Why doesn't a revision reach the sub who needs it?

Because most small builders route changes through whatever channel is fastest at the moment: a text, a verbal note at a walkthrough, an email attachment buried in a thread. None of those create a record the next person checks before they cut material. The change is real and documented somewhere. It just isn't in the one place the trade looks.

How long does a change usually sit undetected on a custom home?

There's no universal number, but the pattern from CII's RT-382 work is consistent: the gap between a documented change and someone acting on the old version is usually weeks, not days, because nobody is re-checking finalized sections until the trade shows up to build them. That's the whole problem: detection happens at install, not at revision.

Does an RFI or VPO process fix this by itself?

Only partly. An RFI or VPO creates a paper trail for the change itself, but it doesn't guarantee the framer, the tile setter, or the electrician actually reads it before they start. You need the record and a way to surface it to the right trade at the right moment. Most builders have the first and are missing the second.

What's the cheapest fix for a small builder without a document control team?

Put every trade on one current set instead of individually emailed PDFs, and tell them what changed since their last visit rather than resending the whole package. A three-line 'changed since your last visit' note catches more conflicts than a full reissue, because subs actually read three lines.

July 21, 2026 · 5 min read · SpecAlign

Half of All Rework Starts With a Change That Never Reached the Field

A framer nails off a wall three feet from where the revised plan puts it. Not because he's careless. Because he's building from the set on his phone, and that set is six weeks old.

Ask ten GCs how their worst rework happened this year and eight of them will describe some version of that story. Not a bad sub. Not a design flaw. A document that changed, and a person who never saw it change. Industry research keeps circling the same figure: something close to half of all construction rework starts exactly there.

Where the chain actually breaks

Picture how a single change is supposed to travel. The architect redlines a detail. The GC approves it. The updated sheet goes into the plan set, the finish schedule gets touched if selections shifted, and every trade working near that scope gets the new version before they order or cut. Four steps, one direction, no gaps.

Now picture how it actually travels on most jobs. The architect redlines it. The GC gets a PDF by email, forwards it to two subs and forgets a third. Someone mentions it at a walkthrough to whoever happened to be on site that day. The framer, who wasn't there, keeps working from the version on his phone. Nobody updated that.

By the time anyone notices, the wrong version is already framed, ordered, or poured. The change was never lost. It just never reached the one person who needed it before he picked up a saw.

The missed-update half, defined

FMI and Autodesk's Construction Disconnected research puts the number at roughly half: about 50% of rework, measured across the industry, comes from poor project data and miscommunication rather than field conditions, weather, or bad craftsmanship. Call that slice the missed-update half. It's the single biggest bucket in the whole rework pie, bigger than material defects, bigger than weather delays, bigger than subcontractor error.

That statistic connects directly to the escalation ladder from detection-timing economics: every dollar of rework was once a Stage 1 problem, a conflict sitting quietly on paper, before it became a Stage 2 miscommunication and then a Stage 3 tear-out. The missed-update half is what happens between Stage 1 and Stage 2. The conflict existed. It just never crossed the gap to the person holding a hammer.

This isn't a people problem dressed up as a process problem. FMI's researchers were explicit that the failure sits in how project information moves, not in who's building it. Framers, tile setters, and electricians all build from whatever document is in front of them. The missed-update half exists because nobody guaranteed that document was current.

Why the gap stays invisible

A 2004 NIST study (GCR 04-867) put the annual cost of poor data interoperability across the U.S. capital facilities industry at $15.8 billion, and its central finding still holds twenty years later: the expense isn't the software gap between systems. It's the information gap between people who each think they're looking at the same thing.

That's what makes the missed-update half so dangerous. Nobody is negligent in the moment. The GC believes the update went out. The sub believes his set is current. Both are acting in good faith on bad information, and good faith doesn't catch a Stage 1 conflict. Only a second set of eyes on the actual document does, and on a lot of custom jobs, nothing plays that role until the walkthrough.

PlanGrid and FMI's field research adds the time cost on top of the dollar cost: construction professionals lose more than 14 hours a week hunting for the right file version, texting subs to confirm which set is current, and untangling conflicts that a synced document would have flagged automatically. That's not lost to bad weather. It's lost to a document control gap that most small builders have simply never had the headcount to close.

What actually closes the gap

Here's how the same change moves with a fix in place, compared to how it moves without one.

Step	Without a fix	With a current-set system
Change happens	Redline in the office, no distribution plan	Redline logged against the live set
Reaches the trades	Emailed to whoever's on the thread	Pushed to every trade tied to that scope
Sub's copy	Whatever PDF they saved last	Always the current version, flagged
What they see	The whole reissued set (rarely opened)	"3 changes since your last visit"
Detected at	The walkthrough, after it's built	The redline, before anyone cuts material

Three habits close most of that gap without adding a meeting to your week. Put everyone on one current set instead of individually emailed attachments; a shoebox of PDFs guarantees three versions of the truth exist at once. Send deltas, not documents. A sub ignores a 40-page reissue and reads three lines. And cross-check the plan against the finish schedule and the selections before the order goes out, because that's still the cheapest place in the whole chain to catch a conflict.

None of that requires a document controller on payroll. It requires the update to actually arrive.

SpecAlign was built around exactly this gap: every trade sees the current set with changes called out since their last visit, so the missed-update half has nowhere to hide. A markup costs minutes. A tear-out costs the number in last week's post.

Fix the distance between the revision and the person building it, and the craftsmanship problem you thought you had mostly disappears.

Sources

FMI / Autodesk, Construction Disconnected (rework attributable to poor project data and miscommunication, ~50%)
Construction Industry Institute, RT-382 (detection timing and measurement of rework)
NIST GCR 04-867, Cost Analysis of Inadequate Interoperability in the U.S. Capital Facilities Industry ($15.8B annual figure)
PlanGrid / FMI field research (14+ hours/week lost to non-productive work)

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.

Frequently asked questions

What is the missed-update half?: It's the roughly 50% slice of all construction rework, industry-wide, that FMI and Autodesk's Construction Disconnected research attributes to poor project data and miscommunication rather than field conditions or craftsmanship. In practice it's almost always the same event: a document changed, and the person building the work never got the new version.
Why doesn't a revision reach the sub who needs it?: Because most small builders route changes through whatever channel is fastest at the moment: a text, a verbal note at a walkthrough, an email attachment buried in a thread. None of those create a record the next person checks before they cut material. The change is real and documented somewhere. It just isn't in the one place the trade looks.
How long does a change usually sit undetected on a custom home?: There's no universal number, but the pattern from CII's RT-382 work is consistent: the gap between a documented change and someone acting on the old version is usually weeks, not days, because nobody is re-checking finalized sections until the trade shows up to build them. That's the whole problem: detection happens at install, not at revision.
Does an RFI or VPO process fix this by itself?: Only partly. An RFI or VPO creates a paper trail for the change itself, but it doesn't guarantee the framer, the tile setter, or the electrician actually reads it before they start. You need the record and a way to surface it to the right trade at the right moment. Most builders have the first and are missing the second.
What's the cheapest fix for a small builder without a document control team?: Put every trade on one current set instead of individually emailed PDFs, and tell them what changed since their last visit rather than resending the whole package. A three-line 'changed since your last visit' note catches more conflicts than a full reissue, because subs actually read three lines.