Should badges be encoded before or on site?
Neither, ideally. Pre encode identical blanks in advance and assign them at the desk. That removes both problems in one move: the accurate attendee list that nobody has three weeks out, and the encoding queue that nobody can afford on the morning of the event.
Encoding in advance
Cheap, calm and done in a room with good light and a working connection. Every badge is written and verified before anything is under pressure.
It requires an accurate list, which is what you never have. Registrations continue, names change, companies substitute people, and a proportion of badges are wrong before they are printed.
Wrong badges then have to be found in a box of hundreds, which is slower than it sounds at eight in the morning.
It also means a walk in has no badge at all, and walk ins are the attendees most likely to matter commercially.
Encoding on site
Flexible and slow. Every badge is correct because it is made when the person is standing there.
The cost is equipment, trained staff, a working connection, and a queue at the moment the venue is busiest.
It also concentrates every failure into the worst hour. A printer jam, a flat reader or a network drop at nine in the morning is an event level problem rather than an inconvenience.
For a small event it is entirely reasonable. For a large one it is where the day goes wrong.
The third way
Write every badge identically, in advance, with a chip that mirrors its own identifier. Each badge is then unique without being personalised.
At the desk, the badge is assigned to the person in the system: read the badge, attach it to the record, hand it over.
Assignment is a database write rather than an encoding operation, so it takes a second and needs no writing equipment at the desk.
Badges are interchangeable until assigned, which means no sorting, no hunting through boxes, and a walk in served from the same stack as everyone else.
Making the third way work
Buy chips that mirror. It is a purchase decision that cannot be added later, and it is the whole basis of this approach.
Verify the batch before it leaves your office. Every badge read once, counted, and any dead ones removed from the stack rather than discovered at the desk.
Make assignment work offline, with reconciliation afterwards. The desk must not stop because the connection did.
And keep the printed face generic or printed on site separately, since a badge that is interchangeable electronically and personalised visually is not interchangeable at all.
Frequently asked questions
Is pre encoding cheaper?
Per badge yes, and the saving disappears if a proportion of badges are wrong and have to be found and replaced on the day.
What about walk ins?
They are the strongest argument for assigning at the desk. A stack of interchangeable badges serves a walk in exactly like anyone else.
Do all chips support mirroring?
No. It varies by reference and cannot be added afterwards, so it belongs in the purchase order.
Can the desk work offline?
It has to. Assignment should queue locally and reconcile when the connection returns, or a network drop closes registration.