QR codes in museums: adoption and best practices.

KEY TAKEAWAYS

The right model is a QR code on the wall card, not on the artwork — placed two to four inches from the label text, at 48–60 inches from the floor, with a one-line call to action that tells visitors what they'll get.
Print QR codes at a minimum of two inches square (5 cm) for typical gallery scanning distances of 18–24 inches, and never under 1.2 inches. Quiet zone and contrast matter more than fancy branding.
Use dynamic QR codes (a redirect link, not a static URL) so you can change destinations, fix mistakes, and track scans without reprinting a single wall card. Static codes are a procurement trap.
The first-scan problem is real: visitors won't scan an unfamiliar code unless a docent, a lobby sign, or a single landing QR explains what they get. Solve it once at the entrance, not at every label.
QR cleared the bar where native museum apps did not because it requires zero install, zero account, and zero commitment. 90% of museum native apps are used by less than 3% of visitors (nuseum.ai, 2025); QR scan rates on well-signed museum tours run an order of magnitude higher.

Five years ago, recommending QR codes for a museum tour would have been embarrassing. Adoption was low, scanning was awkward, and most visitors didn't have the camera-app shortcut. That changed during the pandemic and didn't change back. By 2026, QR is the load-bearing piece of every phone-based audio guide deployment we see — including ours — and the implementation details are what separate a tour that gets used from one that gets ignored. This guide is the practical version: where to put the code, how big to print it, what kind of code to generate, and how to solve the one problem that still trips up most institutions on day one.

I'm the founder of Convo, so we have a horse in this race. But QR isn't a Convo thing — it's the dominant delivery pattern across the entire phone-based audio category, used by essentially every serious vendor we get compared to. The choices below apply regardless of which platform a museum picks.

Why did QR codes win where museum apps didn't?

Because QR asks for nothing. A native app asks the visitor to find it in a store, accept an install of several hundred megabytes, grant permissions, and often create an account — at the threshold of a building they came to look at. A QR code asks the visitor to point a camera. That's the entire difference, and it's enough to change adoption by an order of magnitude.

The data on native museum apps is brutal. A 2025 analysis of 157 German and Austrian museum apps with measurable visitor counts found that 90% are used by less than 3% of visitors, with an average pickup rate of 1.3% across all apps (nuseum.ai, 2025). The best app in the sample hit 10.5%, and there was no correlation between museum size and adoption.

QR-based tours on well-signed phone platforms regularly clear that 10% ceiling, and the leading deployments run substantially higher. The mechanism isn't magic — it's that an app download is a five-step commitment and a QR scan is a one-step transaction. The visitor decides to engage with the tour in the same gesture as scanning, instead of deciding ten minutes earlier in a store.

Where should the QR code actually go?

On the wall card next to the label text, not on the artwork or the floor. The wall card is where visitors are already looking when they pause at a stop, and it's where the rest of the interpretive context lives. Floor decals get walked on. QR codes printed onto the artwork or its mount damage the curatorial frame. The wall card is the only surface that consistently makes sense.

Specific placement guidance that we've seen hold up across deployments:

Height: 48–60 inches from the floor. Tall enough for an adult to scan at chest height without bending, low enough for a seated visitor or a child to reach. Vary heights across a gallery if the visitor population is wide.
Distance from label text: 2–4 inches. Close enough to read as a unit with the label, far enough that visitors can frame the code without the label text confusing the scanner.
Lighting: Avoid direct spotlights that cause glare on glossy stock. Matte or satin finish reads better in low-light galleries. If a code is behind glass, test the scan from a typical visitor distance under actual gallery lighting before opening day.
Surface: Flat, non-reflective, and stable. Movable label stands work; tape-on stickers slumping off plexi do not.

One pattern to avoid: the temptation to put one QR code per object across an entire collection wing. Two hundred QR codes in one gallery confuses visitors and creates a sign-management problem. Most institutions are better served by a single landing-page QR at the gallery entrance that opens the tour, plus object-level codes only on the stops with bespoke interpretation. The MuseumNext write-up on QR resurgence makes the same point: avoid "overwhelming visitors with too many QR codes in a single space" (MuseumNext, 2025).

What size should a museum QR code be?

Minimum two inches (5 cm) square for a typical 18–24 inch scanning distance. The print industry rule of thumb is a 10:1 distance-to-size ratio — at 50 cm scanning distance, the code needs to be at least 5 cm per side (GoCreateQR, 2025). In practice, we recommend museums design to 2 inches as the floor and 2.5–3 inches as the comfortable default for low-light galleries.

What actually matters more than absolute size:

Quiet zone: The blank border around the QR pattern. Leave at least four module-widths of white space on every side. Scanners need this to find the code; a tight crop kills scan reliability.
Contrast: Dark code on light background, always. Reverse codes (white on dark) work in theory and fail in practice across enough phone cameras that we don't recommend them. Avoid trendy "gradient" codes for the same reason.
Print quality: A code printed at 150 DPI on uncoated stock will scan; a code printed onto fabric, woodgrain, or textured wallpaper often won't. Test on the actual substrate, not on office paper.
Brand mark in the center: Supported by the QR spec (error correction level H absorbs about 30% of the code without breaking it), but every brand mark you add reduces scan reliability in marginal light. If you must brand, keep the mark to 15% of the code area or less.

