QR Code Forms: Print One Code, Collect Voice Answers at Events With No Sign-In

The biggest problem with a QR code survey at an event is rarely the code. It is the sign-in screen that appears after the scan. A printed code only has to clear one technical bar, a blank margin 4 modules wide, while a login prompt asks a stranger at your stall for an account they did not plan to use.

Print one code. Take the login away. Get the voice answers. This post is the checklist we would hand anyone running a stall. It covers a QR code form people can answer in under a minute, a print every phone can read, and what you give up when answers arrive without a name.

Why a sign-in wall empties a walk-up form

Think about the 30 seconds after someone scans your code. They are standing at a stall, holding a coffee, half listening to a conversation. Every extra screen between the scan and the first question is a reason to put the phone away.

A login prompt is the worst of those screens. It asks for an account, a password or a permission dialog, and it signals that the form wants something before it gives anything. We would not put a sign-in wall in front of any form people meet by walking past it.

Who scans a code at an event

Walk-up respondents are different from people who click a link in your newsletter. They did not choose your form in advance. They have no relationship with your brand yet, and they will give you about as much time as it takes to read one question.

That changes the design brief. The form has to start answering immediately, ask for one thing at a time, and finish before the person's attention moves to the next stall. A one-question-at-a-time layout helps, because nobody sees a long scroll of fields on the first screen.

What anonymous mode changes in Would You Rather

By default, a Would You Rather form asks respondents to sign in with Google before they answer. That is right for hiring screens and research panels, where you want to know who answered and stop repeat submissions. For a stall, it is the wrong default.

Each form has a switch in its settings called "Answer without signing in". When it is on, anyone with the link or the QR code goes straight to the first question. Voice recordings and file uploads work for those respondents too. The settings panel says the trade-off plainly: responses arrive without a respondent identity, so they can't be de-duplicated and respondent emails aren't captured.

The sign-in column wins on data quality, and that is not a small win. If every answer needs to be tied to a real person, keep sign-in on and accept the lower walk-up rate.

How to set up a QR code form in 5 steps

The process is short, and most of it happens before the event. Plan for about an hour including a test print, and do the test print at least a day ahead.

Step 2 is the one people forget. Test the printed code while signed out, or in a private browser window, because a signed-in creator never sees the sign-in wall their visitors would hit.

The QR code printing checklist

A QR code is a grid of dark and light squares called modules. The standard behind it, ISO/IEC 18004, fixes how those modules encode data. Everything that goes wrong at an event comes from the physical print: margins cropped, contrast too low, or modules too small for a phone to resolve from where people stand.

Keep the quiet zone clear

DENSO WAVE, the company that created the QR code, is blunt about the margin. A QR code requires a four-module wide margin at all sides of the symbol, and nothing may be printed in itDENSO WAVE, qrcode.com, 'Information capacity and versions' how-to page, accessed 8 October 2026.. The margin is how a phone finds where the code starts and stops.

The usual mistake is a designer cropping the white border to fit a poster layout, or placing the code against a busy photo. A margin measured in millimetres is also a trap. The quiet zone scales with the module size, so a code printed twice as large needs a margin twice as wide. The PNG that Would You Rather downloads already includes the 4-module margin, so print the whole image and do not trim it.

Choose error correction level M

Every QR code carries spare data so it can still be read when part of it is damaged. DENSO WAVE lists four levels: L restores about 7% of the code, M about 15%, Q about 25% and H about 30%. On its own site, DENSO WAVE says level M is the one most frequently selected, and suggests Q or H for dirty factory environmentsDENSO WAVE, qrcode.com, 'Error correction feature' page, accessed 8 October 2026. Levels: L about 7%, M about 15%, Q about 25%, H about 30%..

More correction is not free. The spare data takes up space, so the same link needs more modules. GS1, the standards body behind retail barcodes, shows the cost in a worked example.

We think level M is the right default for indoor events. Choose H only when you plan to cover part of the code with a logo or expect it to be scuffed, rained on or half hidden behind a table edge. The Would You Rather download uses level M.

