Does "scan me" text increase scans: the honest answer

What the evidence actually is

Search for a number and you will find plenty: paid QR platforms citing lifts of 30–80% from frames and CTA text. None that we can find publish their methodology, sample sizes, or controls, the figures come from marketing pages, not studies, and are not verifiable. There is no peer-reviewed or independently replicated dataset on QR label lift that this page could honestly cite.

So the honest answer has two halves: the direction of the effect is well-supported by ordinary reasoning about behaviour, and the size of the effect is unknown and almost certainly varies enormously with context.

Why the mechanism is plausible

A bare QR code asks a passer-by to spend effort on an unknown outcome. A label that names the destination ("Scan for the menu", "Scan to pay") changes the transaction:

• It answers "what happens?" before the reader commits, removing the main reason to walk past.
• It filters honestly. People who do not want the menu will not scan, which is fine: scans were never the goal, outcomes were. See what scan rate to expect for base-rate honesty.
• It addresses caution. Awareness of QR phishing is growing; a specific stated destination reads as more trustworthy than an unlabelled square, though a label is obviously not a security guarantee.

Note what the reasoning supports: naming the destination. It says nothing for the literal text "scan me", which restates the mechanism while leaving the outcome unknown, the distinction drawn in frames and CTA labels.

Run the test yourself

Context beats any published average, and the test is cheap:

1. Two codes, one variable. Same artwork, same size, same destination page; only the label differs (or label vs none).
2. Distinct tracking. Give each code the same URL with different UTM parameters (utm_content=label-a vs utm_content=label-b) following adding UTM parameters. UseQR generates static codes and does no tracking of its own, so the measurement happens in your analytics; the tracking guide covers the setup.
3. Control for position. Placement effects swamp label effects, so alternate the two codes between locations weekly, or print both versions of the same poster and distribute them randomly.
4. Wait for volume. With typical scan counts, differences under a few dozen scans per variant are noise. Set a threshold before starting (for example, run until the combined total passes 200 scans), rather than stopping when the result looks pleasing.

Both variants should pass the validator before printing; a marginal code that scans unreliably will corrupt the comparison in whichever position has worse light.

FAQ

Does adding "scan me" to a QR code increase scans?

Probably a little, possibly a lot, and nobody has published rigorous evidence either way. The reasoning favours labels that name the destination over the literal words "scan me". If the number matters to you, run a two-code UTM test on your own audience.

Why are vendor statistics about QR frames unreliable?

They come from marketing pages without published methodology, sample sizes or controls, and they cannot be independently checked. That does not make them false. It makes them unusable as evidence, especially for your specific context.

What is the best text to put next to a QR code?

The destination and the benefit, in a few words: "Scan for the menu", "Scan to pay by UPI", "Get the free guide". Adding a visible short URL beneath gives people who cannot or will not scan a second path to the same place.

How many scans do I need before an A/B result means anything?

More than intuition suggests. Differences of a handful of scans are noise; as a rough floor, wait until the two variants together have a couple of hundred scans, and decide the stopping rule before you start rather than when a result looks good.