Glass Unit Failure Warranty Claim Standard

Purpose

A fogged insulating glass unit is a routine warranty event with a slow, manufacturer-controlled production clock behind it, and most of what goes wrong on these files has nothing to do with the glass. It is a claim submitted without the one photo that gets it bounced, a customer never told a production window who concludes the shop is stalling, a replacement installed without checking it matches the order, or a file closed without asking whether the same failure is starting on a neighboring unit. Each costs a second truck roll or a customer calling the manufacturer directly to ask why their contractor never followed up.

This procedure guarantees three things: every claim leaves the shop complete on the first submission, the customer has a written production estimate before they have to ask for one, and the replacement is verified against the order before it goes back in the opening.

Scope

Covers an in-service insulating glass unit reported fogged after the original installation is complete and closed out, from the customer's first report through the reglaze and file closure. Applies regardless of whether the unit is inside a manufacturer glass warranty, a shop workmanship period, or being replaced outside any warranty at the customer's cost.

Does not cover the field diagnosis of whether reported fogging is a between-the-panes seal failure or ordinary surface condensation, which the condensation between panes decision tree owns; this procedure assumes that diagnosis is already made before step 1 opens a file. Does not cover day-of-installation warranty education and registration, owned by the customer walkthrough and warranty handover SOP. Does not cover receiving inspection on a full order delivery, owned by the delivery inspection and damage claim SOP; step 6 borrows its inspection discipline for a single unit rather than restating it. Does not cover the physical reglaze technique, owned by the fogged insulated glass unit replacement technique HowTo.

Roles and responsibilities

The handoff most likely to fail is the second row. A field tech who photographs the failure but not the spacer code has handed the claims administrator a packet that cannot be submitted, and the file sits until someone notices.

The procedure

1. Intake the report and open a file the same day. Capture the address, the customer's description, and search the job history for the original install record and date. Acceptance: a file opened with either a matched install record or a documented reason none was found. Wrong looks like scheduling a visit off "my window looks foggy" with no address cross-check. Stop rule: no matching install record does not stop intake; it flags the file for the proof-of-purchase search in step 3 rather than being discovered later. Hazard: none at this step, it is a desk decision.

2. Run the field verification visit and confirm this is a real seal failure before treating it as one. Perform the interior wipe test and confirm the pattern persists between the panes, per the condensation between panes decision tree's own diagnostic step, then photograph the spacer-bar etch code under raking light and the exterior sash configuration. Acceptance: a legible spacer code photograph and a confirmed between-pane failure state, both recorded the same visit. Wrong looks like a spacer photo shot straight-on under flat light, which reads as a gray smear instead of a legible code. Stop rule: an illegible code photo is reshot before the tech leaves, not filed blurry and flagged for a return trip. Hazard: where the unit is upper-floor, this is ladder or lift work under the fall protection SOP if the drop reaches its trigger height; where the sash must be tilted to reach the code, handle it per the glass handling SOP, since a decade-old sash can carry a chip that was not there at install.

3. Assemble the claim packet against the manufacturer's required elements and treat any gap as a stop, not a note. The packet needs the failure photograph, the spacer code photograph, the exterior configuration photograph, and proof of purchase, typically the original sales order or invoice. Acceptance: all four elements present and legible before submission. Wrong looks like submitting three of four with a note to follow up on the fourth, which is how a claim gets bounced on its own clock instead of the manufacturer's. Stop rule: proof of purchase not found in the current job system escalates to the office for an archive search before submission; a claim filed without it is routinely rejected and the resubmission restarts the manufacturer's processing clock from zero. Hazard: none at this step.

4. Submit the packet and log the reference number the same business day. File through the correct channel for that manufacturer, whether a dealer portal, a direct homeowner submission, or a dealer-rep phone intake, and record the claim or reference number against the file. Acceptance: a reference number recorded within 3 business days of submission. Wrong looks like submitting and moving on with nothing recorded to check against later. Stop rule: no reference number within that window triggers a follow-up call rather than waiting silently. Hazard: none at this step.

5. Set the review cadence and put the production estimate in writing before the customer has to ask for one. State the claim date, the manufacturer's typical production window, and the resulting estimated install date, in a written message the customer keeps. Set a review date at a defined interval; a check-in every 2 weeks is a common starting point, tune it to your own claim volume. Acceptance: dated written communication in the file and a review date set. Wrong looks like a verbal "should be a few weeks" at the visit and nothing after. Stop rule: no written communication within 2 business days of the reference number being logged holds the file open at the coordinator's desk rather than moving to tracking. Hazard: none at this step.