Print dark on light, and skip red

GS1 published a colour and quality guide for 2D barcodes in February 2025. Its advice is simple. Black and white give the same contrast under any light source and are ideal; colour gradients and similar shades should be avoided as they offer poor contrastGS1 2D Barcode Colour and Quality Guide, Release 1.0, February 2025, section 3.. The guide also warns against red modules, because red-light scanners read red as white.

Phone cameras use white light, so a red code may still work on a phone. That is not a reason to risk it. Brand colours belong around the code, on the poster, not inside it. If you must use colour, keep the modules very dark and the background very light, and test it.

Size the code for the distance, then test it

Here is the uncomfortable truth about QR size: the standard does not tell you how big to print for a given distance. ISO/IEC 18004 and DENSO WAVE's pages define modules and margins, not scan distances. Phones differ in camera resolution, focus and lighting, so the only reliable answer is a test print.

QR vendors often quote a rule of thumb that the code should be about one tenth as wide as the scan distance. We could not find that ratio in DENSO WAVE's material or in GS1's guidance, so treat it as a starting guess, not a rule. GS1's retail size tables are written for checkout scanners held close to a pack, not for a phone held across a table.

What you can control is the number of modules. A code with fewer modules has bigger ones at the same print size, and bigger modules read from further away.

Short links make easier codes

A QR code grows in steps called versions. Version 1 is 21 by 21 modules, and each version adds 4 modules per side. How many characters fit in each version depends on the error correction level. For a web link at level M, DENSO WAVE's capacity table gives these limits.

A Would You Rather form link has the shape wyr.humanbit.in/f/ followed by a 10-character code. With the https prefix, that is 36 characters, which fits version 3. By our calculation from the capacity table, the printed image is 37 modules wide including margins.

Add tracking parameters such as ?utm_source=event&utm_medium=qr and the link passes 62 characters. It now needs version 5, 45 modules wide. At the same print size, each module is about 18% smaller. That is the hidden cost of long links in print.

What the downloaded PNG gives you

The QR tab on the Would You Rather share page shows the code on screen and offers a PNG download. The file is 1,024 pixels square, at level M, with the 4-module margin included. At 300 dots per inch, 1,024 pixels prints about 8.7 cm wide, which suits a table card or a flyer.

For a larger poster, the image is scaled up. QR modules are plain squares, so moderate scaling usually prints cleanly, but check the test print for blurred edges. One honest limit: the QR tab encodes the plain form link. Tracked links exist on a separate tab, but the share page does not draw a QR code for them. If you want to compare scans from two posters, you would need to make the second code elsewhere.

The trade-off: anonymous answers can't be de-duplicated

Removing sign-in removes identity. That is the entire point, and also the cost. Without an account, a form cannot tell whether two submissions came from the same person.

What the research says about repeat and bogus entries

The best evidence on this comes from online polling, not events. In 2020, Pew Research Center studied more than 60,000 interviews across six online sources. Bogus respondents made up about 4% to 7% of opt-in samples, against 1% in panels recruited by addressPew Research Center, 'Assessing the Risks to Online Polls From Bogus Respondents', February 2020. More than 60,000 interviews, at least 10,000 from each of six online sources.. Taking the survey more than once was one of Pew's criteria for a bogus respondent.

Pew's opt-in samples were paid panels, where people have a cash reason to answer twice. A visitor at your stall has no such reason, so we would expect fewer repeats. We have not measured this on our own product, so we cannot give you a number. The point is directional: open forms let some junk in, and you should plan for it.

What happens in Would You Rather

On an anonymous form, every submission lands as a new response. The "Already applied" screen, which normally stops a second submission, never shows, because there is no respondent to match. If a stale draft comes back from an old browser tab, it is saved as a fresh response rather than overwriting an earlier one.

Respondents can still resume a half-finished form, but only on the same phone and browser, because the draft is kept on that device. There is no CAPTCHA or per-device limit on anonymous forms today. If someone wants to submit 20 times, nothing stops them.

How to live with it

For most event feedback, we think the trade is worth making. A few repeat answers matter much less than the visitors you would lose to a login page. Three habits keep the data usable.

First, ask for an optional email or phone number, and use it to spot duplicates after the event. Second, check submission timestamps; a burst of near-identical answers in one minute is easy to see. Third, pause the form when the event ends, so a photo of your poster on social media does not keep collecting answers for weeks.

