QR Codes: When Static Beats Dynamic — and When It Doesn't
Every QR code project starts with the same wrong question: "which generator should I use?" The right first question is whether the code is allowed to ever change its destination. That single property decides the whole architecture, and picking wrong costs real money — usually after the code is already printed on two thousand flyers.
The two families
A static QR code encodes your destination directly. Scan it and you go straight to the URL, the Wi-Fi credentials, or the payment details baked into the pattern. The code is self-contained. No middleman, no subscription, no expiry. But the target is frozen the moment you export the PNG.
A dynamic QR code encodes a short redirect URL owned by the vendor. You scan it, land on their server, and they forward you to the current destination. This is why dynamic codes can change destinations, log scan counts, and geo-target. It is also why every scan routes through infrastructure that bills a subscription — stop paying, and printed codes all over the world quietly die.
Where static wins, permanently
- Wi-Fi credentials:
WIFI:T:WPA;S:NetworkName;P:password;;printed at a reception desk. Nothing about this belongs on someone else's server. - Tickets and boarding passes: payload is cryptographically signed and scanned offline anyway.
- Payment details: a bank transfer QR carries IBAN and amount. Renting that from a QR SaaS vendor makes no sense.
- Anything laminated or engraved: machine labels, equipment panels, industrial settings. The code must outlive the company that made it.
For these, a plain QR code generator that exports a self-contained PNG is the entire requirement. Zero maintenance, zero cost, zero risk of the vendor shutting down.
Where dynamic wins
- Marketing campaigns: the landing page will get redesigned, AB-tested, or sunset. With dynamic you swap the destination; with static you reprint.
- Restaurant menus and store signage: prices change weekly, and a code with built-in analytics tells you the laminated code on table 6 actually still gets scanned.
- Printed URLs that were mistyped: static means the error is permanent. Dynamic lets you fix it remotely.
The trade is real: dynamic costs a recurring fee, every scan is an extra redirect (a few hundred milliseconds of latency), and the vendor becomes a dependency. If they shut down or you churn, the physical codes are already distributed — you own a wall of dead QR codes. Budget for that risk, and if you switch vendors, test that their redirect preserves UTM parameters so analytics don't break.
The decision rule
Ask: will this exact destination still be correct in two years? If yes — static. If no, or if you need scan analytics — dynamic. There is almost no overlap. "Dynamic because it feels more professional" is not a reason; you're buying a subscription for nothing. "Static because it's cheaper" on a marketing campaign is false economy; one landing-page migration equals the price of several years of subscription in reprint costs.
The static-generation mistakes that matter
Most bad printed QR codes fail for reasons fixed in the source data, not the printer.
- URL with special characters: query strings with
&, spaces, or non-ASCII characters must be percent-encoded before encoding. Encoding?ref=summer saleraw truncates the URL at the space on many scanners. Pass it through a URL encoder first, then generate the code. - Error correction left at default: QR supports ECC levels L (7%), M (15%), Q (25%), H (30%). Printed material that will get scratched, sun-bleached, or sticker-covered needs H. A code on a warehouse shelf at level L fails within a season.
- Insufficient quiet zone: the white border around the code is not decoration. Four modules minimum. Designers love to crop it, and the code that scans perfectly on screen dies on paper.
- Contrast failure: QR scanners read luminance, not hue. White code on a dark background inverts the pattern and fails on many devices unless the generator explicitly supports inverted codes.
- Wrong data type: encoding a Wi-Fi password, payment details, or calendar event as a plain URL string loses the structured action. Use the correct payload schema (
WIFI:,bitcoin:,SMSTO:...), and verify the payload structure with a JSON validator if you build the JSON-based payload types (like vCard or contact details) by hand — one malformed field and the scanner shows a raw string instead of opening the contact form.
Verify before you print
The moment a QR code goes to print, changes cost money. So the verification loop belongs before the PNG export, not after: scan the generated code with two different devices (iOS and Android), confirm the decoded payload matches character-for-character, and check the destination URL resolves — a 404 behind a perfect-looking code is still a dead QR. For the destination URL, run it through a URL redirect checker if there's any redirect chain (short links, marketing redirects) — QR scanners don't follow expired or broken chains gracefully.
One last practical note: for any code going onto physical media, print a test at the actual size and material. A code that's flawless at 300px on screen can be unscannable at 2cm wide on textured kraft paper. Minimum scanable size is roughly scan distance divided by 10 — a code meant to be read from 1.5 meters away needs to print at least 15cm on a side.