A property manager in Rotterdam replaced her lobby controller in March. Four months later 1,800 of her 2,300 key fobs would not open the door, and re-credentialing cost €9,400. The software was fine. The fobs were 13.56 MHz and the new readers wanted 125 kHz.
When you search for door access control companies you expect one company. There are four, and only one touches the plastic people hold up to the door.
When A Door Stops Reading, Who Gets Blamed
Ask a facility team where access control problems come from and they will name the software. Industry data points elsewhere. A 2026 review found more than 90% of facilities report an access control failure every six months, and about 34% trace back to wrong technology choice or poor implementation.
That 34% is a procurement number, not a maintenance number. It gets locked in at the buying stage, long before anyone stands at a door. Most door access control companies on page one sell software rather than plastic, so nobody in the buying room argues for the credential spec. The door access control companies blamed first are almost never the ones that caused it.
The Four Types Of Door Access Control Companies
Platform and software companies. Cloud dashboards, mobile credentials, per-door subscriptions. Verkada, Brivo, Kisi, Salto. They set the rules your credentials must obey.
Controller and reader manufacturers. The panel in the closet and the reader on the wall. This layer picks the frequency and card format for the site.
Credential manufacturers. Cards, key fobs, wristbands, tags. 125 kHz, 13.56 MHz, dual frequency. Toptag Group sits here.
Integrators and installers. They wire it, commission it, and run enrollment.
Buyers push hard on the first two of those door access control companies and treat the third as a line item. The credential layer is the only one you can get wrong and not discover until the doors are live.
Why The Credential Layer Fails First
Four things go wrong, all decided before production.
Frequency mismatch is the big one. A 13.56 MHz fob is invisible to a 125 kHz reader. No weak signal, no partial read.
Chip generation comes next. EM4100 and TK4100 are read-only UID chips. T5577 is rewritable. Mifare Classic 1K carries 16 sectors of 64 bytes. Some F08 chips sold as 1K compatible mimic the UID but not the sector map. If your controller stores credentials in sector data instead of the UID, an F08 card reads at the door and logs as unknown. Buyers evaluate door access control companies on software features, then hand the spec to whoever answers the phone fastest.
Printed serials against encoded UIDs are the third trap. Some factories print one number and encode another. Enrollment types the printed number, the system searches for the UID, and every card needs hand-enrollment.
The environment is the fourth. Metal frames, aluminium phone cases and thick dashboards eat read range. Low frequency prox cards tolerate it, high frequency cards do not.
A security integrator in Johannesburg learned all four at once. He re-ordered 40,000 fobs for a mine by copying the old spec, 13.56 MHz Mifare 1K. The mine’s readers were 125 kHz. About 6,000 fobs ever opened a gate, the shipment sat in a warehouse for eleven weeks, and the final 60% payment was withheld.
Six Questions That Reveal Who Makes Your Cards
Which exact chip, and can you send the datasheet? Compatible with Mifare is not a specification. Mifare Classic 1K S50 is.
Can you encode a sample batch against our reader before production? If the answer is no, you are buying blind.
Do you print the serial numbers, and do they match the encoded UID? Get this in writing.
What read distance should we expect on our reader model? A number, not a range.
Do you make the cards in your own plant? Trading companies are not automatically bad, but you cannot audit a process you have never seen.
If 3% of the batch will not enroll, what happens? Free replacement, a credit note, or nothing.
Most door access control companies answer the first three without blinking. The last three are where the conversation gets quiet. Our RFID key fob guide covers the enrollment side in more detail.
What A Wrong Credential Order Really Costs
The cards are never the expensive part. A 125 kHz fob costs cents. The rework does not.
A retail chain in Poland needed 12,000 cards across 320 stores. The factory printed serial numbers that did not match the encoded UIDs, so every card had to be enrolled by hand at every site. The cards came in under €4,000. The enrollment labor landed near €38,000, and 320 store managers each lost half a day.
Add shipping on the replacement run. This is the cost door access control companies never put in the quote. Weigh it against one extra week of samples.
Den RFID cards for access control page breaks down the frequency question in detail. Low frequency sites can start with our LF 125 kHz ABS keyfob sourcing guide.
Quick Answers On Sourcing Access Credentials
Do we need 125 kHz or 13.56 MHz?
Match the reader you already own. Ask three door access control companies and you get three answers, so start from your hardware. For a new build, 13.56 MHz gives more chip options and stronger security, while 125 kHz is cheaper and reads further.
Can we run both frequencies in one building?
Yes, with dual frequency credentials or two access groups. Some door access control companies sell a mixed batch as a shortcut. Mixing silently is how half a shipment stops working.
How long should a sample batch take?
Two to three weeks with custom printing, faster on stock housings.
What MOQ should we expect?
Stock housings start around 500 pieces. Custom tooling or custom encoding pushes that up.
Are rewritable cards a security risk?
T5577 and similar chips can be rewritten by anyone with a copier. For high-security doors, buy read-only or encrypted chips.
Can we print logos and employee names?
Yes, and it costs less than buyers expect. Just confirm the printed number matches the encoded UID first. It is also the question door access control companies least enjoy answering.
The next time you compare door access control companies, ask which of the four layers each one actually owns. The company that makes your cards decides whether 2,300 people get through the lobby on Monday morning.
If you want a straight answer before you commit to a batch, send us your reader model and the chip you are using. We will tell you within a day whether those cards read at your doors.