Is a shared Google Sheets roster good enough for a pub or cafe?

The short version: a shared Google Sheet is legal, it ends the five-versions problem, and for a single-room venue with one editor and a handful of casuals it can be enough. At 20 staff with a kitchen and a bar it fixes the wrong half of the job: the file was never the expensive part, the editing and the checking were. This is general information, not legal or financial advice; the Fair Work Ombudsman is the source of truth for the Hospitality Award (MA000009).

Six months ago you retired roster_FINAL_v3.xlsx. You made a Google Sheet, shared the link with the team, and for the first fortnight it felt like getting a weekend back. Nobody asked which version was current, because there was only one. The head chef stopped emailing you his copy. The screenshot in the group chat matched the grid.

And yet here you are, on a Sunday afternoon, reading about rosters, because the roster is still what Sunday afternoon is for. That is not a sign you made the wrong move. It is a sign the move fixed the problem it could fix and left the other one where it was.

What a move to Google Sheets fixes

We want to be fair to the Sheet, because it earns its place in a lot of venues. Compared with a file that lives on the office laptop, a shared Sheet fixes five real things.

• One current version. The link is the roster. Whoever opens it sees what is there now, not what was there when the screenshot was taken.
• Phone access. The Sheets app opens on the bus. A casual can check Saturday without texting you, which is a text you no longer answer.
• A history of who changed what. Version history shows which account edited which cell, and when. "I swear I never moved him" has an answer now.
• Two managers in the grid at once. The head chef and the venue manager can both be in it on a Wednesday night without anyone emailing a copy, and each can see the other's cursor.
• Free. A Google account is all it takes, and the venue almost certainly has one already.

If the file was the thing hurting you, that list is the cure, and plenty of venues should stop here. The trouble is that for most 20-staff venues the file was the symptom.

What a shared Sheet does not fix

Five things, and they are the same five whether the grid is in Excel, in Sheets or on the wall by the till. The Sheet holds a plan. Everything below is about the rules around the plan, the inputs to it, and what happens after it.

It cannot read the award against the grid

Clause 15.5 of the Hospitality Award is headed Rosters (Full-time and part-time employees), and it holds three rules a grid will let you break in silence. A permanent's roster can be changed by mutual agreement or by the employer giving 7 days' notice (15.5(d)). An employee must have a minimum break of 10 hours between finishing ordinary hours on one day and starting them the next, or 8 hours on a changeover of rosters (15.5(e)). And the whole clause applies to full-time and part-time employees only (15.5(a)), so at a 22-staff pub with six permanents it covers six names, and the Sheet does not know which six.

A close at 1am and an open at 9am is two adjacent cells that look completely normal. A shift moved on Wednesday for Saturday is a cell that changed colour. The Sheet cannot tell a casual from a part-timer, a Level 2 from a Level 3, or an agreed change from an imposed one. We have been through the 10-hour break rule and what the award says about notice separately. Both pieces exist because a grid cannot apply either rule, and sharing the grid does not teach it to.

It does not collect availability or leave

The Sheet is where the answer goes. The questions still arrive the old way: "can't do Fridays until exams finish", "I'm away the long weekend", "mum's in hospital, can I swap Thursday", in a group chat, a DM, or at the pass on Tuesday. The manager carries them in their head to the grid on Sunday, and the one that arrived three weeks ago to the other manager, who is on days off, does not make it. A shared Sheet has not changed one thing about how availability is collected, only where it ends up once someone remembers it.

It does not become a timesheet or a payroll export

Rostered hours are not worked hours. The Fair Work Ombudsman's record-keeping page lists, among the hours-of-work records an employer must keep, "the hours an employee works if the employee is a casual or irregular part-time employee who is paid based on time worked". At a casual-heavy pub that is most of the team, and it is a record of what happened, not of what was planned. The Sheet holds the plan. The hours still come off paper timesheets or a clock-in app, and someone still reads them and types 22 people's hours against the right pay item, once a fortnight. The re-keying line in why free roster templates cost more than you think is unchanged by the move to Sheets, and so is the error rate it carries.

Anyone with the link can edit anything

Sharing the link gave the head chef a live roster. It also gave him the main bar. Google Sheets can protect a tab or a range so that only named people edit it, or show a warning instead, and Google's own help page says this "shouldn't be used as a security measure" and that people can still copy, print and export a protected spreadsheet. More to the point, a protected range is a rectangle on a grid. The one person who works the bar on Friday and the kitchen on Saturday sits in two rectangles with two owners.

Here is how the Sarah problem goes on a shared Sheet. Sarah is a Level 2 casual on the main bar every Saturday. On Wednesday a 21st upstairs firms up at 80 guests, so the head chef, who runs functions, moves her to the function bar. On Thursday the venue manager opens the Sheet, sees the main bar a person short, cannot see why, and moves her back, because she is always main bar. On Friday the head chef sees functions short again, assumes he forgot, and moves her a third time. Whichever edit was last wins. Nobody told Sarah any of the three times, and the version history shows three edits by two accounts and nothing about why. The two managers are no longer holding two rosters. They are holding one, and each is quietly undoing the other inside it.

A shared Sheet solves the file problem and doubles the editor problem.

It does not tell you what the fortnight costs until payroll

A Sheet will hold a cost formula, and the formula multiplies whatever number is in the cell. It does not know that Jack moved from Level 2 to Level 3 in April, that a casual's Saturday is 150% of the minimum rate while a part-timer's is 125%, or that Anzac Day fell on a Saturday this year and the Saturday formula was $27.08 an hour short for every Level 2 casual on it. So the running total on the grid is a guess with a dollar sign in front, and the first real number for the fortnight is the pay run, after the hours have been worked and the roster can no longer be changed.

