QR codes on screen: how long, how big, and why most TV and video codes fail

In the 2022 Super Bowl, Coinbase ran a single QR code drifting across a black screen for a full minute. It drove more than 20 million landing-page hits in about sixty seconds and briefly crashed the app. The lesson agencies took from it was often the wrong one. The code did not work because it moved; it worked because it was huge, alone on an empty screen, and held for a very long time. Most on-screen QR codes do the opposite: tiny, cluttered, and gone in a few seconds. Then everyone wonders why the scan numbers are flat.

A QR code on a screen is a different discipline from a QR code in print, and almost everything that makes a printed code work has a screen-specific twist. This guide is for agency teams putting codes into TV adverts, connected-TV (CTV) campaigns, social video and live presentations. We will cover the four things that decide whether an on-screen code scans, the broadcast traps that are invisible in print, the awkward truth about codes in mobile social video, and why on-screen campaigns need a dynamic code more than any printed one does.

On-screen is not print: four factors decide everything

A printed code sits still in someone's hand under steady light for as long as they like. A code on a screen has none of those luxuries. It appears for a few seconds, at a distance, possibly over moving video, and the screen showing it is usually not the screen that will scan it. Four factors carry almost all of the outcome.

What the diagram shows:

1. Dwell time. Keep the code on screen long enough for a viewer to notice it, reach for a phone, open the camera and scan: at least 5 to 7 seconds, and ideally 10 or more.
2. Size and safe area. Make it large, around 15 to 20% of the screen for TV, and keep it, with its quiet zone, inside the title-safe margin so a television's overscan cannot clip the edge.
3. Hold the frame still. A scanner needs a stable, in-focus code, so place it on a held shot or an end card, never over moving footage.
4. A second device scans it. The screen showing the code is not the one that reads it, so design for a viewer with a phone in their hand, and always give them a fallback.

The rest of this guide is those four factors in detail, plus the broadcast and platform realities around them.

Dwell time: the factor that makes or breaks it

If you change one thing about your on-screen codes, change how long they stay up. Scanning is a four-step chore: notice the code, decide to act, find and raise the phone, point and let the camera lock. That takes real seconds, and viewers are slower than the creative team imagines.

The widely repeated practitioner minimum is 5 to 7 seconds, with 10 seconds or more treated as comfortable. For lean-back television, where the viewer is across the room, some go further to 10 to 15 seconds. Be honest with clients about what these numbers are: rules of thumb, not a published standard. No broadcast or advertising-standards body mandates a QR dwell time. They are sensible defaults drawn from how long the physical task takes, and you should treat a real scan test, not a number, as the final word.

One implementation detail is telling. In Google's own connected-TV QR format, the code is only revealed after the first 5 seconds of a non-skippable ad, then held. The platform that arguably has the most scan data does not even surface a code in the opening seconds. The practical reading: give the code the back half of a 30-second spot, hold it through to the end, and do not expect a brief flash to convert.

This is also why a 6-second YouTube bumper is the wrong unit for a QR call to action. Bumpers are non-skippable and built for branding, not response. Subtract the time a viewer needs just to notice the code and there is no scan window left. If response is the goal, you need a format that can hold the code for long enough, and you need to pair it with something for the people who miss the window.

That something is a spoken or memorable short URL. A held broadcast moment happens once and cannot be paused, so "scan or visit example.com/offer" rescues the viewer whose camera would not focus, who was across the room, or who simply looked up too late. Every on-screen code should travel with a readable fallback.

How big, and how far: the 10:1 rule on a screen

Size on screen is really a question about distance. The reliable engineering principle is the 10:1 distance-to-size rule: a code's width should be roughly one tenth of the scanning distance. A code scanned from one metre wants to be about 10 centimetres across; from three metres, closer to 30. Television is a lean-back medium watched from perhaps 2 to 3 metres, so the on-screen code has to be genuinely large to clear that distance.

In screen terms, aim for the code to occupy roughly 15 to 20% of the screen area for a TV spot, and in pixel terms treat 400 to 600 pixels square as a floor for large displays. A small code tucked in a corner is decoration, not a call to action. There is a hard limit underneath all of this: QR's inventor, Denso Wave, specifies that each module (the small squares that make up the code) needs to render at four or more pixels to scan dependably, so once a code is scaled down past that point on a distant screen it simply stops working. The fundamentals of QR code size and the specs that make it scan still apply; screens just push the size requirement up.

