Why Won't My Barcode Scan? 12 Causes, Ranked by How Often They Happen

Updated 2026-09-07 · ParityLabel editorial

The barcode looks perfect on screen. You print it, hold the scanner to it, and nothing happens. Or worse: it scans on your phone, sails through your own tests, and then fails at a customer's goods-in gate. Scanning failures almost never come from a broken scanner — they come from the artwork, the print, or the surface. Here are the twelve causes, ordered by how often they actually turn up, with how to spot each one and what to change.

Start with the symptom

1. The quiet zone has been trimmed

Every linear barcode needs clear space at both ends so the reader can tell where the symbol starts and stops. It is empty white space, so it looks like wasted room, and it is the first thing to go when a layout gets tight.

• EAN-13 and UPC-A: 11 X-dimensions on the left, 7 on the right — roughly 3.6 mm and 2.3 mm at 100 % magnification. The human-readable digit printed at the bottom left sits inside that left margin by design; the > character you sometimes see on the right marks where the right-hand quiet zone ends.
• Code 128, GS1-128, Code 39: at least 10 X-dimensions each side, and GS1-128 additionally asks for no less than 2.54 mm.
• QR and Data Matrix: 4 modules all round for QR, 1 module for Data Matrix (GS1 application guidelines commonly ask for 2 to 3).

Measure from the outermost bar to the nearest printed element of any kind — a border, a rule, a block of colour, or the trim edge. If anything intrudes, widen the space rather than shrink the symbol.

2. The symbol is below 80 % magnification

EAN-13 has a nominal size of 37.29 × 25.93 mm at 100 %, which corresponds to an X-dimension of 0.330 mm. GS1 permits scaling between 80 % and 200 %, so X may range from 0.264 mm to 0.660 mm. Anything below 80 % is out of specification for retail point of sale. Distribution symbols carry their own, larger floors: ITF-14 and GS1-128 on cases are normally specified at an X-dimension of 0.495 mm or above.

The quickest test: an EAN-13 narrower than about 29.8 mm overall has been scaled below 80 %. If it will not fit, reduce the bar height (cause 3) or move to a smaller symbology such as EAN-8 or a 2D code — never squeeze the width.

3. The bars have been cut short

Bar height does not matter for a careful handheld scan, but it decides whether omnidirectional scanning works. The bioptic scanners at a supermarket checkout sweep several beams across the symbol at once; if the bars are short, no single beam crosses the whole code. EAN-13's nominal height is 22.85 mm. Cutting it down — "truncation" — is not recommended by GS1 for retail symbols, and in practice failure rates climb sharply once you go below roughly two-thirds. Handheld warehouse scanning is far more tolerant.

4. Colour and contrast

Most linear scanners illuminate with red light around 660 nm. Red ink reflects that light, so to the scanner a red bar is as bright as the white space beside it and the symbol simply is not there. The same applies to orange, gold and pale pink. Dark blue and dark green bars are usually fine.

The formal measure is symbol contrast, graded under ISO/IEC 15416. A quick desk test: look at the symbol through a piece of red transparent film — if bars and background still separate clearly, a red-light scanner will cope. Keep the barcode black on a light background as a rule, and do not let it follow the brand palette.

5. The X-dimension is not a whole number of printer dots

This is the classic "perfect on screen, useless on the label" failure. A thermal printhead is a row of fixed dots: one dot is 0.125 mm at 203 dpi and 0.0847 mm at 300 dpi. The X-dimension has to be a whole multiple of that. Specify EAN-13's nominal 0.330 mm on a 203 dpi printer and it needs 2.64 dots per module — impossible, so some bars come out at 2 dots and others at 3, and bar width deviation goes out of tolerance immediately.

Set the module width in dots, not millimetres: 2 or 3 dots at 203 dpi, 3 or 4 at 300 dpi. For EAN-13 on a 203 dpi printer, 3 dots gives X = 0.375 mm, about 114 % magnification, which is well inside spec and the usual choice. If you place a vector symbol, scale it by whole multiples only.