If you're sharing one code at multiple stops (rare but valid for short tours), size up to 3 inches and increase the quiet zone. If the code is behind glass, size up by 25% to compensate for parallax and glare.

Should you use dynamic or static QR codes?

Dynamic, almost always. A static QR code encodes the destination URL directly in the printed pattern — once you print it, the destination is permanent. A dynamic QR code encodes a short redirect URL that resolves through a service you control; you can change the destination at any time without touching the printed code.

For museums, dynamic is the right default for four reasons:

You will change the URL. Tour structure moves. Slugs change. The platform you're on may move from /tour/123 to /t/abc between versions. A static code freezes you to one URL forever; a dynamic code lets the underlying platform evolve without you reprinting wall cards.
You will make mistakes. Someone will print the QR for the wrong exhibit. A dynamic code is a five-minute fix in the admin; a static code is a reprint cycle.
You need scan analytics. Dynamic platforms report scan counts, times, and rough locations. Static codes report nothing — a scan goes straight to the destination and you have no record of it.
You may switch platforms. Most museums switch audio guide vendors at some point. A dynamic QR layer that you own — pointing to whichever platform is current — lets you swap underlying tools without reprinting signage.

The only argument for static is paranoia about a redirect service going offline. That risk is real but small if you use a major QR platform; it's smaller than the certainty of needing to update something.

A defensible setup: own your QR redirects on a short, branded domain (yourmuseum.org/q/abc) and point each one to the current tour URL on your audio guide platform. If you ever change platforms, the QR layer stays put and only the destination moves.

How do you solve the first-scan problem?

By telling visitors what they get before they scan the first code. The hardest QR scan to earn is the first one, because a visitor approaching an unfamiliar code doesn't yet know what's behind it. The second and third scans are easy if the first one paid off. The whole adoption curve hinges on the threshold.

Three things solve it, in order of effectiveness:

A docent or front-desk staffer mentions the tour at admission. A single sentence — "We have an audio guide on your phone; scan the QR in the lobby to start" — moves adoption more than any signage choice. Where there's no staffed admission, a lanyard-wearing docent at the gallery entrance does the same job.
A large lobby sign with a single landing QR. This is the one QR that has to be unmissable — a vertical sign at the entrance with a 6-inch QR, a one-line call to action ("Start the audio guide — works in 10 languages"), and a logo of the platform if recognizable. Visitors who scan here are pre-committed by the time they reach the first object-level code, and the audio guide page can teach them what subsequent codes will do.
A consistent visual treatment. Every QR in the building should look the same: same border, same icon, same call-to-action language. Visitors learn the visual once and recognize it everywhere. Mixed treatments (one code with a logo, another without, a third with different wording) make every code feel like a new decision.

A useful pattern we've seen: the first wall card in a gallery includes a longer call-to-action ("Scan to hear the curator on this piece — 90 seconds, no app needed"); the rest of the codes in that gallery use a short visual identifier ("Listen"). The first card teaches; the rest deliver.

What's the right call to action next to a QR code?

A single sentence that answers "what do I get and how long does it take?" Generic prompts like "Scan for more" earn nothing; specific prompts earn scans. The best museum QR captions we've seen all share the same shape: a verb, a payoff, and a time signature.

Examples that work:

"Hear the curator on this piece — 90 seconds, 10 languages."
"Listen to a conservator describe the restoration — 2 minutes."
"Ask a question about this object."

Examples that don't:

"More information." (No payoff.)
"Powered by [Platform]." (Branding, not invitation.)
"Scan QR code." (Imperative without reward.)

The wall label is the right surface for the payoff. The QR code itself should be silent — the code is the door; the label is the doorman.

How well do museum QR codes actually scan?

For phone-based audio guides on well-signed deployments, scan rates routinely hit 10–25% of visitors at the entry QR and 30–60% of visitors who started the tour scan additional object-level codes. This is roughly an order of magnitude higher than the 1.3% average native app pickup rate, and the gap is the entire reason the category moved to QR.

The numbers vary substantially by institution and signage quality. The variables that move the rate, in our experience:

Visibility of the entry QR. A 6-inch code on a lobby sign at admission gets scanned by a meaningful share of visitors; a 2-inch code on a small placard near the gift shop gets almost no one.
Staff mention at admission. A single docent or staffer mentioning the tour at the door reliably doubles or triples the entry-scan rate.
Language coverage signaled at the entry. Visitors traveling on a different first language are more likely to scan if the entry sign explicitly lists languages. We've seen entry-scan rates jump significantly when the language flags are added to the lobby sign.
Object-level signage consistency. Once a visitor starts the tour, subsequent object-level scans follow a power law: if the first three object scans are rewarding, the rest of the tour gets used; if the first one is broken or wrongly labeled, the visitor stops scanning.

