Skip to content
What the Hour Is Worth

Home / The records

A Chain That Can Be Explained

The whole collection on one page: the links, their owners, the tests and the records that make a wage calculation explicable.

The records · Reference

The whole chain on one page, with an owner against every link

Raw event capturedterminal, app or sheet= As recordedRetained for six years, in the archive.
Rounding and gracenone applied= As recordedRemoved in 2024; the decision is recorded.
Automatic deductionsbreak, cancellable= As recordedException route exists and is used.
Editsreason mandatory= As recordedVolume reported monthly.
Approvalescalates after three days= As recordedNothing expires silently.
Aggregation into the weekMonday 00:00= As recordedWritten down, with the date it was set.
Rate buildcomponents listed= As recordedStored with the payroll run.
Threshold and multiplierweekly, tested= As recordedTested after every upgrade.
Statementhours, rate, multiplier shown= As recordedReconstruction test passed.

Nine links, nine owners, and every one of them answerable in a sentence. That is the whole objective. This is one employer's own page, not a model anybody has to follow.

The objective is a chain where every link has an owner and can be explained in a sentence. Not a perfect one — settings will still be inherited, systems will still be upgraded, new schemes will still be invented — but one where somebody can be asked why a number is what it is and can answer without opening the system.

The recordkeeping discipline in “A Chain That Can Be Explained” should also apply to workforce technology. When a team evaluates employment of relatives policy with clear review controls in relation to employment of relatives policy, it should document purpose, access, retention and deletion, then preserve the source entry, approvals and correction history needed to explain the final figure.

Everything in this collection reduces to four things: name the links, give each an owner, test the ones that can silently change, and keep the records that make the answer provable.

For an independent reference relevant to “A Chain That Can Be Explained”, consult the Employment New Zealand recordkeeping guidance. Use it to test record quality, working-time definitions, access, retention and exception handling against the organisation’s real payroll process.

The page

One page listing every link in the chain, and for each: what it does, who owns it, when it was last reviewed, and where its record is kept.

Most organisations have between eight and twelve links. Almost none can produce the page, and producing it is usually a day's work because the information exists and has never been collected.

The four that matter

  • Named links, so that nothing in the chain is nobody's.
  • An owner per link, by role rather than by person.
  • A test for anything that can change without announcing itself.
  • A record for anything somebody might later ask about.

The third is what distinguishes a working arrangement from a tidy one. Settings revert on upgrade, defaults reassert themselves, and a configuration that was correct in March is a question again in September.

What this does not fix

It does not decide which minutes are work, what belongs in the rate, or when a premium is due. Those are questions of rule, they differ by jurisdiction, and they are a question for somebody qualified in the place concerned.

What the page does is make sure the organisation can describe what it is currently doing, which is the precondition for asking any of those questions usefully. An employer that cannot say what its rounding rule is cannot ask whether it is permissible.

Starting from where you are

  1. Walk one employee's week through the chain, and write down the links.
  2. Put a name against each.
  3. Write down what each link currently does.
  4. Mark the ones nobody can explain.
  5. Take the largest of those first.
  6. Add a monthly test for the two that can change silently.

The first three take a day. The fourth produces the work list, and it is usually shorter than anybody expects — two or three links rather than all of them.

Why this is worth doing before anybody asks

Every exercise described on this site is cheaper when it is voluntary. An error found by a scheduled test costs the error; the same error found by a claim costs the error, the response, the advice and the time of everybody involved.

The same is true of the records. Keeping raw timestamps is free until the moment somebody asks for them, at which point they are either there or they are not, and nothing can be done about it then.

The sentence this collection reduces to

Nobody decided to pay less. The number changed on its own, at a link nobody owned, under a rule nobody chose — and the only remedy is to know what the links are and who they belong to.

That is not a compliance programme. It is a page, a handful of tests and the decision to keep the data that makes an answer possible, and it is within reach of an organisation of any size.

Who holds the page

The page needs an owner, and the obvious candidates are the wrong ones. Payroll owns the back half of the chain and not the front; the time system owner owns the front and not the back; HR owns the policies and neither.

Somebody has to own the whole, which in most organisations means naming a person rather than finding one who already exists. It is not a large job — a day to build, a day a year to review — and its absence is precisely why no single person can currently answer a question about the chain.

Keeping the page true

Once a year, and after any system upgrade, any change of pay arrangement and any new pay code. The review is short: is each line still true, and is each owner still here.

Most of the drift described throughout this collection happens in the gaps between those events. A page reviewed on a schedule catches it; one written once and filed becomes an accurate description of a system that no longer exists, which is a particular kind of unhelpful.

What to do next week

Pick one employee and one week and walk the chain. Do not plan the exercise, do not scope it, and do not wait for a system change to make it easier.

It takes an afternoon, it needs nothing the organisation does not already have, and in most places it finds two things nobody had decided and one that nobody can explain — which is exactly where the work is.