Who May Use It, and How Often
The fallback exists on every site and the rule governing it exists on almost none, which means it is governed by whatever each supervisor thinks is reasonable.
A site with no written rule about the fallback does not have no rule. It has several, one per supervisor, applied inconsistently to different people, and nobody has ever compared them.
The record in “Who May Use It, and How Often” becomes useful only when people understand what it proves and how to correct it. For teams exploring self report bias, read the provider overview can connect time and project context with manager review, provided collection is proportionate, access is limited and consequential decisions remain subject to human explanation.
The paragraph that fixes this is short. What it mainly does is take a set of daily judgements away from individuals who should not be making them and replace it with something the same for everybody.
For an independent benchmark relevant to “Who May Use It, and How Often”, consult the ICAEW audit resources. 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.
The four questions the rule has to answer
Who may use the second method. In what circumstances. How it is recorded. And what, if anything, happens when the volume for one person gets high.
That is the whole scope. It fits on half a page and it is the single most useful document in this collection after the enrolment record.
Who: a named list plus a general case
The named list is people with a recorded reason: an enrolment note, a documented failure rate, an accommodation. They may use the fallback whenever the reader refuses them, without asking, without explaining, every day if that is what happens.
The general case is everybody else, when the circumstances apply: the reader is out of service, the network is down, a badge has been lost and the replacement has not arrived. That covers the whole population on the days it is needed and nobody the rest of the time.
In what circumstances: say them
Reader failure. Network outage. Credential lost, damaged or pending replacement. A refused read by an enrolled person. New starter not yet enrolled. Visitor or transferred worker from another site.
Six lines. The point of listing them is not to be exhaustive; it is that anything not on the list is an exception somebody has to decide, which is exactly the right place for the discretion to sit. Without the list, everything is an exception, so nothing is.
How it is recorded
As a normal time entry, with a flag showing which method was used, and a reason code where the system supports one. Not as a correction, not as an override, and not as a blank to be filled in later.
The reason this matters is covered elsewhere in this section: a record that cannot distinguish a read failure from an administrative amendment is a record that supports very little. The flag is a configuration setting and it is usually available and unused.
What happens when one person's volume gets high
Nothing disciplinary. That is the important half of the sentence.
What should happen is a re-enrolment appointment or a change of method, triggered by a threshold — say, five uses in a month — and owned by whoever runs the system rather than by the person's line manager. Framing it as maintenance rather than as a conversation about the individual is what keeps people reporting problems instead of working around them.
The sentence that should not be in it
"The second method may be used with the permission of a supervisor." That one sentence reintroduces everything described in the rest of this section: the wait, the judgement, the ambiguous record and the asymmetry between people whose hands read and people whose hands do not.
Permission belongs in the exceptional cases, which the list has already separated out. For the ordinary case — an enrolled person whom the reader will not read this morning — permission is being sought for a fact about a sensor, and the only thing it adds is delay.
Reviewing it
Once a year, with two numbers in front of you: total second-method entries as a share of all entries, and the count of distinct people using them.
A rising share with a stable headcount means the enrolment base is degrading. A rising headcount with a stable share per person means the workforce changed. Both are ordinary operational findings with ordinary fixes, and neither is visible without the rule that made the entries identifiable in the first place.
Telling people the rule exists
A rule nobody has read governs nothing. The people who need this one are at a terminal at six in the morning, and they will not be reading a policy document.
So it needs to exist twice: in full, where policies live, and in three lines on a card at each terminal. The card version is the one that does the work. Circulating the full version by email and considering the matter communicated is how sites end up with a written rule and an improvised practice that differ in every particular.