How to compare one hotel key card supplier against another so that an incompatible batch is caught before production, not after freight.

Three quotes arrive for 10,000 hotel key cards, within a few cents of each other. It looks like a rounding exercise, so the decision gets made on the lowest line.
Eleven weeks later the pallet lands, the drawer behind reception is nearly empty, and the night auditor finds that the new cards read on room 112 and nowhere else. The supplier was asked whether the cards were compatible and said yes. The word covered the frequency and the card body, not the chip.
That failure mode has little to do with price per card. This guide is about comparing one hotel key card supplier against another in a way that catches it before production, not after freight.
Price per card is one line in a landed cost, and rarely the biggest. Check you are comparing like with like.
Divide the all in figure by the number of cards that actually open doors. That is the only unit price worth comparing.
Most compatibility failures start with a quote request that did not contain enough to answer properly, and a supplier who answered anyway. Five items separate a specification from a guess.
Not the brand alone. The same manufacturer ships several lock families and they do not all read the same credential. The model number is usually on a label under the battery cover. Photograph it.
Lock software gets upgraded and credentials change with it. A hotel refurbished in two phases can run two generations of the same brand on different floors. Send the version if you know it.
One card that opens doors now, posted to the supplier. Nothing else resolves as much ambiguity: it lets a hotel key card manufacturer read the chip type, memory layout, UID length and antenna geometry directly instead of inferring them.
Rooms, keys issued per arrival, and whether cards are collected at checkout and reused. A supplier who knows you reuse cards sizes the order differently.
Blank for your own encoder, or pre encoded at the factory to a scheme you supply. Different products, different checks. Decide before the quote.
When a supplier writes back the word compatible, ask what it refers to. It can mean three things.
Physically compatible. The card is the right size and fits the slot or reader face. Standard geometry is ID-1 in ISO/IEC 7810 and nearly every key card meets it, so it tells you almost nothing.
Radio compatible. The card answers on the same frequency and air interface, usually 13.56 MHz under the ISO/IEC 14443 family. Also insufficient. MIFARE Classic, Ultralight, Plus and DESFire all run at 13.56 MHz and all speak ISO/IEC 14443, and they are not interchangeable. Memory structures and security models differ, and a lock expecting one rejects another.
Chip confirmed. The supplier names the exact chip in writing, on the order confirmation, before production. MIFARE Classic 1K is a confirmation. 13.56 MHz RFID is not. The same applies at 125 kHz, where EM4200 and T5577 behave differently.
Treat the written chip designation as part of what you are buying. If a supplier will put it on the order confirmation, somebody has engineered the answer. If the answer stays at works with most hotel locks, nobody checked.
Insist even when it feels pedantic. It gives you something to hold the supplier to if the batch fails, and forces somebody at the factory to look at your sample. The chip families are compared in the hotel key card compatibility guide.
On MIFARE Classic installations the card carries data in sectors protected by keys the lock system writes and reads. On DESFire installations it holds applications protected with AES, with keys diversified per property or per card. Either way the blank card is half the credential. The other half is the data written onto it with your keys.
So the question for a hotel key card supplier is not can you encode our cards. It is: given this lock model and software generation, must the cards arrive blank, and if so, what has to be true of the blank card for our encoder to accept it. A supplier who answers that has seen your system. One who offers to encode everything without asking who holds the keys has not.
Size the reserve from attrition, not arrivals. If cards are collected at checkout and reused, consumption is not the keys you issue. It is the ones that never come back, plus the ones you retire as cracked, delaminated or no longer reading reliably.
Three times monthly attrition is a defensible floor, roughly 350 cards here, with the reorder triggered at that level rather than when the drawer empties. Size against your peak month: a resort loses far more cards in August than in February. Minimum order quantity usually overshoots this anyway, which is fine, since a 500 piece minimum at 100 cards a month is five months of cover.
Production time is short and measurable: 13 calendar days standard at our plant, 8 express, 12 for wood. What stretches the calendar is everything around it: sample approval, artwork sign off and freight. Sea freight from South China is weeks, not days. Air freight is days, at a multiple of the cost.
The worse cost of a rush is the step that gets dropped. When the drawer is empty nobody waits for a sample, and nobody tests 50 cards on three lock models before the pallet goes into service. The schedule removes the only two checks that would have caught the wrong chip. Transit is broken down in the guide to lead times and freight.
Never move a full delivery into service on one card read at the front desk. Run an acceptance routine the week the pallet arrives, while old stock is still there to fall back on.