Any vendor that quotes you a single scan-rate number across all deployments is selling, not reporting. Institutional variables — signage, staffing, gallery layout, visitor demographic — dominate the platform choice.

Where QR codes don't fit

A few honest cases where the QR-on-the-wall-card model is the wrong answer:

Outdoor heritage sites with weatherproofing constraints. Direct sun fades printed QR codes, and rain damages all but high-spec laminated stock. A dynamic QR setup helps because you can replace the printed code without losing the destination, but outdoor sites need a more robust hardware spec than a typical gallery.
Dim lighting with no surface contrast. Some galleries — particularly low-light installations and sound-and-light environments — make any printed signage hard to scan. The fix is usually a single, illuminated entry QR at the gallery threshold rather than per-object codes inside.
Visitors without smartphones. A small fleet of loaner phones at the front desk handles the gap. Smartphone ownership in typical museum-visiting demographics is high but not universal, and a no-questions-asked loaner program is the right accommodation.
Accessibility cases where QR isn't reachable. A visitor with low vision or a visitor in a wheelchair shouldn't have to find and scan a code that's been placed for the median adult. NFC tap-to-start is a complement (not a replacement) for QR in these cases, and many phone-based platforms support both.

QR is not the visitor experience. The audio guide is the visitor experience; QR is the doormat. Implementations that confuse the two tend to over-engineer the doormat.

What's the right next step?

If you're evaluating a phone-based audio guide platform, the QR layer is mostly a procurement and signage decision rather than a platform feature. Own your QR redirects on a short branded domain you control, point them at whichever tour URL the platform serves, and use the signage and call-to-action guidance above. That setup outlives any single vendor.

For more on the category around QR-based tours, see the pillar guide on AI audio guides and the deep comparison of AI audio guides versus traditional studio production. For the related decision on whether to keep any rented handsets at all, see BYOD versus rented handsets.

See the full visitor-experience hub.

FAQ

Yes — that's the entire point. Modern iOS and Android camera apps recognize QR codes natively and open the URL in the default browser. No app install, no account, no permissions. A QR-based audio guide is a web page the visitor reaches by pointing their camera. This is the structural reason QR cleared the bar that native museum apps could not.

A QR scan resolves a URL, which requires connectivity. In galleries with poor cellular coverage, free guest Wi-Fi is the standard solution and one that most museums already provide. Some platforms support offline mode where audio is cached after the first stop, which helps with intermittent coverage but doesn't replace connectivity for the first scan.

Fewer than you'd think. A 20-stop tour usually needs one lobby entry QR, one entry QR per major gallery, and an object-level QR only at stops with bespoke audio. Most museums over-print QR codes at first and dial back. The signal that you have too many: visitors stop scanning by the third one in a single room.

Partially. A QR code at the median adult height is harder to reach for a seated visitor and harder to see for a visitor with low vision. Best practice is to vary placement heights, offer NFC tap-to-start as a complement, and pair the QR signage with a verbal cue at admission. The audio guide itself, once started, inherits the phone's accessibility stack (VoiceOver, TalkBack, dynamic type), which is generally stronger than rented-handset hardware.

A static code encodes the destination URL directly into the printed pattern — once printed, the destination is permanent. A dynamic code encodes a short redirect that resolves through a service you control, so you can change the destination without reprinting. For museums, dynamic is the right default: it lets you fix mistakes, change platforms, and track scan analytics without touching signage.

Increasingly yes, with the right signage. Smartphone QR scanning is now mainstream behavior — over 80% of US smartphone users have scanned a code in the past year. In museum settings, the variables that determine scan rate are mostly about signage clarity, staff mention at admission, and a clear payoff in the call to action. A well-signed entry QR routinely earns scans from a meaningful share of visitors; a buried one earns almost none.

No. The right model is a single landing QR at the entry, a gallery-level QR at each major room, and object-level QRs only where there's bespoke audio worth stopping for. Visitors who see two hundred QR codes in a wing tune them out the same way they tune out wall text density. The QR signals "here's something specific"; over-deploying it kills that signal.

The verdict

QR codes won the museum audio guide threshold problem because they ask for nothing: no install, no account, no commitment. The hard work isn't on the QR side — it's on the signage, the placement, the call to action, and the choice between dynamic and static. Get those four right and a phone-based audio guide outperforms a native app by an order of magnitude on adoption alone, before you've made a single argument about content quality.

If you're past the threshold question and into platform evaluation, the visitor-experience pillar walks through BYOD, dwell time, and the rest of what changed in how visitors expect to engage in a gallery. If you're ready to look at numbers for your own institution, our pricing is published in full and the pilot tier is free.

About the author

Eric Duffy is the founder of Convo, a platform that lets museums and cultural institutions publish multilingual audio tours their visitors can have a conversation with. He writes about visitor experience and museum interpretation from inside the category, drawing on deployment data, RFP responses, and conversations with curators and visitor-experience leads. Reach him at eric@convo.app or on LinkedIn.