QR code with logo wont scan
SmartQRCode editorial · Updated October 2026
A QR code with logo wont scan when the centre mark spends more error correction than the level can restore, or when the mark sits on a finder. Logo versus error correction H and a smaller mark at print millimetres are the levers. Confirm with a test scan of the real millimetre size, not a zoomed laptop.
Why a QR code with logo wont scan still shows finders
The three nested squares are how a camera locates the code. A logo in the centre is recovered by error correction, not by a special logo mode. If the logo covers more of the symbol than the chosen level can restore, finders may still be visible and the code still fails.
Levels are L about 7 percent, M about 15, Q about 25, H about 30. Higher correction survives more damage and makes a denser pattern. On a small print, density fights you.
Product-launch cartons often want a huge mark. That is a reprint problem, not a dashboard edit. Destination can change on hosted types; the logo cannot.
A Canva one-off without a logo is the right tool for a one-shot poster that only needs to open a URL.
Logo covering more than error correction H can restore
Cause: the brand mark is a wide rectangle that wipes a third of the modules, then someone picks L to "keep it looking open." Scuff on the carton spends the rest of the budget.
Fix: set H if you insist on a mark. Shrink the logo so data modules remain around it. Recreate from the with-logo generator. Do not paste a PNG over a finished QR in InDesign and call it error correction.
Confirm: print at final millimetres. Scan at about 30 cm. If a logo-free twin of the same payload scans and the logo twin does not, the mark is the cause. Reduce the mark 20 percent and print again. Stop when both phones pass.
Logo colliding with a finder square
Cause: the mark is off-centre, or a circular crop pulls a finder into the logo plate. Designers see "three corners still there" because two finders remain. Many cameras need all three.
Fix: keep the logo in the middle and smaller than the finder-to-finder gap. No frames that recolor a corner.
Confirm: on the press proof, you should see three intact nested squares. Cover the logo with a finger; if it suddenly scans, the mark was eating data or a finder. Move or shrink, reprint, test-scan.
Low correction L with a large centre mark
Cause: L is chosen because the preview "looked cleaner." L restores about 7 percent. A logo that looks modest on a 15 cm screen is huge at 2 cm on a label.
Fix: either drop the logo or raise correction and enlarge the whole symbol so modules stay resolvable. Uncoated stock and shrink wrap spend correction too; do not spend it all on branding.
Confirm: export two PDFs, L with logo and H with a smaller logo, both at the label size. Scan both. Ship the one that decodes on the formed pack, not the prettier flat proof.
Confirming a smaller mark with a test scan at print size
Cause: the laptop zoom lies. A mark that looks like 10 percent of the square is 35 percent of a 18 mm label.
Fix: rule the logo on a paper print. If it covers finders or most of the centre third, it is too big. Keep the quiet zone. Send SVG to the printer so the logo edges stay hard.
Confirm: three phones, real lighting, final substrate. If only the newest phone scans, the code is marginal; shrink the mark again. A plain website file without a logo is the control sample.
Finder collision under a mark also happens with round frames that recolor two corners in brand red. Those corners cease to be finders. Keep frames off the three nested squares. Correction L with a large centre is the second common pairing: L restores about 7 percent, which a wide lockup spends immediately.
Export PNG, JPG, SVG, or PDF from the generator. Do not flatten a logo in Photoshop on top of a raster QR and send that to flexo. The printer will see grey. Confirm by zooming until one module is a square with a hard edge. Then print at label size and scan three phones.
If brand insists on a mark larger than H can restore, put the mark beside the code, not in it. A logo-free symbol plus a printed wordmark is readable. A wordmark inside an unread square is neither.
H correction densifies the grid. On an 18 mm label that density can drop you below what the camera resolves even when the logo is legal. Enlarge the label, or drop the logo. Confirm both variants at the true millimetres, not at 400 percent zoom.
When a logo-free Canva file is the right one-off
If this is one poster and you will never count scans or change the URL, skip the hosted logo tooling. Export a high-contrast static square from Canva or iOS, no mark, at 2 cm or larger. Put the brand next to it in type.
If you need a mark and a destination that will move, keep hosted URL plus a small H-level logo, and reprint when the mark changes. Test-scan every new die. Visible finders are not a pass.
Questions this raises
No. A centre logo is recovered by error correction, not by a logo mode. If the mark covers more than the chosen level can restore, the code fails with finders still visible.
H restores about 30 percent, Q about 25, M about 15, L about 7. The logo plus scuff plus blur all spend that budget. Leave margin.
Finders locate the symbol. The data modules under the logo are what correction must rebuild. Visible corners do not mean the payload is recoverable.
When the mark has to be large for brand rules, or the print is small, or the stock is uncoated. A logo-free Canva one-off is better than a pretty unread square.
No. Logo is in the modules. You can change a hosted destination without reprinting. You cannot change the mark without new ink.
Keep going
The tools and playbooks this post refers to, one click away.
Make the code this post is about
$1.99 for 7 days, then a paid plan. Pick a type, brand it, and edit the destination whenever you like, even after it is printed.