The one-week test

Take the pub from our guide to moving off Excel: 22 staff, six permanents and 16 casuals across bar, kitchen and floor, two people editing the roster, payroll fortnightly, and a salaried venue manager whose time costs about $45 an hour once super and workers compensation are added. That guide costed the file version of this roster at about $13,700 a year. Here is the same week, six months after the move to a shared Sheet. Everything is an estimate except the award figures, which are MA000009 rates from 1 July 2026.

The difference between $13,700 and $12,300 is the 35 minutes a week that used to go on chasing versions, re-sending screenshots and fixing the wrong copy. The Sheet took that off, and it is a real $1,400 a year. The other $12,300 is the half it never touched: the collecting, the checking, the re-keying and the one error that recurs because nothing is reading the award. The Sunday afternoon in the opening paragraph is mostly rows three and four, and we have written about where those hours go.

When Sheets is honestly enough

Keep the Sheet, with a clear conscience, if all of these are true.

• One room and one editor. Everyone else has the link to read it. Not one person who is supposed to edit, one person who does.
• A small, settled team. Roughly ten names or fewer, mostly the same people on the same days, so you are copying last fortnight forward rather than solving it from scratch.
• Hours are recorded somewhere else. A timesheet process, paper or app, that someone reconciles against the roster before pay, rather than paying the plan.
• Someone checks the award by hand on every change. The 10-hour break, the day's rate, who is a permanent and who is a casual, which classification. Every time, not most times.
• Changes are rare. One or two swaps a fortnight, not one or two a day.

If that is your venue, two small habits make the Sheet a better record. Name a version at the end of each fortnight, because Google's version history page says revisions "may occasionally be merged" and that only people with edit access can see the history at all. And protect the published fortnight so it can only be changed on purpose, which is the Sheet's version of "not altered except to correct an error".

Where Shiftly fits

We built Shiftly for the venue where two of those five have stopped being true. The roster is one roster: it is published once, it lands on every phone through Shiftly Me, and a change reaches the person it touches, which is the half of the Sarah problem that matters. You can view and filter it by position, so the kitchen and the bar are views of the same grid rather than two tabs with two owners. Each manager gets a permission matrix: which locations they run, and whether they can view or edit the roster, people and pay. It is scoped by location rather than by department, so in a one-room pub two managers still share the roster, but they share one published version with a notification to the person moved, not a grid they quietly undo each other in.

Casuals enter their own availability and leave in the app, so the group chat stops being the intake. Each person carries an award level, and the award interpretation estimates what a Saturday, a Sunday or a late-night hour costs at that person's rate as you build, and a double-booking or a shift on someone's unavailable day shows up on the roster before you publish. Worked hours come from the timesheets staff clock on the app, approved hours flow into the pay calculation and sync to Xero, and nobody types 22 people's hours by hand. All of it is free at any headcount, with no per-employee fees, so 22 staff costs the same as six.

Shiftly is a calculation tool and a staffing network, not a payroll provider. It helps estimate pay against the award, your business stays responsible for verifying it, and the Fair Work Ombudsman remains the source of truth. The staffing side, where an open shift goes out to local hospitality workers from inside the roster, is coming soon, starting in Sydney, with a waitlist open.

Frequently asked questions

Is it legal to roster in Google Sheets?

Yes. Neither the Fair Work Act nor the Hospitality Award names a tool. Clause 15.5 of MA000009 requires, for full-time and part-time employees, a roster showing each employee's name and their start and finish times, posted in a conspicuous place that is easily accessible by the employees. A shared Sheet the team can open on their phones meets that at least as well as a printout by the till. The law cares about outcomes and records, not the software: a permanent's roster changed with agreement or 7 days' notice, a 10-hour break between ordinary hours, and hours-worked records for casuals kept for 7 years. A Sheet can sit inside all of that. It just will not check any of it for you.

Does version history count as a Fair Work record?

It is a record of edits, not the record the law asks for. Fair Work's time and wages records cover pay, loadings, penalty rates, leave and, for casuals paid on time worked, the hours actually worked. They must be kept for 7 years, be readily accessible to a Fair Work Inspector, be legible and in English, and not be changed except to correct an error. Version history shows which account changed a cell of the plan and when. It does not show that a permanent agreed to the change, and it does not record what anyone worked. Google also says revisions may occasionally be merged, and only editors can see them. Keep the hours record separately, and export a dated, read-only copy of each fortnight's roster.

Can I stop one manager editing the other's department in Google Sheets?

Partly. Protected sheets and ranges let you restrict a tab or a block of cells to named editors, or show a warning that does not block anything. Google's help says the feature should not be used as a security measure, and in practice a protection is a rectangle drawn on the grid that has to be redrawn when the layout changes, while the person who works both the bar and the kitchen sits in two rectangles with two owners. It is a useful guard against accidents. It is not per-department permissions.

At what point does a shared Sheet stop being enough?

When two of three things arrive: a second department, a second editor, or a casual-heavy team whose availability changes weekly. Headcount on its own is not the trigger. A 30-person cafe with one room and one editor can run on a Sheet, and a 14-person pub with a kitchen, a bar and two managers can fall apart on one. The honest tell is the week: if the roster still takes most of a Sunday afternoon six months after the move, the Sheet fixed the file and the rest is waiting.

The bottom line

If you are one room and one editor, keep the Sheet. Name a version each fortnight, protect the published one, record hours somewhere else, and bookmark the complete guide to moving off Excel for the day a second manager or a second room arrives. If they already have, the fortnight plan in that guide is the next read, and Shiftly's rostering is what the plan moves you onto. The Sheet fixed the file. Sunday afternoon is the other half.