Social video flips the distance maths. There the phone is in the hand, around 30 centimetres away, so size is rarely the binding constraint. A different problem takes over, and it is the one almost nobody is honest about.

Hold the frame still

A QR scanner needs a stable, in-focus frame to lock the three finder patterns in the corners and read the timing grid. Pan, zoom, drift or animate the code and you defeat the scanner and add motion blur on top. The fix is simple: treat the code as a static element on a held shot or a dedicated end card, not as something laid over moving footage.

This is where the Coinbase example is so often misread. That code did move, slowly, like a bouncing DVD logo, and it worked. But it moved across an otherwise empty black screen, it was enormous, and it held for a full sixty seconds with nothing else competing for attention. The takeaway is "slow, uncluttered and long," not "motion is fine." For a normal 15 or 30-second spot with a busy frame, give the code a still, high-contrast moment of its own.

The single-device problem in social video

Here is the uncomfortable truth for anyone planning a QR code inside an Instagram Reel, a Story or a TikTok: if the viewer is watching on their phone, they generally cannot scan an on-screen code with that same phone. The camera cannot point at its own screen. The documented workarounds are clumsy, such as screenshotting the frame and scanning the saved image from the gallery, and almost nobody does that mid-scroll.

So in mobile-first social, a QR code is usually the wrong tool. Use the platform's native link affordance instead: Instagram and Facebook Story link stickers, TikTok's link stickers and bio links. Those put a tappable destination one thumb-press away, with no second device required. Reserve QR codes for the cases where the content is genuinely consumed on a shared or second screen: a Reel playing on a TV in a shop, a video on a desktop, a slide projected in a room, or content people screenshot to revisit later. If you would not scan it yourself while holding the phone that is playing it, your audience will not either.

Broadcast gotchas that never show up in print

On-screen codes fail for reasons that simply do not exist on paper.

• Title-safe and overscan. Televisions still overscan, cropping the outer edge of the frame, and many do it by default. Broadcast convention, derived from SMPTE practice, keeps important content inside a title-safe area of about 90% of the frame (action-safe is around 93%). Put the code, and its quiet zone, well inside title-safe or a real set may clip the edge and break the scan.
• Compression and motion blur. Video compression introduces banding and softens edges, and any movement smears the modules. A code that is crisp in your export can degrade on a streaming feed, so keep it large and static to give the modules margin.
• The quiet zone still matters. The code needs a clear margin of at least four modules on all sides. Do not let busy video imagery bleed into that border; sit the code on a solid, light, contrasting plate.
• Glare and brightness. On a physical panel, maximum brightness and screen glare wash out the contrast a scanner relies on. Codes read best with strong dark-on-light contrast at a sensible panel brightness, not a blown-out white. Many on-screen scan failures are really contrast and focus problems in disguise.

TV is actually the best screen for a QR code

For all those constraints, television is the surface where on-screen QR makes the most sense, precisely because the scanning device is separate from the viewing one. The second screen is already in people's hands. Around 83% of viewers use a second device while watching TV, and roughly two thirds of social users will watch TV or streaming while on another screen. A meaningful share already act on it: about 31% of viewers have used a QR code or on-screen link to shop for something related to what they were watching.

Connected TV sharpens this further. CTV viewers tend to sit a little closer, are more likely to have a phone to hand, and the format can support interactive or system-generated QR overlays with cleaner attribution. Linear broadcast is the harder mode: one shot, no overlay, no interactivity, no pause. On linear especially, dwell time, size and a spoken fallback URL are not nice-to-haves, they are the campaign. This connects to the wider work of getting more scans through the offer and the call to action: on screen, the call to action has to do its job in seconds.

Why on-screen codes need to be dynamic

On-screen campaigns benefit from a dynamic QR code more than printed ones do, for two reasons that are specific to broadcast and video.

First, you can never re-encode a shipped asset. A printed poster is fixed once it is on the wall, but an aired or uploaded video is just as locked, and far more expensive to recut. A dynamic code keeps the on-screen pattern fixed while letting you change the destination after the ad ships: fix a broken link, point it at a live offer during the flight, or redirect it to an evergreen page when the campaign ends. For a creative asset you cannot edit, that flexibility is the difference between a recoverable mistake and a wasted media buy.

