A shower door that leaks weeks after installation is usually reporting a decision made during it. The joint that was sealed before the door was adjusted, the fixing that was never sealed, the threshold that drains away from the shower: none of them produces water on the day of the handover test, and all of them produce water later. Recognising which mistakes behave that way is what makes a short handover test meaningful rather than reassuring.
The mistakes that matter are not the obviously wrong ones, which are caught during installation, but the ones that are technically complete and functionally wrong. Sealing a joint before the door has been adjusted is the clearest example: the door moves afterwards, the sealant does not, and the joint opens where it was meant to be closed. Applying silicone to a moving edge belongs to the same family, because the profile can no longer flex and water is trapped behind it instead of being directed back into the shower.
The second family is about water leaving the enclosure. A threshold that is not level, or that is level but drains to the wrong side, sends water where the enclosure cannot stop it. A fixing that was made without sealing leaves a hole rather than a fixing. A track fixed to a surface that moves under load shifts over time, and the shift changes the door's position relative to its seals without anything appearing to have failed.
Mapping the mistakes to the installation sequence is what makes them preventable. Errors of position are made when the enclosure is set out: the base, the track and the fixed panel. Errors of containment are made when the edges are treated: the seals, the joints and the drainage path. Errors of finish are made at the end: sealing before adjustment, adjusting after sealing, and the tightening that closes a gap by moving a part away from where it was set. Each group has its own test, which is why a handover that only runs water for a minute misses two of the three.
Water behaves differently in the first hour and the first month, and the difference is almost always timescale rather than volume. A handover test with a new seal and a clean threshold will not show a slow path behind a wall, a joint that only opens once the door has settled, or a drainage path that is blocked by the debris of construction. The table below is the practical version of that idea.
| Mistake | Why it stays hidden at handover | First sign later |
|---|---|---|
| Joint sealed before the door was adjusted | The sealant looks continuous on the day and the door has not yet moved | A crack in the sealant line, then water at that joint |
| Fixing made without sealing | Nothing is visible once the plate covers the hole | Damp patch or staining on the far side of the wall |
| Threshold level or drainage direction wrong | A short test does not build up enough water to show which way it runs | Water at the ends of the run rather than along the edge |
| Track fixed to a surface that moves | The movement only appears under the door's weight and use | The door starts to drag or the seals change contact |
| Seal fitted without checking contact along its length | The profile looks correct from the front | A leak at one point of an otherwise dry edge |
| Silicone applied to a moving edge | The profile appears sealed rather than bonded | Water appearing behind the profile, or a stiff door |
The one mistake that the table cannot capture is a joint that only fails under a particular spray direction. That is not an installation error at all, and it is covered by the diagnostic route rather than by the installation sequence, which is why it is worth separating before parts are ordered.
Several of these mistakes are recoverable without removing the enclosure, and the order in which they are approached matters more than the speed. Adjustment comes first, because a door that is not in its final position makes every seal test meaningless. Contact along the edges comes second: the seal can be re-seated, trimmed or replaced, and the joint it makes with the glass can be checked along its length rather than at a point. Drainage comes third, since a clear path changes where standing water goes.
Adjust, then seal contact, then drainage, then the static joints - and only then re-test with the same method used at handover, so the two results are comparable. Re-sealing a static joint is the last step rather than the first, because any movement after sealing reintroduces the original fault. Where a joint has to be reopened to be corrected, the old sealant comes out completely rather than being covered, which is slower and is the difference between a repair and a repeat.
Three mistakes cannot be corrected in place. A base or threshold that has been set to the wrong level has to be reset, which means the enclosure comes out. A fixing that cannot hold, or that was made into a substrate that cannot carry the door, has to be resolved at the wall rather than at the door. And a drainage path that was built the wrong way cannot be reversed by working on the visible surface, because the water is already travelling where the construction sends it.
Recognising that boundary early protects the parts that are still correct. A door that is removed for a base correction is a door that can be reused; a door that is adjusted repeatedly in an attempt to fix a drainage problem usually ends up needing seals as well. Where the symptom has already appeared rather than being anticipated, the shower door problems and fixes guide routes it by symptom, the water containment guide explains how the enclosure is meant to hold water as a system, the threshold slope guide covers the base, and the hinge-side leak guide and the bottom leak guide cover the two edges where installation mistakes most often show. Replacement seals are listed by profile in the shower door seals collection.