Plan revisions and change orders: catching what changed before it costs you

Where revision losses actually come from

The expensive revision almost never announces itself. It gets issued mid-bid or just after contract, as “Rev 2 — see clouds” in an email with six attachments. The cloud on sheet A-4 marks a moved pantry wall; what nobody clouded is that moving it changed the header over the relocated opening, added four feet to a plumbing wall, and bumped two windows one size on the schedule sheet the drafter edited on the way through.

The job gets built to the new sheets — the field always builds the latest set — but the budget and the bid still describe the old ones. Materials arrive per the revised plans, cost more than the bid carried, and by the time the numbers disagree loudly enough to notice, the work is in place and the negotiation is retroactive. Retroactive change orders are the worst kind: the leverage is gone, the memory is fuzzy, and the difference usually comes out of margin.

The loss, in other words, is not caused by the revision. It is caused by the gap between when the documents changed and when you knew what the change did to your numbers.

Why sheet-diffing by eye fails

Revision clouds mark what the drafter remembered to cloud. That is not cynicism about drafters; it is how document production works under deadline. A re-issued sheet is a fresh plot of a living model — everything that changed since the last issue is in it, whether or not it got a cloud, and the revision block says only “revised per owner comments.”

What slips through a visual comparison, reliably:

• Dimension changes without geometry changes — the wall looks identical at plot scale; the string above it went from 12'-4" to 13'-0". Your eye reads shapes, not numbers.
• Schedule edits — one cell in a window schedule changes a rough opening, a header, and a glass spec. Two dense tables side by side are where attention goes to die.
• Ripple effects on unclouded sheets — the clouded change on A-4 has consequences on S-2 and the elevations that were re-plotted without clouds because “nothing changed there,” except what did.

Eyeballing catches the moved wall. It misses the moved number — and the moved number is what you buy materials with.

Compare quantities, not pixels

The dependable comparison is downstream of the drawings: run a full takeoff on the bid-set revision, run the same takeoff on the new revision, and diff the results. Which quantities moved, on which sheets, and what the change does to the material budget — as numbers, not as a feeling that sheet A-4 looks different.

A quantity diff has a property no visual check has: it does not care whether the change was clouded. A schedule cell edit and a dramatic clouded wall move show up the same way — a line whose quantity is different between revisions, traced to the sheet that moved it. The honest boundary matters too: lines that could not be proved on either revision should be flagged as unproved rather than silently compared, because a diff of two guesses is noise wearing the costume of information.

Turning a diff into a change order

“The lumber line grew 6%” is an observation. A change order is that observation dressed in evidence and priced before the material is bought:

• What changed — the specific lines whose quantities moved between Rev 1 and Rev 2, each traced to the sheet and the element that moved it.
• What it costs — the quantity deltas priced at current supplier numbers, plus the labor the change implies, stated as a total the customer can react to.
• When it was caught — dated to the revision issue, before procurement. Timing is credibility: a change order raised the week the revision landed reads as diligence; the same dollars raised at invoice time read as an excuse.

Customers push back on vibes; they rarely push back on a moved wall they asked for, shown on their own drawings with the quantity difference next to it. The evidence does the negotiating.

A revision workflow for small builders

None of this needs a document-control department. It needs a habit with three rules:

• Log every issue — a one-line register: date received, sheets affected, what the transmittal claims changed. Two minutes, and it establishes the timeline every later conversation leans on.
• Re-measure before re-pricing — quantities first, then costs. Pricing an unmeasured revision is guessing with confidence.
• No revised sheet reaches the field unexamined — the set the crew builds from is the set your budget has already answered for.

PlanProof’s Revision Compare does the measuring half of the habit: the same proved takeoff runs on both revisions, and the diff lists which lines changed, by how much, and where on the drawings — with unproved lines flagged for review instead of silently compared. What stays yours is the judgment: which moved quantities become a change order, and how to price it.

← All articles