Second, attribution. Broadcast and video have no native click. A distinct dynamic code per spot, network, daypart or platform is the only clean way to know which placement drove the scans, which feeds straight into the scan metrics that actually matter. Without per-placement codes you are guessing; with them, on-screen QR becomes one of the few measurable bridges from a TV moment to a tracked visit.

A pre-flight checklist for on-screen QR

Before any code goes into a screen asset, run through this:

1. On screen for at least 5 to 7 seconds, ideally 10 or more, held through to the end of the spot.
2. Large: around 15 to 20% of the screen for TV, never a corner afterthought.
3. Static, on a held shot or end card, not over moving footage.
4. Inside the title-safe area (about 90% of the frame), with its quiet zone clear.
5. Strong dark-on-light contrast on a solid plate, at sensible screen brightness.
6. A spoken or short fallback URL alongside it for the viewers who miss the scan.
7. A dynamic, per-placement code so the destination stays editable and scans are attributable.
8. Tested on a real TV and a real phone from couch distance, not just on the edit suite monitor.

For mobile-first social video, replace the QR with a native link sticker and keep the code for shared-screen contexts.

Frequently asked questions

How long should a QR code be on screen?

At least 5 to 7 seconds, and ideally 10 or more, with 10 to 15 seconds recommended for lean-back television. These are practitioner rules of thumb, not a broadcast standard; the real test is whether someone can actually notice, raise a phone and scan in the time given. A 6-second bumper is too short to convert.

How big should a QR code be on a TV screen?

Aim for roughly 15 to 20% of the screen area, and apply the 10:1 distance-to-size rule for couch distance: the code's width should be about a tenth of the viewing distance. In pixels, treat 400 to 600 square as a floor for large screens. A small corner code will not scan from across a room.

Can you scan a QR code on the same phone you're watching on?

No, not directly. A phone's camera cannot point at its own screen, so a code in a Reel or TikTok you are watching on that phone is effectively unscannable without screenshotting it first. On-screen QR works when a second device is involved, most naturally a phone scanning a TV. On mobile social, use a link sticker instead.

Do QR codes work on Instagram Stories and TikTok?

Rarely, for in-feed mobile viewing, because of the single-device problem above. The native answer on those platforms is a link sticker or bio link, which is one tap away and needs no second device. Keep QR codes for content that plays on a shared or second screen, or that people screenshot to use later.

Why won't my QR code scan on TV?

The usual causes, in order: it is too small, it is on screen too briefly, it is moving, it has weak contrast over busy video, the screen has glare or is too bright, or it sits outside the title-safe area and is being clipped by overscan. Work down that list and most on-screen failures resolve.

Can a QR code be in a moving video?

It can be present, but it will not scan reliably while it moves. Cameras need a stable, in-focus frame to read the code, and motion adds blur. Hold it still on a dedicated shot or end card. The Coinbase Super Bowl code drifted slowly, but it was huge, alone on a black screen, and held for a full minute, which is not a normal spot.

Should I put a QR code in a 6-second YouTube ad?

No. Six-second bumpers are non-skippable branding formats with no room for the notice-and-scan sequence. If you want a scannable code, use a longer format that can hold it, or rely on YouTube's system-generated CTV QR overlays rather than burning a code into the video file.

The short version

A QR code on a screen lives or dies by four things: dwell time, size, stillness and the second device. Hold it on screen for at least 5 to 7 seconds and ideally 10 or more, make it 15 to 20% of the screen so it clears couch distance, keep it static on a held shot inside the title-safe area, and remember that the screen showing the code is not the one that scans it. Pair every on-screen code with a spoken fallback URL, because a broadcast moment happens once.

Television is the best surface for a QR code precisely because the phone is already in the viewer's other hand, and connected TV makes the attribution cleaner still. On mobile-first social, drop the QR and use a link sticker. Above all, make on-screen codes dynamic: you can never recut a shipped asset, and a per-placement code is the only honest way to measure which screen earned the scan. Build your next on-screen code from a dynamic, trackable link, run the eight-point checklist, and test it on a real TV from the sofa before it airs.