How to Answer a Returned Submittal and Resubmit
Why this matters
The day a submittal comes back is the day the job's schedule is decided. Read the stamp right and the order goes out that afternoon. Read it wrong and one of two things happens: the shop orders material the reviewer has not accepted and eats a return, or the shop sits on a package it was free to release and loses a week of lead time nobody gets back. The meaning of each box on the stamp, and the disclaimer printed under it, is covered in What a Submittal Review Stamp Means and What It Does Not. This card is what to do next, box by box, and how to send a resubmittal that closes the comments instead of starting a second argument.
First, the same three moves whatever the stamp says
Read every page, not the stamp. Reviewers write comments on the cut sheets, in the margins of shop drawings, and on a separate comment sheet. A stamp reading Make Corrections Noted with a correction buried on page 7 is a correction you are building whether you saw it or not.
Log the return the same day. Date back, action taken, number of comments, and the date your next move is due. The submittal log is where a returned package either moves or quietly stalls (see related: The Submittal Log SOP).
Send the field a copy only of the final, stamped version. A returned package with red marks on it is not a build document until you know which marks are instructions. The copy that belongs in the truck is the one that releases the work (see related: The Drawing Revision Control on the Truck SOP).
No Exceptions Taken: release it
Place the order from the stamped package itself, model string and options exactly as marked, and note the stamp date and submittal number on the purchase order. That one line is what lets anyone later trace the product on site to the document that released it. Then check the supplier's acknowledgment against the stamped page before the order is locked, because order-entry errors happen after review, not during it.
What breaks if you skip the acknowledgment check: the stamped submittal is correct, the product delivered is not, and the mismatch is found at installation, by which point the reviewer's stamp protects nothing.
Make Corrections Noted: build the corrections, and say so in writing
This box releases the work on the corrected version. Every correction the reviewer marked is now part of what you are furnishing.
- List each correction from every page, numbered, in a short acknowledgment.
- Change the order to match, line by line. A correction that changes a part number has to reach the purchase order, not just the file.
- Send the acknowledgment back through the same channel, stating that the corrections will be incorporated. Some Division 01 submittal sections (commonly 01 33 00, Submittal Procedures) or stamps require a corrected copy for record; where they do, the corrected copy is a resubmittal for record, numbered as one.
- Push any correction you cannot build back immediately. If a reviewer's note asks for an option the manufacturer does not make, or would change the price, say so in writing before ordering. Building around it silently is how a correction becomes a dispute at closeout.
What breaks if you skip it: the supplier ships the uncorrected configuration because it was already in their system, and you have installed a product the reviewer conditioned on a change you never made.
Revise and Resubmit: do not order
This is the box that costs schedule, and the reason not to order anyway is simple: the reviewer has said they need to see it again, so nothing in the package is accepted yet, including the items they did not comment on. A second review can raise new comments on pages the first review passed, especially once a revised item changes something else.
- Write a response to every comment, numbered to match the reviewer's numbers, saying what you changed and where. Agree, change it, and show the page. Or disagree, with the spec paragraph or the data that supports you. "Noted" is not a response.
- Revise the pages, and cloud every change on the resubmitted sheets so the reviewer reads only what moved. Unclouded changes force a full re-review and invite new comments.
- Number the resubmittal as a revision of the original, not as a new submittal: if the first was 26 51 00-002, the resubmittal is 26 51 00-002.1, or -002 Rev 1, whichever format Division 01 prescribes. A new number breaks the history, and a reviewer looking at a fresh number has no reason to limit the review to the comments.
- Resubmit the complete package, unless the stamp or Division 01 says only the revised items are wanted. A partial resubmittal leaves the reviewer unsure which version of the untouched pages is current.
Rejected: change the product or the package
A rejection says the product does not comply or is not what was asked for. You have two answers. Submit a compliant product, as a new revision of the same number, with the response sheet explaining the change. Or, if the reviewer misread compliance, answer with the spec paragraph and the data, once, in writing, through the general contractor. If the product you want is not on the spec's manufacturer list, the right paper is a substitution request, not a resubmittal (see related: How to Write a Substitution Request the Engineer Can Approve). Arguing a rejection twice in resubmittals costs two review cycles to reach the same refusal.
