The addendum-day playbook:
revisions without the panic

Why addenda blow up estimates

Every estimator knows the feeling. The takeoff is 80% done, the numbers are starting to hold together, and then the email arrives: "Please find attached Addendum 2." Forty new sheets, a two-page narrative, and a bid date that didn't move.

The damage rarely comes from the changes themselves. It comes from the scramble: someone measures from the old set because the new one wasn't clearly labeled, someone else re-does an entire floor when only one wall type changed, and nobody writes down what was different — so a week later, when the GC asks why your number moved, the answer lives in somebody's memory.

The fix isn't heroics. It's a boring, repeatable process you run the same way every time an addendum lands. Here's the one we'd suggest.

Step 1: version discipline before anything else

Before a single measurement happens, the new set gets quarantined and named. A convention like ProjectName_Add2_2026-03-14 beats drawings_final_v3_NEW every time. Three rules keep a team honest:

• One current set. The moment the addendum arrives, the previous set is marked superseded. Nobody measures from it again — not "just to finish this sheet."
• Never overwrite. Keep every issued set forever. Addenda get argued about after award; the old sheets are your evidence.
• Announce it. A one-line message to everyone touching the bid: "Add 2 is in, old set is dead, changes being logged." Thirty seconds of communication prevents the classic two-estimators-two-sets disaster.

Step 2: diff every sheet — not just the ones the addendum mentions

Here's the part most teams skip, and it's the part that costs money. The addendum narrative says "revised sheets A2.1, A2.3, S1.0." Experienced estimators will tell you the narrative is a starting point, not a contract. Architects reissue whole sets, and quiet changes ride along on sheets nobody flagged — a bumped ceiling height, a wall type swapped in a schedule, a door that grew a rating.

So the rule is: overlay-compare every sheet in the new issue against its predecessor, whether or not the addendum mentions it. With a decent comparison tool this takes seconds per sheet. Without one, you're doing the two-monitors-and-squint method, and you will miss things.

Step 3: cloud it and log it

Every real change gets two things: a revision cloud on the sheet, and an entry in a change log. The cloud is for eyes — anyone opening the sheet sees instantly where to look. The log is for memory — what changed, which addendum introduced it, who found it, and whether it's been re-measured yet.

The log matters more than most teams think. A useful one has five columns:

That last row is the whole argument for step 2. Nobody told you about A2.4. The diff did.

Step 4: re-measure the deltas — only the deltas

The instinct under deadline pressure is to re-do everything "to be safe." Resist it. If your original takeoff was organized by sheet and by system, an addendum is surgery, not amputation: delete the measurements the change invalidated, re-measure the affected area on the new sheet, and leave the other 90% of the takeoff alone.

This is where takeoff hygiene from week one pays off. Measurements named and grouped by sheet can be filtered to "everything on A2.1" in one click. Measurements dumped into one undifferentiated pile mean you genuinely can't tell what's still valid — and then you really do have to start over.

Step 5: close the loop with the team

Before the revised numbers go into the bid, someone owns a ten-minute summary: which sheets changed, what the cost movement was by trade, and which changes carry risk worth a note in the proposal. On a multi-estimator bid, this is also where you confirm nobody's scope silently absorbed or dropped anything. Addenda are where scope gaps are born; a short structured handoff is where they die.

Where the tooling earns its keep

Everything above is process — you can run it with any software, or with paper and a light table if you hate yourself. But the diff-every-sheet step is where tools separate. In Groundwork Takeoff, plan compare is a live-canvas overlay, not a flattened export: removed work renders red, added work renders blue, with a blend-mode opacity slider to fade between issues. Sheets auto-align using text anchors, and when a reissued sheet shifted its title block, you drag-align it manually and the diff recalculates.

The revision toolset picks up from there: cloud the change, and the Revision Log panel records who changed what and when — so the change log from step 3 isn't a spreadsheet someone forgets to update, it's a panel that's already current. Because measurements live on the same canvas, re-measuring a delta means deleting the old run and drawing the new one; the estimate grid re-prices as you draw. And since collaboration is real-time and multi-user on the same sheet, the "close the loop" step happens on the drawings themselves instead of over screenshots.

If you're comparing tools for revision handling specifically, our comparison hub covers how the major takeoff platforms handle overlays, and the Revu 20 end-of-life guide is worth a read if your current overlay workflow lives in Bluebeam.

The playbook on one page

1. Quarantine and name the new set. Kill the old one. Tell everyone.
2. Overlay-diff every sheet against its predecessor — mentioned or not.
3. Cloud every change on-sheet; log it with source and cost impact.
4. Re-measure only the invalidated areas. Leave the rest alone.
5. Summarize movement by trade before numbers enter the bid.

Run it twice and it stops being a fire drill. Addendum day never becomes fun, but it can become routine — and routine is what protects a bid.