Where voice answers fit at a stall

A QR code gets someone to your form. A good question keeps them there. At an event, people are standing, holding things and talking, which is close to the ideal setting for a spoken answer and a poor one for thumb typing.

Our view is that one open voice question is the right number for a stall form. "What would make you come back next year?" earns a 30-second recording. Ask taps and ratings first, then the voice question, and set a time limit in the question text. We explain why spoken answers run longer, and why they also lose more people, in our analysis of voice surveys and drop-off.

Noise is the practical worry. A busy hall can drown a recording. You can switch on the optional "or type your answer" box on a voice question, so people in a loud corner type instead. It is off by default, and at events we would turn it on. For more on running forms outside an office, see our guide to forms for field research.

Results arrive in the dashboard as they come in, with each recording next to the person's other answers. If your team wants them in a spreadsheet during the event, sending form responses to Google Sheets covers the live sync, including audio links.

Where QR code forms are the wrong tool

QR forms are good at one thing: turning a passer-by into a respondent quickly. They are weak at several others, and the evidence for some of our advice is thinner than we would like.

First, phones are a harder place to answer surveys. A 2023 Cochrane review of randomised trials found that the odds of response were about 1.6 times higher when an e-questionnaire was completed on a computer rather than a smartphoneEdwards et al., Cochrane Database of Systematic Reviews, 2023, Issue 11, MR000008.pub5. Pooled randomised trials of electronic questionnaires; odds ratio 1.62, 95% CI 1.36 to 1.94.. A QR code commits you to the phone. Keep the form short to compensate.

Second, the same review found response rose when people could choose their mode, electronic or paper, rather than being given electronic only. At events with older visitors or poor signal, a stack of paper cards beside the code is a fair backup.

Third, our printing advice on size and distance rests on a vendor rule of thumb and on testing, not on a published study. We looked for a controlled study of phone scan distance and did not find one. Fourth, anonymous mode trades away data quality. If you need one response per person, an audit trail, or follow-up with every respondent, keep sign-in on.

Before the doors open

Do three things the day before. Print the code at its final size, scan it signed out from where visitors will stand, and submit one test answer by voice so you hear how the hall sounds in a recording. Then delete the test response.

On the day, put the code at chest height, away from glare, with one line of text saying what people get and how long it takes. "2 questions, 1 minute, speak or type" sets expectations better than a logo. If you want the poster and the form to look like they belong together, we cover logos, colours and themes in our post on branded forms. After the event, pause the form, export the answers, and listen to the recordings while the day is still fresh.

References

1. DENSO WAVE, Error correction feature, qrcode.com, accessed 8 October 2026. Used for the four error correction levels and the note that level M is most frequently selected.
2. DENSO WAVE, Information capacity and versions of QR Code, qrcode.com, accessed 8 October 2026. Used for the 4-module quiet zone.
3. DENSO WAVE, Version list and data capacity, qrcode.com, accessed 8 October 2026. Used for module counts per version and binary capacity at level M.
4. DENSO WAVE, History of QR Code, qrcode.com, accessed 8 October 2026. Used for the 1994 introduction date.
5. GS1, 2D Barcodes at Retail Point-of-Sale Implementation Guideline, Release 1.1.0, December 2025. Used for the error correction width example.
6. GS1, 2D Barcode Colour and Quality Guide, Release 1.0, February 2025. Used for contrast and colour advice.
7. Pew Research Center, Assessing the Risks to Online Polls From Bogus Respondents, February 2020. Used for bogus respondent rates and the repeat-entry criterion.
8. Edwards, P. J., Roberts, I., Clarke, M. J., DiGuiseppi, C., Woolf, B. and Perkins, C., Methods to increase response to postal and electronic questionnaires, Cochrane Database of Systematic Reviews, 2023. Used for the computer vs smartphone and mode choice findings.
9. Would You Rather source code, share page and form settings, reviewed 8 October 2026. Used for anonymous mode, the QR download format and pause behaviour.

The weakest point in this source base: nothing here measures QR scan distance or repeat submissions at real events. The size advice relies on testing, and the duplicate risk is inferred from paid online panels.