Working Time Records

It Was Always Documented

The objection to recording trust-based working time is that it destroys the trust. It is worth noticing that the hours were already being documented, thoroughly, by systems nobody chose for the purpose. For a separate operational perspective on measuring work and activity, see this overview.

What already exists

Email timestamps. Every message carries the time it was sent, retained for years. For a separate perspective on work, technology and organisational practice, see Haufe.

Chat and collaboration logs. Message times, presence status, activity indicators.

Access logs. Building entry, VPN connections, system logins.

Calendar entries. Meetings, with start and end times.

And file modification times, on every document touched.

Between them these describe a working day in considerable detail, they are retained longer than a time record would be, and nobody agreed to them being a record of working time.

Why that matters

The choice was never between documentation and no documentation. It was between an agreed record and an accidental one.

The accidental one is worse in every respect. It was not designed to measure working time, so it measures something adjacent — presence, activity, availability. It is scattered across systems, retained under different policies, and nobody can see their own version of it.

And it is available to the employer and not to the employee, which is the exact asymmetry the ruling was addressing.

Where the accidental record actually gets used

In disputes, retrospectively. An employer defending an overtime claim will look at badge logs and email times. So will an employee's representative.

Which means both parties already rely on it, in the situation where it matters, and both do so with a dataset that was not built for the question.

A proper record replaces that with something designed for the purpose — narrower, agreed, visible to both, and retained for a defined period rather than indefinitely.

The argument to make internally

Not "the law requires it", which produces resistance.

"Your hours are already visible in six systems. This replaces that with one record you can see."

That is accurate, it is checkable, and it reframes the introduction from surveillance to consolidation — which is the honest framing given what the record is for.

The condition attached

It only holds if you actually stop relying on the accidental record.

An employer who introduces time recording and continues to reference login times in performance conversations has added a record rather than replaced one, and the argument above becomes untrue.

Purpose limitation is the principle here, and it is enforceable as well as fair: the time record is for working time, and the accidental data was never collected for it either.

What this does not resolve

The visibility of overtime. A trust-based arrangement quietly producing fifty-hour weeks now produces a record of them, and that is uncomfortable for both parties in different ways.

But it was happening. The record does not create the hours; it makes them answerable. And an arrangement that only worked while nobody counted was not a trust arrangement — it was an untested one.

The short version