6. Receive the replacement unit and inspect it against the order before it goes anywhere near the opening. Verify size, glass makeup, coating position, spacer color, and the tempered stamp where the hazard-location check required one, and inspect for shipping damage using the same arm's-length, raking-light method the delivery inspection and damage claim SOP applies to a full order, scaled to this single unit. Acceptance: every order field matches and no chip, scratch, edge crack, or seal gap is found. Wrong looks like accepting a unit that is the right size with the wrong coating surface because it is otherwise correct and the customer has already waited weeks. Stop rule: any mismatch or damage red-tags the unit, holds it out of the install schedule, and reopens with the supplier before a reglaze is scheduled; installing a wrong-spec unit to close the file faster produces a second claim on a unit the shop just installed. Hazard: still glass; handle per the glass handling SOP even for a single unit outside a full delivery.

7. Schedule and perform the reglaze per the fogged insulated glass unit replacement technique, and treat its own acceptance checks as this file's gate, not a formality. Confirm no gap at the sightline, correct coating orientation, and full operation on a cycle test where the sash is operable. Acceptance: every check in that HowTo's verification section passes. Wrong looks like closing this step because the visit happened, without confirming the HowTo's own acceptance was met. Stop rule: any failed check there holds the file open and the reglaze is corrected before closure; a fogged-unit file does not close on a visit completed, it closes on a reglaze verified. Hazard: covered in full by the reglaze procedure; cut-resistant gloves and eye protection stay on regardless of how routine the unit count looks.

8. Close the file: customer sign-off, related-defect scan, and record update. Confirm the unit is clear and operating with the customer, run the related-defect scan the decision tree describes (thermal exposure, missing sill pan, field modification, transit-damaged storage) against this project's other units, and update the manufacturer registration if the replacement carries its own serial. Acceptance: a signed or documented closure and a recorded scan result, even "no related condition found." Wrong looks like closing on the customer's thanks with no scan recorded, since the scan's value is in the next file, not this one. Stop rule: a second unit failure on the same project within 18 months reopens as a pattern claim per the decision tree's own rule, escalated to the owner rather than filed as a routine repeat. Hazard: none at this step.

The record this produces

The claim file carries: the install-record match or proof-of-purchase search result, the verification visit's failure confirmation and both photographs, the four-element packet as submitted, the reference number and submission date, every written customer communication with its date, the receiving inspection result against the order, the reglaze completion and its verification checks, the customer closure record, and the related-defect scan result.

It stays on the job record, not a separate warranty log. The coordinator reads it every 2 weeks against the review date. The claims administrator reads the reference number when the manufacturer calls with a production update. The owner reads the related-defect scan across a project's other files when a second claim comes in, since that is the only place the pattern becomes visible before a third unit fails.

Worked pass: a ten-year-old install, one failed unit

A customer reports a foggy bedroom window on a job installed a decade back. Step 1 opens the file the same day; the address search finds no matching job in the current system, which only holds six years of history after a software change. The file is flagged rather than stalled.

Step 2 confirms the between-pane failure on the visit; the spacer code photographs clean under raking light on the first attempt, no reshoot needed.

Step 3 fails on its first pass, exactly as the stop rule anticipates for a job outside current system history. No proof of purchase is on file. The escalation runs to the office, which locates the original paper invoice in archived records from that installation year. The packet is completed with the recovered invoice and submitted whole, not with three of four elements and a promise to follow up, the version of this step that gets a claim bounced.

Step 4 logs a reference number within 2 business days. Step 5 sends the written production estimate the same week, review date set at the 2-week interval, and the replacement ships roughly six weeks later, consistent with what the customer was told.

Step 6 finds the replacement matches every ordered field with no shipping damage. Step 7's reglaze passes its own acceptance checks on the first visit. Step 8 closes with sign-off and a related-defect scan finding nothing else on this project, since the original install used a proper sill pan and the affected unit sits on a shaded north elevation with no thermal-cycling exposure. That result is recorded as "no related condition found" rather than left blank, so the next reader knows the scan happened.

The step 3 failure is worth keeping in mind on any job older than the current record system. It is the easiest step in this whole procedure to skip past by improvising, and the improvisation is exactly what costs the manufacturer's clock a full restart.

References

1. See related: the condensation between panes decision tree, owning the surface-versus-seal diagnosis in step 2 and the related-defect scan criteria in step 8.
2. See related: the delivery inspection and damage claim SOP, whose receiving-inspection method step 6 applies at single-unit scale.
3. See related: the fogged insulated glass unit replacement technique HowTo, owning the reglaze procedure and its checks in step 7.
4. See related: the glass handling and two-person lift SOP, for crew size and handling method throughout.
5. See related: the fall protection for upper-floor window work SOP, for visits above the construction trigger height.