Pre-Existing Seal Failure Discovered During Cleaning - The Response Standard

Purpose

This SOP guarantees that when a technician notices a hazy or foggy insulating glass unit mid-clean, the response is confirm, disclose, and document, in that order, before the customer has a chance to conclude on their own that today's visit caused it. A fogged unit is invisible to most homeowners until something makes them look closely, and a cleaning crew scrubbing the same pane is often the first person to look closely enough to notice. Left unmanaged, that timing coincidence becomes a claim: the customer remembers the window was fine before the crew came, sees fog after, and connects two events that share nothing but a calendar date. The procedure exists to break that false connection at the moment it forms, not weeks later in a phone call the office did not see coming.

Scope

Covers a between-the-panes seal failure, a fogged or hazy insulating glass unit, discovered during a routine cleaning visit, from the moment it is noticed through confirmation, disclosure, and record close. It does not cover the physical confirmation test itself or the plain-language explanation of how a sealed unit fails, both owned in full by the condensation-between-panes decision tree, which this SOP's step 2 routes into directly rather than restating. It does not cover repairing or replacing the failed unit, a glazing job outside a cleaning-only shop's trade, this SOP ends at a referral, not a repair. It does not cover a haze that turns out to be on a glass surface rather than trapped inside the cavity, whether from hard water or a cleaning-chemical residue, both of which are removable and route to their own response rather than this one.

Roles and the handoff between them

The procedure

1. Stop cleaning that pane the moment the haze looks like it will not clear, and do not escalate product or effort to fight it before confirming what it is. Set the tool down and step back from the pane rather than reaching for a stronger chemical or a second pass with more pressure. Acceptance: no additional product or technique is applied to that pane before the confirmation test runs. Wrong looks like reaching for a stronger glass cleaner or working the same spot harder on the assumption it is just stubborn hard water. Stop rule: if a stronger product has already gone on the glass before the tech realizes it is not responding, stop entirely, note in the record that a product was applied before the pre-existing nature was clear, and run the confirmation test on the glass as it now stands rather than pretending the attempt did not happen. Hazard: an unconfirmed haze is not yet known to be safe for a stronger product, some low-e and other surface coatings react to acidic or abrasive treatment, so nothing beyond the shop's standard glass cleaner touches that pane until the source of the haze is confirmed, and if the coating side cannot be confirmed on-site, no escalated product goes on today regardless of how close the deadline feels.

2. Run the confirmation test per the condensation decision tree's own method before saying anything to the customer about what it is. Press a dry cloth firmly against the interior face for several seconds, then the exterior face, and check whether the haze pattern responds to either wipe. Acceptance: a confirmed between-pane result, unchanged after both faces are wiped dry, or a confirmed surface result, cleared or shifted by the wipe, recorded either way. Wrong looks like calling it a seal failure from a glance without running the wipe test, which is exactly the kind of claim a customer can contest if the haze turns out to be ordinary surface condensation that clears within the hour. Stop rule: if the result is ambiguous, a humid morning fogging every pane in the room, hold off on any pre-existing claim, finish the visit, and reassess later in the day once ambient humidity has settled rather than asserting a diagnosis the conditions cannot support yet. Hazard: none directly in the test itself, but where the pane sits above standard reach height, running the wipe test from a ladder or a water-fed pole platform still falls under the shop's standard access and fall-protection controls, a quick test is not an excuse to skip a stable base or a spotter.

3. Photograph the confirmed result before moving on, closing out the pause from step 1. Capture both faces after the wipe test and the haze pattern itself, with a location note, room, elevation, unit identifier. Acceptance: photos exist showing the wipe-test result, not just the fog on its own. Wrong looks like a single photo of foggy glass with no evidence the test was actually performed, which proves the fog exists but proves nothing about which side it sits on. Stop rule: none, this step always completes once step 2 has produced a result. Hazard: none, a camera or phone in hand carries no risk of its own beyond the same access controls already covered in step 2.

