Why 80% of residential GCs lose money on every project (and how to stop it)

A 4% margin on a residential build is considered healthy. A 6% margin is considered a great year. So when the same crew that hit 7% last year shows up at 1.2% this year, the difference is almost never a single line item — it is dozens of $400, $1,800, and $6,000 hits that nobody saw coming until the close-out reconciliation. Most residential GCs lose money on most projects. The honest reason is that the job ran the GC, not the other way around.

The underlying problem is what we call information debt: the accumulating gap between what is actually true about the project and what the GC currently knows. Every undocumented change, every field decision made over the phone, every spec conflict that everyone notices but nobody flags — those are interest payments on information debt. They compound silently for weeks. When they hit the balance sheet, they hit hard.

The five categories of profit leak

After looking at thousands of residential projects, the leaks fall into five repeating categories. None of them are exotic. All of them are fixable.

1. Rework from spec conflicts caught after framing

Architectural shows a 9-foot ceiling. Structural shows a 22-inch deep beam in the same span. Mechanical shows ductwork that needs 14 inches of vertical clearance. Somebody notices in the field, six weeks after framing started. The fix is a 2x4 soffit drop you never quoted, or a value-engineered beam that triggers a structural revision, or both.

The dollar damage on a typical 3,200 sq ft single-family is $8,000 to $50,000 per missed conflict, and a clean residential set usually has three to seven of them. Catching these at the plan-review stage — before the lumber package is ordered — is the single highest-ROI move a residential GC can make.

2. Verbal change orders that never got papered

The homeowner asks for a slightly bigger pantry on a Saturday walk- through. You say sure, we will work it out. Your framer reframes it on Monday. The reconciliation conversation in October is now a dispute, because everyone's memory of the conversation is different and there is no signed change order. You eat the upgrade or you fight for it; either way you lose three hours of accounting time and probably some goodwill.

The cost of capturing a change order in writing is a 45-second decision-log entry. The cost of not capturing it is sometimes the margin on the entire project.

3. Sub no-shows and re-sequencing

The drywaller pushes Tuesday to Friday because a commercial job ran long. Your taper was on the calendar for Monday after that, and now loses two days that he charges you for anyway. Your painter was stacked behind the taper. The dominoes fall for two and a half weeks. None of that is on the schedule you presented at contract signing — it is being absorbed by the GC.

4. RFIs that stall for days because nobody knows they exist

A framing question goes from foreman to PM by text on Thursday night. The PM has 200 unread texts. By the time the architect sees it on the following Wednesday, the crew has built around the problem incorrectly, and the answer triggers a partial demo. The cost is two crew-days plus material. The fix is not faster architects — the fix is a tracked RFI system with a deadline, a priority, and an audit trail. (We wrote a whole post on the five RFI mistakes that compound here.)

5. Punch list creep at close-out

You finish the project, but the homeowner walks the house with a notepad and finds 67 items. Some are legitimate. Some are buyer's remorse. Without photos and dated documentation of the as-installed condition, every item is a negotiation, and most of those negotiations the GC loses or splits 50/50.

Why this keeps happening

The pattern in all five is the same: the project is producing information faster than the GC can capture it. A residential build generates 200-400 decisions, 50-150 RFIs, 20-60 change events, and 1,000+ field communications. A GC with two phones, an email account, and a spreadsheet is operating at roughly 10% capture. The other 90% is information debt accruing interest.

The interest rate on information debt is not constant. It compounds fastest at the moments where the project changes phase — design to framing, framing to MEP rough-in, MEP to finish. Those handoffs are when undocumented decisions surface as conflicts. Catching them in the prior phase is 10x cheaper than catching them in the next one.

The five-step fix sequence

You do not need a transformation program to plug these leaks. You need five process changes, in this order:

1. AI cross-sheet plan review before lumber order. Run the architectural, structural, and MEP sets through a cross-sheet conflict pass. Even catching one $20,000 issue per year pays for the tooling 20x over. Days saved: 5-15 per project.
2. One-tap change-order capture.Every field change gets a 60-second decision-log entry with the homeowner's initials photographed. No exceptions, no "we will write it up later." Dollars saved: $5,000-$40,000 per project on disputed scope.
3. Sub performance scoring. Track on-time start, punch count, and responsiveness across 3-5 jobs for every sub. Drop the bottom 20% of your bench. The variance between a good sub and a bad one in the same trade is enormous. Days saved: 8-20 per project.
4. RFI tracking with deadlines.Every RFI gets a subject line written like a question, a deadline, a priority, and an assignee. No texts, no "I will call him." Days saved: 3-10 per project.
5. Dated photo log at substantial completion. Photograph every elevation, every room, every transition the day you call substantial completion. Punch-list disputes evaporate when there is timestamped evidence of the as-installed condition. Dollars saved: $2,000-$15,000 per project.

What success looks like

A residential GC running the five-step sequence will typically move from a 1-3% project margin to a 6-9% project margin within two full project cycles. That is not because the work got cheaper — the materials and labor are the same. It is because information debt stopped being absorbed by the GC and started being captured, priced, and either paid back or refinanced as change orders.

The other thing that happens, less obvious but more important long-term: your relationship with homeowners stops being adversarial at close-out. You walk the house with documentation. They walk it with their notepad. Most of the items are already accounted for. The conversation is short. They give you a referral. The next project costs you 30% less to source.

How Buildra fits

Buildra is built specifically around closing the information-debt gap for residential GCs. The plan reader runs the cross-sheet review. The voice-to-decision tool captures field changes in 8 seconds. The RFI module enforces subjects, deadlines, and assignees. The sub scoring runs in the background as you close out each job. You do not have to remember to do any of it — the friction is what kept it from being done at all.