Who changed the roster? Version control for the most used document on site

Every prestart has heard a version of that exchange. Some workers miss shifts because they could not attend. Others miss shifts because they were looking at the wrong roster. The second kind, call it coordination absenteeism, has nothing to do with commitment, and it is the most fixable absence in mining.

The roster keeps changing, and nobody is sure which one is right

On most sites the roster is a spreadsheet, and spreadsheets travel. One copy goes out by email, another gets printed for the crib room, a third lives on someone's phone from three weeks back. Then a swing changes, a name moves, and the versions quietly disagree. Half the crew is planning their week from a document that is no longer true, and there is an award-shaped problem hiding in there too: if a published roster changed without proper notice, notice most awards set at seven days, nobody can even prove which version was published when.

Nobody is wrong, which is why it never gets fixed

When someone no-shows off a stale spreadsheet, it looks like an attendance failure and gets managed like one. It is not. The information existed, the superintendent did their job, the worker believed the version in front of them. Discipline fixes none of that.

It also does quiet damage between people. When the crew cannot trust which version is current, every roster gets read with suspicion and every change becomes a rumour. The document that exists to coordinate the team starts unsettling it instead. The UQ turnover researchers heard where that ends from a site manager: when the numbers go down, we do not know if it is absenteeism or turnover.

What fixes it: one roster, drafts, and a change log

• Drafts and one published roster. The planner works in draft freely. The crew only ever sees the published version, so "which roster are we following" has exactly one answer.
• Notify the team, or stay silent until ready. Publishing a change is a deliberate choice, not an accidental leak through a forwarded spreadsheet.
• A change log. Every edit records who, what, and when. "Who changed the roster" gets answered in seconds, not argued at prestart. It is also the notice-period evidence if a dispute ever reaches Fair Work.
• Read versus edit. Everyone can see the roster. Only accountable people can change it.

One tool for one type of absence

Absenteeism has many causes and no single fix. Health, fatigue, and morale need their own tools, and confirmation windows handle the human no. But absence born of version confusion has a complete, boring, technical cure: one centralised published roster with a change log the whole squad can trust.

What we see once there is one roster

The surprise, watching published rosters replace travelling spreadsheets, is that the change log's daily value is not disputes. Fair Work evidence matters perhaps once a year. What changes daily is trust. Crews stop screenshotting the roster "just in case," because versions can no longer drift. The question at prestart shifts from "which roster are we following" to "who is covering Thursday," which is the conversation a roster was always supposed to start. Coordination absenteeism does not get managed down. It simply stops having anywhere to live.

Source: Beach et al., Workforce Turnover in FIFO Mining Operations in Australia, UQ Centre for Social Responsibility in Mining, full report.