The Cost of Rework in Custom Home Building: What the Data Actually Says

On a $2 million custom build, roughly $180,000 disappears into rework. Tearing out and redoing work that was already "done." On a 15% margin, that single line item can erase more than half the profit on the job.

That number isn't a scare tactic and it isn't ours. It falls straight out of the most-cited research in the industry, applied to the build sizes custom home builders actually run. This post walks through the data: how much rework really costs, where it starts, why it gets more expensive the longer it hides, and what actually prevents it.

How much rework really costs

The figure that shows up again and again across independent studies is about 9% of total project cost once both direct rework (labor and material to redo the work) and indirect costs (schedule slip, rush orders, supervision hours, callbacks) are counted:

• The SpecAlign rework-cost model puts direct plus indirect rework at about 9% of total project cost. It is built from Dodge Construction Network's Not by Design (2024) as the primary anchor, corroborated by KPMG, McKinsey, and Deloitte, and cross-checked against peer-reviewed residential rework research.
• The direct rework figures builders typically track (2-5%) roughly double once indirect effects (schedule slip, rush premiums, supervision hours, callbacks) are counted. That's why the "official" rework number on most jobs dramatically understates the real one.
• The time dimension compounds it: field and office teams lose a large share of the week to non-productive work, looking for project information, resolving conflicts, and dealing with mistakes and rework. Procore's own research puts the data-hunting alone at roughly 18% of the workweek.

What 9% looks like at your build size

Build size	Planned profit at 15%	Lost to rework at 9%	What's left
$1.5M home	$225,000	~$135,000	$90,000
$2M home	$300,000	~$180,000	$120,000
$3.5M home	$525,000	~$315,000	$210,000
$5M home	$750,000	~$450,000	$300,000

The ratio never changes: at a 15% margin, rework at 9% takes 60 cents of every profit dollar. The checks just get bigger.

Custom residential margins generally run 10-20%. At any point in that range, unmanaged rework is not a rounding error. It's the difference between a profitable year and a break-even one.

Where rework actually starts

The build rarely fails inside one trade. It fails in the gaps between them, and most of all in the gap between the documents and the field.

Coordination gaps and missed updates are the largest single driver: the plan and the finish schedule that disagree, the revision that exists in the office but not in the framer's hands, the selection that changed after the order went out. Dodge's Not by Design (2024) traces 33% of quality problems to coordination and found only 11% of field crews always have the current set. The rest splits across the familiar suspects, but they cascade from the same root:

1. Rework and callbacks. Conflicting documents get built, torn out, and redone on a compressed schedule.
2. Wasted hours. Half a day lost to finding files, texting subs, and retyping specs into spreadsheets.
3. Wasted material and labor. Rush orders at 25-50% premiums; crews idle while they wait on the right spec.
4. Slipped schedule. A long-lead item nobody flagged stalls the whole sequence.
5. Unhappy clients. The daily "what happened?" call, and the referral that never comes.

Why it costs so much: the escalation ladder

The same conflict gets more expensive at every stage it goes unnoticed. This is the single most useful mental model in the rework literature:

Stage 1: Born on paper. The plan and the finish schedule disagree. The conflict sits in the documents, unread. Cost to fix: a markup. Minutes. No materials, no labor.

Stage 2: Changed, not communicated. The revision exists, but the sub is building from last month's set. Nobody knows until the walkthrough. Cost to fix: rework. The missed-update gap is the single largest source of rework, and it starts exactly here.

Stage 3: Built wrong. Caught at install, it's physical now. The fix is a tear-out on a compressed schedule, with rush-order materials and a client watching. Cost to fix: the numbers in the table above.

Every dollar in the 9% was once a Stage 1 problem that would have cost nothing to fix. The economics of rework are really the economics of detection timing.

What actually prevents it

Strip away the vendor noise and the prevention playbook is short:

1. One current set. A single source of truth for plans, specs, and selections. Not a shoebox of email attachments, texts, and Dropbox folders where three versions of the truth coexist. On a small crew that comes down to five checkable revision-control habits, not a document-control hire.
2. Deltas, not documents. Trades don't need the full set re-sent; they need to know what changed since they last looked. "Three changes since your last visit" prevents more tear-outs than a hundred-page reissue.
3. Read the documents before the field does. Most conflicts are visible on paper weeks before install, if anything actually cross-reads the plan against the finish schedule against the selections. Historically nothing did, because that reading was a human Saturday that never happened.
4. Check the money and the clock at decision time. A quote 17% above your own history, or a lead time that no longer fits the schedule, should surface when you can still act on it. Not in the post-mortem.

This is the gap SpecAlign was built for: AI that actually reads every plan, spec, and selection, keeps every trade on the current set (with the deltas called out), and flags the cost and schedule risks while they're still Stage 1. A markup, not a tear-out. The industry data above is the reason the product exists; catching one conflict on paper pays for a year of software many times over.

Want to see it on your own plans? Watch the 4-minute demo. Free during beta for founding builders.

---

Sources

• SpecAlign rework-cost model (the ~9% of total project cost figure), built from Dodge Construction Network's Not by Design (2024) as the primary anchor, corroborated by KPMG, McKinsey, and Deloitte, and cross-checked against peer-reviewed residential rework research
• Dodge Construction Network, Not by Design (2024) (only 11% of field crews always have the current set; coordination named among the leading root causes of quality problems)
• Procore (teams lose roughly 18% of the workweek hunting for project data; competitor-reported, corroborating the time cost)

Figures are industry averages; individual results vary. SpecAlign's own claims are limited to what the product demonstrably does: reading project documents, delivering current sets to the field, and flagging cost and schedule risks at decision time.