4. Disclose to the customer before the crew leaves the property, in plain terms. State that the moisture is sealed inside the unit, that no cleaning tool reaches it, that the visit did not cause it, and that cleaning the exterior face will not fix it and may make the fog more visible by contrast against clearer glass around it. Acceptance: the customer has heard the explanation and, where possible, seen the photo before the crew departs. Wrong looks like a note left on the invoice with no verbal conversation, which reads as the shop quietly disclaiming responsibility rather than explaining a real, common condition. Stop rule: if the customer is not present, this goes out by call or photo text the same day, and if that does not hold before the crew leaves the property, the office places the call before end of day rather than letting it wait for the next scheduled contact. Hazard: none, this step is a conversation, though it happens with the tech's tools already set down per step 1, not balanced in one hand while the other holds a phone photo up to the customer.

5. Complete the rest of the cleaning on that pane's reachable surfaces, resuming the work paused in step 1. Clean the interior and exterior faces normally; only the trapped moisture itself is out of scope. Acceptance: both reachable surfaces are cleaned to the same standard as the rest of the job. Wrong looks like skipping the whole pane because part of it cannot be fixed, which shortchanges the customer on the surfaces that are cleanable and were paid for. Stop rule: none. Hazard: the same access and coating considerations that applied to every other pane on this job apply here too, a confirmed seal failure does not change what is safe to put on the glass surface, standard product only, since the confirmation only ruled out a surface cause, it did not rule in an unconfirmed coating.

6. Offer a glazier referral and log the answer. Ask directly whether the customer wants a referral for repair or replacement of the unit. Acceptance: a stated yes or no recorded in the file, with contact information passed along if yes. Wrong looks like assuming disinterest and never asking, which leaves a customer who did want help with no next step and a shop with no record that help was offered. Stop rule: none. Hazard: none.

The record this produces

• Confirmation: the wipe-test result on both faces, and the photo pair showing it.
• Location: room, elevation, and unit identifier for the affected pane.
• Disclosure: who was told, how, and when.
• Referral: yes or no, and the contact shared if yes.

This record is what the office reads if the customer calls back later asking why "the window the cleaners looked at" is still foggy, the answer is already on file, dated to the visit, not reconstructed from memory during the call.

One run of this standard, filled in

A semiannual residential account, kitchen window over the sink, clean at the prior visit six months ago.

• Step 1: the tech notices persistent haze along the lower third of the pane mid-pass and, assuming it is heavy hard water from sink splash, reaches for a stronger acidic glass cleaner and works it in before checking anything else. Fail. The acceptance calls for no escalated product before confirmation; the stronger cleaner went on first. Stop rule taken once recognized: the tech stops entirely, notes in the record that a stronger product was already applied, and runs the wipe test on the glass as it now stands.
• Step 2: the wipe test on both faces shows no change, the haze is unchanged after both are dried, between-pane failure confirmed. The stronger cleaner made no difference either way and left no visible mark on the glass.
• Step 3: photos of both wiped faces and the haze pattern are taken, with a note identifying the window as "kitchen, over sink, interior/exterior."
• Step 4: the homeowner is in the kitchen. The tech explains the sealed-unit fog in plain terms, shows the photos, and is direct that today's visit did not cause it.
• Step 5: both faces of the pane are cleaned normally for the rest of the visit; the trapped moisture is left alone.
• Step 6: the homeowner asks for a referral. A local glazier's contact is logged in the file and passed along before the crew leaves.

The stronger cleaner reached for in step 1 did not damage anything here, but it easily could have on a coated pane, and it added a step to the record that should not have existed, an application made before the source was known. Confirming first costs two minutes; explaining an unnecessary chemical application after the fact costs more than that in trust.

References

• See related: the condensation between panes decision tree, which owns the wipe-test diagnostic method used in step 2 and the plain-language explanation used in step 4.
• IGMA (Insulating Glass Manufacturers Alliance) TB-3001, Guide to Understanding Sealed Insulating Glass Unit Failures.
• ASTM E2188, Standard Test Method for Insulating Glass Unit Performance.