The Reader Nobody Cleans
A time clock is the only piece of equipment on site touched by every employee every day and maintained by nobody. Its performance degrades accordingly.
Every other device on a site with that duty cycle has a maintenance schedule. The time clock has none, because it was bought by whoever buys software, installed by a contractor, and inherited by nobody.
The failure described in “The Reader Nobody Cleans” 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 explore the software for task switching cost should test retries, missed punches, corrections and employee review while preserving a non-biometric fallback that does not depend on finding a supervisor.
The consequences are gradual and therefore invisible. Nothing fails; everything gets slightly worse, over about eighteen months, until the refusal rate is high enough that somebody proposes replacing the system.
For an independent benchmark relevant to “The Reader Nobody Cleans”, consult the FTC data-security guidance. 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.
What actually degrades
The sensor surface, from residue. The mounting, which works loose and changes the presentation angle. The keypad, where the most-used digits wear smooth and then stop registering. The cable, at the point where it enters the wall. And the firmware, which is several versions behind and may contain the fix for the exact matching problem being experienced.
None of these announce themselves. A loose bracket that has tilted the reader by fifteen degrees produces a rising refusal rate and looks like nothing at all.
The routine, which is short
Weekly: wipe the sensor and the keypad with the cloth and the cleaner the manufacturer specifies, not with whatever is in the cupboard. Some sensor coatings are damaged by solvents, and a well-meant clean with the wrong product is a known way to destroy a reader.
Monthly: check the mounting is tight and the angle unchanged, look at the cable entry, and test a read yourself. Quarterly: check the firmware version against the current one, and read the release notes, which occasionally say something directly relevant.
Annually: review the refusal rate against the same month last year. A device that is worse than it was has something wrong with it that the daily experience has normalised.
Who owns it
This is the real question, and the answer on most sites is nobody, because the device sits across three boundaries. Facilities own things attached to walls. IT own things on the network. HR or payroll own the data. The time clock is all three and therefore none.
Assign it to whoever already has a maintenance round on that building, with the routine written on a card. It is five minutes a week. The alternative is not that it gets done less often; it is that it never gets done at all, which is the current state almost everywhere.
The terminal that is in the wrong place
Maintenance also means noticing what has changed around the device. A reader that was well sited in 2019 may now be behind a stack of pallets, next to a new roller door that lets cold air in, or in a corridor that became the main route when the entrance moved.
Walk the approach to the terminal once a year at shift change. The questions are whether people can reach it without queueing past something, whether the light has changed, and whether the thing it was mounted next to is still there. Sites change constantly and the reader stays where it was bolted.
Spares, and the hour after a failure
Hold one spare terminal, configured, on the shelf. A site with four readers and no spare is a site where a hardware failure means a week of paper, because the replacement has to be ordered, shipped and commissioned.
The spare needs to be configured and tested when it arrives, not when it is needed. A boxed unit with factory settings and no network configuration is a reassuring object rather than a spare, and the difference is discovered at the worst possible moment.
What a maintained reader looks like in the figures
Flat. The refusal rate for a given month should look like the same month last year, allowing for the seasonal pattern.
That is the entire test, and it requires only that somebody writes the number down once a month. A site that has twelve numbers can see degradation coming a year before it becomes a crisis; a site that has none finds out when a supervisor says the clock has got worse, which is both true and far too late to be cheap.
What to do with the old unit
When a reader is finally replaced, the old one goes into a cupboard rather than through any process, and it may still hold templates and event data. That is covered properly elsewhere in this collection and it belongs on the maintenance routine too.
Add one line to the card: a unit taken out of service is wiped, recorded and disposed of, not shelved. The shelf is where this problem accumulates, because an object in a cupboard is nobody's decision and stays there until the cupboard is cleared by somebody who has no idea what is in it.