Worked example: the lighting package that came back Revise and Resubmit
An electrical shop submits 26 51 00-002, interior luminaires for a retail tenant, in week 3. Lead time on the fixtures is 7 weeks from release; the general contractor needs them on site in week 17; Division 01 allows 15 calendar days per review. Values are illustrative.
The package returns in week 6, Revise and Resubmit, with four comments:
RESPONSE TO REVIEW COMMENTS
Submittal 26 51 00-002.1 (resubmittal of -002)
Original returned: week 6, Revise and Resubmit
1. Reviewer: Type A color temperature shown 3500K;
schedule E-601 calls 4000K.
Response: Agreed. Revised to 4000K. Sheet 2,
clouded.
2. Reviewer: Type B dimming driver not marked.
Response: 0-10V driver marked per 26 51 00
2.3.B. Sheet 4, clouded.
3. Reviewer: Type C emergency battery pack required
per E-601 note 6, not shown.
Response: Agreed. Battery pack added, 90-minute
rating. Sheet 6, clouded.
4. Reviewer: Provide photometric data for Type D.
Response: Photometric file added as sheet 9.
Sheets unchanged: 1, 3, 5, 7, 8
The temptation. The project manager wants to release Types B, C and D in week 6, since comment 1 only touched Type A. The lead refuses, for the reason above: nothing in an R&R package is accepted, and comment 3 shows the reviewer was reading the emergency requirements across types. Ordering Type C in week 6 would have ordered it without the battery pack.
The resubmittal. The shop sends -002.1 in week 7, complete, with four of the nine sheets clouded and the response sheet on top. It returns in week 9, Make Corrections Noted, one correction: Type D finish to match the ceiling grid, white. The shop acknowledges it in writing the same day and releases the order in week 9 with the finish corrected.
The calendar. Released in week 9 plus 7 weeks of lead time puts the fixtures on site in week 16, one week ahead of the week 17 need. Had the first package come back Make Corrections Noted in week 6, release in week 6 would have landed them in week 13. The single R&R cycle cost 3 weeks (week 6 to week 9) of the 4 weeks of float the shop started with (week 13 against week 17). A second R&R cycle would have cost roughly another 3 weeks, putting them on site in week 19, two weeks late.
What changes the answer. If the reviewer had applied an action to each fixture type instead of one stamp to the package, with Type A marked Revise and Resubmit and Types B, C and D marked No Exceptions Taken, releasing B, C and D in week 6 would have been proper, unless Division 01 says a section is released only as a whole. Some reviewers do stamp per item. Read the stamp at the level it was applied: one stamp on the cover covers the package; an action beside each item covers that item.
When a comment is really a change
Some comments are not corrections at all. A reviewer who asks for a battery pack the drawings never called for, a finish upgrade, or a larger unit than the schedule shows is adding scope through the submittal, and accepting it on the response sheet is, in practice, agreeing to furnish it at no change. Answer that comment differently: say the requested item is not in the contract documents, cite the sheet and paragraph that show what was, and say the shop will furnish it on a change instrument (see related: The Paper That Changes a Job: RFI to Change Order). Then keep the rest of the response moving, so the dispute over one item does not hold the other items' review hostage. In the worked example, comment 3 was a correction because E-601 note 6 already required the battery pack. Had note 6 not existed, the same words would have been a change.
How to know the resubmittal will close
- Every comment number has a response, and the count of responses matches the count of comments.
- Every change is clouded, and the response names the sheet.
- The submittal number is the original number with a revision suffix.
- Nothing has been ordered from a Revise and Resubmit or Rejected package.
- The log shows the new due-back date and what it does to the slowest item's on-site week.
References
- See related: What a Submittal Review Stamp Means and What It Does Not
- See related: How to Prepare a Submittal for Your Own Scope
- See related: How to Read a Submittal Package
- See related: The Submittal Log SOP
- CSI MasterFormat, Section 01 33 00 Submittal Procedures, as written in your project manual