QR codes for museums, galleries and heritage trails

The job is a cross-product, not a list

Thirteen exhibits in two languages is not thirteen codes and it is not twenty-six unrelated ones either, it is a cross-product: every exhibit, times every language, each pairing pointing at its own page. That shape is what makes museum work harder than a single-language poster run and easier to get wrong. Number the exhibits once, list the languages once, and every code follows a rule instead of being typed from scratch.

Doing today's job with the tool that exists

There is no dedicated multilingual set-builder yet, and it is better to say that plainly rather than pretend otherwise. What does the job today is the bulk generator, which takes a pasted list and turns it into a whole batch of named, downloadable codes in one pass. Build the cross-product once, in a spreadsheet or a text editor, then paste it in.

Per-exhibit scan analytics is the sellable insight

The codes are the delivery mechanism. The thing a curator cannot get anywhere else is the answer to which exhibits people actually stop for. A collection has a hierarchy of attention that nobody in the building can currently see: the object every visitor photographs, the one nobody reads past the headline of, the language split that says your Italian-language visitors cluster around three specific cases. That is a genuinely new instrument for interpretation planning, not a side effect of having codes at all, and it is only available on dynamic codes because a static code cannot be counted.

Be precise about what a scan proves, because overselling it costs trust. A scan is a phone pointed at a label near an exhibit, aggregated with the date, a rough location and a device class, and no scanner is ever identified: no cookie, no device ID, nothing that links two scans to one visitor, and no geography is ever named below five scans in a window. It tells you engagement by stop and by language. It does not tell you dwell time, whether the visitor read to the end, or anything about who they were, and it should never be reported as if it does.

Labels that survive a gallery, not a desk

An interpretation panel is read from further away than a business card, under lighting a print shop never sees, by someone who may not want to touch the panel to get closer. Three things decide whether the code actually scans in that room rather than just looking correct on a proof sheet.

Full sizing detail, including how distance and quiet-zone margin interact, is in the print size guide, and the trade-off between correction level and code density is in the error correction guide.

No app, no barrier

For a public-sector or funded venue, accessibility is not a nice-to-have, it is a condition someone will actually check. A QR code that requires downloading a scanner app is a barrier by definition: it excludes any visitor who won't install something for a single visit, and it adds a step between curiosity and content that a genuinely accessible interpretation scheme should not have. Every phone made in roughly the last eight years reads a QR code directly in its camera app, no download, no account. That is the whole pitch to a visitor: point the camera you already have open.

Budget: pilot free, commit for £9 a month

Static codes are free, unlimited and never expire, and they're the right choice while a collection is still being interpreted or a gallery is being trialled: paste the list, download, print, done, no account needed. The moment you want to repoint a mislabelled exhibit without reprinting the panel, or you want the scan data that makes the per-exhibit engagement picture possible, that's a dynamic code, and dynamic codes need an account. Three are free, which is enough to pilot one small gallery or a single trail stop end to end before deciding anything. A thirteen-exhibit, two-language collection is twenty-six codes, which is past the free allowance and needs Pro at £9/month, one of the smallest line items an interpretation budget will see all year, and considerably less than the labour of retyping labels by hand across accounts because nobody looked at the plan first.