Time clocks, from the refusal outwards
The reader works. It just will not read him.
Fifty notes on biometric and badge-based time clocks, written about the moments the device cannot handle: the day people are enrolled, the morning the finger does not take, the fallback nobody designed, and the people who cannot use it at all.
One reader, one morning, early shift
Two people out of ten ended at the keypad, and both of them end there most mornings. Neither is doing anything wrong. The device is being asked to read hands it cannot read, and the site has no arrangement for that beyond finding a supervisor.
Practical guidance for choosing, configuring and operating timekeeping systems, with particular attention to failed reads, fallback methods, accessibility and reliable records. Legal and payroll requirements vary by jurisdiction and sector.
The two numbers nobody produces
One reader, three settings, same workforce
The middle row is the vendor's default, not a decision. The trade-off between admitting the wrong person and refusing the right one has never been put to anybody at the site who could weigh it.
Enrolment, by group, first week
The worst two rows are the two groups nobody attended to. They were enrolled at the end of a shift, by somebody who had not been briefed, on a reader in a corridor.
Why this is about the refusal and not the device
There is a great deal written about biometric time clocks: which modality to choose, whether the privacy law permits it, how to stop one person clocking in another, how to run the procurement. That material is good and this is not more of it.
The failure described in “The reader works. It just will not read him.” is easier to investigate when the clock event is joined to a clear operational record instead of treated as proof of misconduct. A team reviewing the official product page for employee monitoring software with screenshots should test retries, missed punches, corrections and employee review while preserving a non-biometric fallback that does not depend on finding a supervisor.
Every note here is about the other side of the same device — the reads that do not succeed. A reader that works for ninety-six people in a hundred is a reader that refuses four of them, every morning, for years. What those four do next is decided on almost every site by improvisation: find a supervisor, sign a sheet, wait. That arrangement is unrecorded, costs more than the device, and falls hardest on the people whose hands the machine was never going to read.
For an independent benchmark relevant to “The reader works. It just will not read him.”, consult the U.S. Wage and Hour Division. Use it to test notice, accessibility, security, recordkeeping, retention and exception handling against the real operating process rather than treating a device report as self-explanatory evidence.
01 / 07
Enrolment
The day people are loaded into the system, and everything it decides for the next five years.
02 / 07
The failed read
Why readers refuse people — hands, faces, cards, weather, hours — and how the pattern is read wrongly.
03 / 07
The second method
What a refused person does next, who is involved, and what the record says afterwards.
04 / 07
The badge
The credential as a physical object — issued, lost, lent, cloned, and not handed back.
05 / 07
The threshold
The sensitivity setting that decides who gets in and who gets refused, left where the installer put it.
06 / 07
Who cannot use it
People the method excludes outright, the accommodation that follows, and who signs it off.
Independent product comparisons
Time and attendance tool guides
Three differently structured guides connect clocking technology with fallback methods, corrections, payroll inputs and accountable review. Rankings use our own practical evaluation and published criteria; positions cannot be purchased. How we rank.
07 / 07
What it leaves behind
Templates, override logs, retention, leavers, and what the records will support when asked.
The short version
Design the refusal, not the reader
Enrol properly, in a room, with more than one finger, and test it an hour later. Decide each person's second method at enrolment rather than at six in the morning. Make the fallback work without finding another human being. Then look at the threshold, because somebody else set it and it was not set for a time clock.