When we first met Lena, she ran operations for a four-location restaurant group with 42 staff. Her toolkit, paid for entirely in stress, looked like this:
- A Google Sheet with 14 tabs (one per week)
- A WhatsApp group called "shifts"
- A second WhatsApp group called "shifts urgent"
- A third group called "sick today"
- A clipboard at each venue for sick notes
- A shared inbox for receipts she had to back into payroll once a month
Payroll day was Friday. It took her until Monday.
What broke first
The math wasn't the problem. The math was a symptom.
The actual problem was that the schedule, the clock-ins, and the leave records all lived in different places, and disagreed with each other by Tuesday. By the time she sat down to pay people, she was reconciling four different versions of reality.
A typical Friday went:
- Open last week's clock-in sheet (paper at the venue)
- Cross-check against the WhatsApp swap log
- Email anyone whose hours looked off
- Wait for replies, often into the weekend
- Manually compute night-shift premiums (one venue runs late)
- Manually compute holiday surcharges (some weeks)
- Build the payroll file
- Send to the accountant
- Field disputes for the next two weeks
The team didn't trust the numbers. Lena didn't either.
What changed
We migrated her in a day. The actual workflow shift was small but the consequences were large:
One source of truth. Every shift is a record. Every clock-in is a record. Every sick note is a record. They share IDs.
The schedule and the clock-in agree. When Sara clocks in at 09:24, that event references the shift she was scheduled for. If she clocks in late, the system knows. If she's not scheduled at all, the system knows that too.
Surcharges are computed, not estimated. Night premiums and public-holiday rates are configured per branch. They apply automatically.
Leave is part of the same engine. A sick day shows up in payroll the same week, not a month later when someone remembers to email the accountant.
What it actually saved
Three things, measured:
| Metric | Before | After | |---|---|---| | Payroll cycle time | 8–12 hours | < 1 hour | | Disputes after pay-day | 3–6 per cycle | 0–1 | | Hours/week on scheduling | ~6 | ~1.5 |
That's roughly 8 hours a week back per manager, which lines up with the industry research showing SMB owners lose 10+ hours weekly to manual computer tasks.
What we'd tell anyone in the same spot
Don't start with payroll. Start with the clock-in.
If your clock-in is honest, your timesheet is honest. If your timesheet is honest, your payroll is honest. Most of the chaos comes from upstream, the schedule that doesn't match reality, the swap that didn't get logged, the sick day that never made it into the file.
Fix the source. The payroll math will follow.
Want to see what this looks like for your team? Book a 30-minute demo, we'll walk through your current Friday and show you what it looks like as a Tuesday.