Cards are produced in sheets and packed in sequence, so a defect can sit in part of a run. Pull from several cartons and different positions. ISO 2859-1 sets out how sample sizes and acceptance numbers are derived.
Use at least three doors chosen on purpose: one from each lock generation, one on a floor refurbished separately, one back of house door if it runs on the same system. A card that works on the newest lock and fails on the 2011 wing is the failure you are hunting.
Encode a card, open a door, cancel it, re encode it for another room, open that door. Some chip substitutions pass a first read and fail on rewrite. Repeat twenty times: a card that survives one write and fails on the fifth will fail in month two of service.
Thickness and dimensions against ID-1, print registration, laminate edges, and whether the print survives a slot reader. If the cards are pre encoded, check the number range does not collide with cards in circulation. Keep 20 or 30 cards aside, labelled with the production date, as a reference sample. Our delivery inspection checklist lists the physical checks in order.
Guest cards get specified carefully. Staff cards get added to the same quote as an afterthought, and usually wrong.
Housekeeping, maintenance and duty manager cards sit on a different part of the lock system, with different access levels and sometimes a different card type. They also have to work with things guest cards never touch: the time and attendance terminal, the car park barrier, the back office door, a lift controller. Those subsystems were installed by different contractors at different times, and many are still on 125 kHz with EM4200 or T5577 while the guest rooms run at 13.56 MHz.
For a design led property the card is the first object a guest holds, and standard white PVC is a wasted surface. The choice is real, with engineering consequences.
PVC at 0.76 mm is the default for a reason. It is dimensionally stable, prints well in offset and digital, and is what slot readers were designed around. Recycled PVC and PET are available when the property has a materials policy, and both run on the same tooling.
Wood and bamboo change the object completely. We work in basswood, cherry, black walnut and bamboo. Our chain of custody certificate FSC-C195226 covers the wood. Wood cards run 12 days in production. Two things to plan for: the body is thicker and less uniform than PVC, so a card for an insert style slot reader must be checked against the actual lock before a full run, and grain means every card differs slightly, which should be agreed in advance.
Any material change is a compatibility change. Thickness, antenna position and the material around the antenna all affect read range. Moving from PVC to wood is a new specification, so approve a sample on your own locks. Never change material and chip in the same order.
Kaway Group manufactures RFID and NFC cards in its own plant in Dongguan, Guangdong, with an office in Hong Kong. Cards have been produced at the plant since 2007, and Kaway Group Limited was incorporated in Hong Kong in 2015. The plant runs 1,300 square metres with 87 people and a capacity of 8 million cards a month, and has shipped to 53 countries.
This is how we work by default: you send the lock or reader model and a sample card, and we confirm the exact chip in writing before anything enters production. We work at 125 kHz, 13.56 MHz and UHF, with MIFARE Classic, Ultralight, Plus and DESFire, NTAG, ICODE, EM4200 and T5577, in PVC, recycled PVC, PET, wood (basswood, cherry, black walnut) and bamboo.
Minimum order is 500 units. Standard production is 13 calendar days, express 8, wood 12. Quote requests are answered the same working day, and a free sample is provided once the request is qualified. All work is white label, with no Kaway branding on the card, the box or the paperwork, and we do not sell to our distributors' customers. Our hotel work is set out on the hospitality page.
A card that currently opens doors, plus a photograph of the lock model plate. Those two items let a hotel key card manufacturer identify the chip, the frequency and the memory layout directly.
Sometimes, when the lock model and software generation are unambiguous. Often not. If a chip is confirmed without a sample and without the system generation, ask what the answer was based on.
Usually yes, as blank cards of the correct chip type, encoded at the property or by the vendor. No factory can write data protected by keys nobody will release.
Count the cards lost or retired over one real month, hold three times that figure, and reorder at that level, not at zero. Size against your busiest month, not the annual average.
Yes, but approve a sample on your own locks first and change only the material. Thickness and the material around the antenna affect read range, and insert style readers are the most sensitive.
Not separately, but they must be specified separately. They often carry a different access level, sometimes a different chip, and often a second chip at another frequency. Give them their own line and chip confirmation.
About this guide
Published by Kaway Group. The date on this page shows when this guide was last updated. If anything here is out of date, or contradicts what your integrator tells you, tell us and we will correct it.
Send the reader or lock model, the quantity and the date you need them. That is enough for us to answer.
Talk to us