If a 100 mm reference line prints as 96 mm at a 100% setting, the calculated correction is 104.17%. Treat it as the next test setting, then measure again.
Measure something with a known intended size
Use a reference line or a box whose original dimensions are stated in the file. Measure between its intended endpoints with a physical ruler. Avoid a thick border if you cannot tell whether its intended measurement includes the stroke.
Measure horizontally and vertically. If one 100 mm line prints as 96 mm while the other prints as 99 mm, one uniform scale cannot correct both. Inspect the document and print settings before applying a single percentage.
Calculate the correction
New percentage = current percentage × intended length ÷ measured length
For a print made at 100% with a 96 mm result where 100 mm was intended, use 100 × 100 ÷ 96 = 104.1667%. If the first print was instead made at 95%, the formula gives 95 × 100 ÷ 96 = 98.9583%. The current setting matters.
This is an arithmetic model: it assumes the same proportional response on the next print. The application or driver may impose another fit operation, cap the size, or alter the printable region. Use a new test sheet to see whether the model matches your output.
Check for clipping after enlargement
Making a page bigger also moves its outer content farther from the center. A correction that fixes a central reference line can push a border beyond the printable area. Compare the final width and height with the usable area before printing a full batch.
For example, a 200 mm wide area enlarged by 104.1667% becomes 208.33 mm wide. It cannot fit inside a 200 mm printable width. A larger sheet or a different layout may be necessary.
Keep a useful record
Record the document, paper size, percentage, measured horizontal length and measured vertical length. If you change the paper, application or printer, repeat the test instead of assuming that the correction transfers.
For an initial geometric fit, start with the print scale calculator. For a physical mismatch after printing, use the measurement method on this page. They answer different questions.
Do not chase more precision than you can measure
If your ruler cannot reliably distinguish 99.8 mm from 100 mm, a correction calculated from that reading may add noise. Repeat the measurement and test one changed setting. A percentage with four decimal places is not evidence that the printer or the measurement is that precise.
The A4 and Letter test PDFs include horizontal and vertical references plus a place to record the setup. For a correct-size print whose reverse is shifted, use the alignment workflow rather than changing scale.
Worked examples use the formulas stated on this page. These are mathematical examples, not physical printer tests.