on this is enormous, and it says the same thing Section 07 just said.

Businesses lose customers constantly because there's no real process to track, nurture, and follow up on leads. Not a marketing problem. An operations problem. The leads exist. The system to work them doesn't.

SPOTIO's 2026 State of Field Sales survey, run across sales professionals in 14 industries, finds CRM adoption at 62% among field sales teams overall — and a turnover picture to match: 41% of teams report annual turnover of 50% or more, while just 16% stay under 10%. The same survey found 65% of field reps spend five or more hours a week on manual data entry, with only 3% reporting that entry is fully automated.

None of this is about motivation. It's the same finding as Section 07, one layer up: a business with no system doesn't find out what it's losing until the person holding it in their head is gone, tired, or busy. A pricing error and an unfollowed lead are the same failure wearing two different hats — silent, until somebody builds something that can't forget.

A field sales rep loses five hours a week to data entry the system should have already captured. A contractor loses the whole estimate to a number nobody validated. Same hole. Different department.

09 — The fix

What price engineering actually is

Not software. Not a template. A rebuild of how the company arrives at a number.

Blue Collar Infrastructure price engineers every company we work with before we automate anything. The order matters — automating a wrong number just produces wrong numbers faster. Here is how we approach it, in the order we run it:

1. 1Extract every real cost inputMaterial unit costs. Coverage rates. Waste factors. Labor converted from “what we pay a guy” into cost per unit of production. Crew rates. Disposal. Permits. Overhead allocation. Tax treatment. Every one of these comes from the contractor's own operation — never from an industry average, never from a benchmark.
2. 2Build the Inputs sheet — one home for every numberEvery cost the business has lives in exactly one named place. Nothing is typed into an estimate twice, and nothing is buried inside a formula. When a supplier raises shingle prices, one cell changes and the entire company reprices.
3. 3Build the Architecture — the calculation layerFor every line item on an estimate, we write the formula that produces its quantity, referencing the named inputs from step 2. Squares from measured area and pitch. Underlayment from square count and coverage rate. Fasteners from square count. Labor from production rate. Every quantity on the estimate is derived, not entered. Nobody is typing numbers into a proposal.
4. 4Validate against real jobs before anything goes liveWe run the contractor's actual completed jobs back through the engine and compare what it produces to what those jobs really cost. Gaps get traced to a specific input and corrected. The system does not go live until it reproduces reality. This is the step almost every estimating tool skips.
5. 5Automate the deliveryThe validated engine gets wired into the CRM. The on-site rep submits the inspection form. That submission drives the calculation, builds the line-item estimate inside the CRM, references it by URL inside the proposal, and emails the deliverable to the rep and the homeowner at the same moment.

Steps 1 through 4 are the work. Step 5 is what most companies think they're buying when they buy estimating software — and it's the only step that's worthless without the four before it.

Automating a number nobody validated doesn't fix an estimating problem. It industrializes it.

10 — What it changes

Three things become possible that weren't before

These follow directly from the architecture. Each one is a consequence of quantities being derived instead of typed.

1. Instant estimates that are actually reliable

The homeowner gets a complete, itemized price while the rep is still standing there — or in their inbox before he's out of the driveway. Not a range. Not a “let me work up some numbers and get back to you.”

Two things happen at once. The speed wins the race — and in home services that race decides most of it. Research on lead response finds that contacting a homeowner within five minutes instead of thirty makes you roughly 21 times more likely to qualify them, and the same body of research consistently finds a large majority of buyers purchase from whoever responds first.

And the number is right, because it came out of a validated engine instead of a rep's memory. Fast and wrong is worse than slow. Fast and right is the whole game.

2. Material and work orders come out of the same math

This is the part owners underestimate.

Because every quantity on the estimate is calculated rather than typed, the document that prices the job already contains the exact quantity of every material the job needs. The takeoff and the price are the same object.

That means the material order, the crew's work order, and the customer's estimate all come from one calculation instead of three separate people re-deriving it and introducing three separate errors. One takeoff. No re-keying. No “how many bundles did we say?” No second trip to the supply house.

3. Lead conversion changes on both ends

Speed converts the top of the funnel. Detail converts the bottom.

11 — The proof

Two clients, one conversation. Reported honestly — including where the window is narrower than usual.

These figures come from the ledger, not a call, and the window below is not the site's usual 12-month-before / 90-day-after standard. That's stated plainly, not dressed up.

On July 22, 2026, Blue Collar Infrastructure held a coaching conversation with the sales teams at two clients — Walnut Hollow Decks & Outdoor Living and The Custom Shower Company — about proposal speed: getting a priced document to the homeowner in hours, not days. That date is not a confirmed operational go-live for a new system at either company. It's the date of the conversation. What follows is what the ledger shows in the days on either side of it — the remaining days of July, split at the 22nd.

Walnut Hollow Decks & Outdoor Living: a clean before-and-after

Inside July, the split is sharper than the headline number. Counting new-job sales only — excluding one change-order line against a job originally sold in April — the company closed $69,730 in new work in the 21 days from July 1–21, three jobs. In the 10 days from July 22–31, it closed $83,674, four jobs. 2.52 times the daily dollar velocity. 2.80 times the deal cadence. Same company, same month, on either side of one conversation. Walnut Hollow's pricing sheet had already been validated and locked on July 13 — nine days before the conversation — so nothing else was changing in the background here.

The Custom Shower Company: real, smaller, and here's why

The second client shows no such clean monthly cliff. Its ledger over the same stretch: March $41,623, April $143,520 — the highest month in the whole run, well before the conversation — May $57,108, June $73,376, July $90,224. Forcing that into the same shape as Walnut Hollow's story would mean ignoring what the ledger actually shows.

The real effect is there, just smaller and located inside the month rather than across it. In the 21 days before the conversation, the company closed $45,941 in new work. In the 10 days after, $44,283 — nearly the same dollar total in less than half the time, or 2.02 times the daily dollar velocity.

The size of an effect is set by how broken the thing was before it, and by how many other things were changing at the same time. Walnut Hollow had one clean variable and a documented gap to close. Custom Shower had a real gain running alongside a second, unfinished change — which is exactly why its number is smaller and shouldn't be credited to proposal speed alone. Both are real. Only one of them is a clean read.

The full twelve months, all three clients, is in Year One: The Entire Book.

12 — The audit

Twelve questions about your own company

Answer from records, not from memory. If the answer is “I'd have to look,” that is the answer.

Check every box you can back up with an actual document today — not one you could put together if you spent a weekend on it.

• I can state my true labor cost per unit of production — not per hour, per unit of work completed.
• I know my monthly overhead to the dollar, and I know what share of it every job has to carry.
• I know whether the percentage I apply to jobs is a markup or a margin, and I know what the other one works out to.
• My material quantities come from a calculation, not from a person's estimate of how much we'll need.
• Every cost input in my pricing lives in exactly one place, and updating a supplier price is a one-cell change.
• I have compared what my last five completed jobs were estimated at against what they actually cost.
• I know which of my job types is the most profitable, by margin — not by revenue.
• Two different reps pricing the same house would produce the same number.
• My estimates state what is included, in quantities, in writing.
• My estimates state what is not included, and what happens if hidden damage is found.
• A homeowner can get a complete priced proposal from us the same day we inspect.
• If my best estimator quit tomorrow, the company could still price a job correctly on Monday.

The last box is the one most owners skip past. It's also the one that decides whether what you've built is a company or a job.