6. Ink spread on press

Flexo, offset and screen printing all push ink outwards, so bars print wider and spaces narrower than the artwork. Printers compensate at plate-making with bar width reduction, thinning the bars by a known amount so the printed result lands on target. If your labels come from a press, say explicitly that the artwork contains a barcode and ask what BWR is being applied. Under a loupe, the symptom is bars with soft, spreading edges and spaces that look filled in; on a verifier it shows up as poor decodability.

7. Wrong check digit or wrong data

A symbol that scans but returns an unknown code — or one that a scanner refuses outright — usually carries bad data rather than bad printing. EAN-13, UPC-A and ITF-14 end in a mod-10 check digit computed from the preceding digits; edit the digits without recalculating and the code is invalid. Let the generator compute it, and validate codes that arrive from supplier spreadsheets. When the data comes out of Excel, check that leading zeros survived and that no 13-digit code turned into scientific notation — the full list of traps is in the bulk barcode checklist.

8. The scanner is not configured for that symbology

• A laser scanner will never read a 2D code. QR, Data Matrix and PDF417 need an imaging scanner. This is the most commonly misdiagnosed "bad barcode" of all.
• Scanners ship with several symbologies disabled. ITF-14, Code 93, GS1 DataBar and Codabar are frequently off by default, and Code 39 full ASCII is normally off — so a Code 39 symbol containing lowercase letters will not read until you enable it. The fix is a configuration barcode from the scanner manual.
• GS1-128 uses an FNC1 start character and application identifiers; Code 128 switches between subsets A, B and C. Get these wrong and the decoded string comes back with extra or missing characters rather than failing outright.

9. The label is on a curved surface, the wrong way round

On bottles, tubes, cans and small drums the bars must run parallel to the axis of the cylinder — the "picket fence" orientation. Every bar then sits at the same distance from the reader as the beam sweeps across. Turn the symbol ninety degrees into the "ladder" orientation and the beam has to travel around the curve, so the bars at each end fall away from the focal plane and distort. The tighter the diameter the worse it gets; below roughly 60 mm, use picket fence and keep the symbol narrow.

10. Glare from lamination, and transparent materials

Gloss laminate, UV varnish and clear PET all reflect the scanner's own illumination straight back at it, so the reader sees a wash of white. Tilt the label about 15° and try again: if it reads, glare is your problem. Specify matt laminate over the barcode area, or leave that panel unvarnished. On clear film, print a white ink layer beneath the symbol — a black barcode on transparent material applied to a dark bottle has no background at all.

11. Thermal print quality: darkness, printhead, ribbon, speed

• Darkness too high bleeds the bars into the spaces; too low leaves them grey and broken. Move one step at a time from the middle of the range.
• A failed printhead element shows as a white line running through every label in the same place, slicing the symbol. Print a solid black test label to find it.
• Ribbon mismatch: wax ribbon on synthetic or PET stock has poor adhesion and the print rubs off. Match ribbon to substrate — resin for films, wax-resin as the general compromise.
• Excessive print speed smears bar edges. Dropping one speed setting often fixes marginal symbols outright.

12. Damage, and how you hold the scanner

Abrasion in transit, direct thermal labels fading in sunlight, tape pressed over the symbol, and codes applied across a box seam are all use-phase problems with no artwork fix — choose a durable material and specify where the label goes. One habit worth changing: do not hold the scanner perpendicular to the label. Straight-on is where specular reflection is strongest; a 10–15° tilt reads more reliably.

A five-minute triage

1. Try a phone first. If the phone also fails, suspect the data or the symbology (7, 8) or a badly degraded print (6, 11). If the phone succeeds, carry on down this list.
2. Measure the quiet zones and the overall width. An EAN-13 under 29.8 mm wide, or anything printed within the margins, is your answer (1, 2).
3. Look at the colours. Non-black bars, dark background, gloss laminate (4, 10).
4. Identify the print process. Thermal: check dots, darkness and printhead (5, 11). Press: ask about bar width reduction (6).
5. Look at where the label sits. Curved surface, seam, under tape (9, 12).
6. Before a production run, verify. Grade the symbol to ISO/IEC 15416 with a verifier. Retail symbols are generally expected to reach at least Grade C — many retailers and trading partners specify Grade B or better — and the report tells you which parameter is closest to the edge, which repeated scanning never will.