Events are where QR codes get used most and where they fail most, and the two facts have the same cause: a show floor breaks nearly every assumption a code quietly depends on. There is a crowd between the person and the sign. There is a network carrying thousands of phones that was specified for a few hundred. The person is walking, holding a bag and a coffee, and they have eleven more stands to get through. Nothing about that is like scanning a table tent in a quiet café.

The good news is that the failures are predictable, which means they are avoidable at planning time rather than discoverable on the day.

Plan as though there is no signal, because there frequently is not

This is the defining constraint and it is the one most often discovered at 9am on day one. Large venues concentrate an enormous number of devices into one space, all contending for the same spectrum and the same backhaul. Conference centres are also often built of exactly the materials that block radio — concrete, steel, glass with metallic coatings — and the halls are frequently below ground.

The result is not a clean failure. The code resolves instantly, because decoding happens on the phone and needs no network at all, and then the page sits there loading. From the visitor's side that is indistinguishable from a broken code, and it costs you twice: once for this scan, and once because it makes them warier of the next one.

What to actually do about it

Put the essential information in print as well. If the only place your stand number, product name or contact details exist is behind a code, a bad network has erased them.
Prefer payloads that need no network. A contact-card code saves your details straight into the phone with no page load at all. For a stand whose real goal is "let them leave with our details", that is strictly better than a link to a contact page.
Keep the destination light. A page that is mostly text loads on a bad connection. A page that opens with a video carousel does not.
Go and stand there. If you can get into the hall during build-up, scan your own codes on the venue network, from where a visitor will actually be standing. This is fifteen minutes and it is the only test that counts.

Size the code for the distance it is read from

Booth graphics are read from across an aisle, not from arm's length, and a code that works on a leaflet is invisible on a banner. The working rule is that a code wants to be roughly one tenth of the scanning distance across: read from four metres, it needs to be around forty centimetres. Codes on pull-up banners are routinely printed at a size that would be fine on a flyer and is useless at aisle distance.

Height matters as much as size. A code at the bottom of a pull-up banner sits near knee level and is blocked by every person standing in front of it — which, at a busy stand, is constantly. Put codes between waist and eye level, and put them where someone can stop in front of them without stepping into the aisle or into your own conversation. The full set of format constraints — sizing, contrast, quiet zone, error correction — is on the design studio page, and QR code design that still scans covers what breaks a code visually.

One code, one job — and a stand has several jobs

Stands accumulate codes. A code on the banner, one on the table, one on the leaflet, one on the demo screen, one on the staff badges. That is fine, and it is different from putting five codes on one surface, which is where it goes wrong.

Each code needs its own line of copy naming its destination — "Scan for the spec sheet", not "Scan me" — and each surface should carry one. Five competing offers in one glance is a decision, and a person mid-aisle will not make it. The reasoning is set out in what to put next to a QR code, and it applies with more force at an event because the person is moving.

The handout outlives the show, and this is the real argument for dynamic

Almost everything printed for an event is scanned after the event. Leaflets go into a bag, the bag goes into an office, and someone empties it a fortnight later. Badges, lanyard inserts, catalogue adverts and branded giveaways can surface months on. So the destination that made sense on the day — a stand-specific offer, a "book a slot this afternoon" page, a live schedule — is precisely wrong by the time most people arrive at it.

A static code cannot be told this. Whatever URL was encoded is encoded permanently, and once the page behind it retires, every leaflet still in circulation points at an error. A dynamic code lets the same printed artwork mean something different afterwards: the show offer becomes the product page, the live schedule becomes the recordings, the "come to stand 42" becomes "here is what was on stand 42". Nothing gets reprinted. The trade-off — that you have taken on a dependency on a redirect continuing to resolve — is set out in static vs dynamic QR codes, and event print is close to the strongest case for accepting it, because the artwork's useful life is so much longer than the campaign's.

Decide the after-the-show destination while you are designing the leaflet. It costs nothing then and it is a scramble later.

Organiser badge scanners are a different thing from your code

These get muddled, and they answer different questions.

The organiser's badge scan is the visitor identifying themselves to you. They have consented to a data exchange, and you get whatever the registration form collected. It works when someone stops and engages.
Your own code is the opposite direction: you identifying yourself to them, or handing them something. It works on people who do not stop — passers-by, someone reading your banner from the aisle, someone going through the bag later.

You want both, and they should not point at the same thing. The badge scan already has their details; your code should give them something. Using your own code as a lead-capture form is asking a person standing in a crowd to type, which is the friction described in where should a QR code point — and the organiser's scanner already does that job with no typing at all.

What the numbers will and will not tell you

Scan counts are genuinely useful at an event for one comparison: which of two placements is working. Same code, same offer, banner versus table — that difference is real and it is actionable next time.

What they cannot tell you is why. A code with few scans might be badly sized, badly placed, in a dead spot, or simply in front of fewer people than the other one — and a number that cannot separate four explanations is not yet an answer. They also tell you nothing about whether a scan became anything: that lives in your own systems, not in the code. How to track QR code scans is explicit about the boundary. Be equally explicit internally, because "we got N scans" has a way of becoming "we got N leads" between the show floor and the debrief.

The short version

Assume the network is unusable. Print the essentials, keep destinations light, prefer offline payloads like contact cards.
Size for the reading distance — roughly a tenth of it — and keep codes between waist and eye level, not at the foot of a banner.
One code per surface, each with a line naming where it goes.
Use dynamic codes on anything printed, because the handout will be scanned long after the show ends.
Decide the post-event destination before you print, not after.
Badge scanners capture leads; your code hands things out. Do not make yours a typing exercise.
Scans compare placements. They are not leads.