The problem
Proposal · based on our own experience, not on measured data
Scheduling sliding shifts by hand is hard because every change ripples: move one person and the coverage, the hours and the rest rules of several others move with it.
What sliding shifts are
In a store with sliding shifts, people do not all start at 9:00 and leave at 18:00. Starts and ends are staggered so that more people are on the floor when traffic peaks, for example at lunchtime and after work, and fewer when the store is quiet. A day might look like this (illustrative):
| Person | Shift | Why |
|---|---|---|
| Anna | 08:00–16:00 | Opening and morning delivery |
| Ben | 10:00–18:00 | Covers the lunch peak |
| Chloe | 12:00–20:00 | Covers the after-work peak |
| Dan | 14:00–22:00 | Evening and closing |
Why it hurts
When we built these schedules, the difficulty was not any single rule. It was keeping all of them in mind at once:
- Coverage. Enough people at each hour, including someone who can open, close or run the till.
- Availability. Students, part-timers and people with fixed days off.
- Hours. Contracted hours per week, overtime limits, fair spread of weekends and late shifts.
- Rest. Minimum rest between a closing shift and the next opening shift.
- Changes. A sick day or a swap request after the schedule is out means solving the puzzle again, usually by the manager, usually under time pressure.
Training helped people understand the rules. It did not make the puzzle smaller. That is the part we think an agent could take over.
What we do not know yet
We have not measured how long managers spend on scheduling today, how often schedules change after publication, or how much over- or under-staffing costs. These are the first things to find out with real stores. See Open questions.