Measure an out-of-square opening by taking the width at the top, middle, and bottom, the height at both sides and the center, and the wall plumb at the mounting line, then recording the worst case in every dimension. An out-of-square opening is normal; the mistake is hiding it behind a single reading. The controlling rule is to give the fabricator the full range, not an average. The complete planning framework is in the complete measurements guide.
The out-of-square condition is measured at the mounting lines, not at the floor or the ceiling. The top rail or hinges mount at one line and the bottom seal at another, and the deviation at those two lines decides the fit. Measuring the floor and the ceiling can show a different number from the line the door actually uses, so the mounting line is the reference for every reading.
Confirm that the opening is finished: tile set, trim installed, and the tub or curb in its final position. An out-of-square reading is only useful if it describes the surface the door will actually meet, and a pre-tile reading describes a different opening. Then confirm the tools: a metal tape that does not flex, a level, a straightedge, and a way to record the numbers on a sketch.
The opening is out of square when the top and bottom widths differ, or when a level held on the wall shows plumb deviation. Both conditions are common, and both are handled by recording the range rather than forcing the opening to look square.
The level and the tape answer different questions: the tape gives the width and height, and the level gives the plumb and the level. A corner can be out of square even when all four widths match, because the walls lean rather than bow, and only the level catches the lean. Both tools are needed, and both readings go on the same sketch.
The straightedge is the tool that finds the bow: a wall can measure the same at the top and bottom and still curve in the middle, and the curve is exactly what the glass must clear. Hold the straightedge flat against the wall and measure the gap under it at the mounting line.
The photos are taken at the same time as the numbers, because the site will not wait for a second visit: the tape measure is held in the frame, the level is visible, and each corner is photographed straight-on. A photo without a scale tells the fabricator nothing, and a backlit shot hides the wall condition. The photo set is the visual record of the numbers.
The comparison is between like numbers: the same mounting lines, the same tools, and the same person where possible. If the two sets agree within a small margin, the opening is stable and the numbers are usable; if they differ, the opening is still settling or the method has an error. The stable set is the one the order uses.
Take the readings twice on two different days and compare, because a measurement that changes between readings means the opening is still moving or the method is inconsistent. The numbers that agree are the numbers to use.
The build tolerance is stated by the product, and it is the line between a fit and a rework. A deviation inside the tolerance is handled with the glass reveal and the hardware adjustment; a deviation outside it needs the opening corrected before the door is made. Comparing the measured range with the stated tolerance is the compatibility check the quote depends on.
An out-of-square opening is buildable within the design's tolerance, but the tolerance is finite. A wall that bows beyond the glass reveal or the hardware's adjustment range cannot be shimmed into a square fit, and the opening must be corrected first. The product's stated mounting requirements define the limit, so the measured deviation is compared with that number before the order.
Safety follows the same rule: a glass panel cut to the average of an out-of-square opening can bind at the narrow end and gap at the wide end, and a binding panel stresses the hardware.
A panel cut to the worst-case width clears the narrowest point and seals at the widest, while a panel cut to the average does neither. The same logic applies to the height and to the wall deviation.
The build is designed around the worst case in every dimension at once: the narrowest width, the tallest height, and the largest deviation all act on the same panel. A door that clears the width but binds the height, or clears both but misses the deviation, still fails. The full set of worst cases is the specification the build follows.
The worst case is the only number that describes the full condition, and the build is designed around it.
The worst case is not pessimism; it is the number the build needs to fit safely.
The out-of-square mistakes are all versions of hiding the truth: one reading, an average, a rough-in, or no photo. Each one lets the opening look better than it is, and the door is built to the fiction. Recording the full range is the opposite habit, and it is the habit that produces a door that fits the real site.
The verification is done against the same mounting lines as the original readings, so the comparison is between like numbers. The second set of readings confirms the opening has not moved during the planning or the lead time, and it catches the change while the door is still in the box. The re-check is the last gate before fabrication.
The out-of-square record is part of the wider measurement framework, covered in the complete measurements guide.
The out-of-square habit is also the quality habit: the opening is measured as it is, not as it should be, and the door is built to the real condition. A buyer who records the full range gets a door that fits the site, and a buyer who hides the range gets a door that fits a fiction. The honest record is the standard for a custom fit.
The dimensions above describe the opening's size; the diagonals describe its shape. A rectangle has two equal diagonals, and a shape whose widths and heights agree at every point but whose diagonals do not is a parallelogram - the opening leans rather than being too small or too large. That distinction matters because a lean and a size difference are corrected in different ways, and no set of widths or heights can separate them on its own.
Take the check on the finished face, from the same four reference points used for the other measurements:
Writing the result down is what makes it usable. Record each diagonal as its own number, then state the difference and the direction in words - for example, that the top-left corner sits back relative to the other three - rather than leaving a bare figure that could be read as a width. Put the difference on the same sketch as the width, height and deviation notes, so that every dimension carries the position it was taken from and the worst case in each direction is visible at a glance.
The diagonal check verifies the record; it does not replace the measurements above, and a single agreement between diagonals is not evidence that the opening is ready - the widths, heights and level readings still have to be taken. What the numbers are then built to is a question for the door maker, and the custom door measurements planning guide covers that side of the process together with the non-standard dimensions range for openings outside a standard size.
For the complete decision path, use the shower-door measuring and fit guide.
It depends on the design's tolerance, which the product states. Measure the deviation and compare it with that number; a small bow is normal, and a large one needs the opening corrected first.
No. Give the three widths and the worst case, because the glass must fit the narrowest point. The fabricator needs the range, not a number that hides it.
The level checks plumb and level, and the tape checks width and height. Together they describe the squareness; neither alone is enough.
Yes. Photos of the corners and the widest and narrowest points show the condition the numbers describe, and they help the fabricator confirm the build. Include a tape measure in the frame.
The record closes the loop: the sketch shows the opening, the numbers show the range, and the photos show the condition the numbers describe. The fabricator reads all three together, and the quote and the build follow the same evidence. The complete record is the standard for an out-of-square order.
Record the full range, compare it with the tolerance, and send the sketch and photos with the order. Request a Custom Shower Door Quote with the worst-case numbers in hand.