BPOs and contact centres
Contact centres staff a queue, not a building. A window opens at six, and the question is whether your seats were filled when it did. PeopleMuster answers that per window. Named shifts are assigned per agent, punches come from a floor terminal or the browser, and late minutes read against the window you gave each person.
The shift pattern
Rotating windows, several of them, and a night one. Each is a named shift with its own start and end. Morning, evening and night are three definitions, not one office day with exceptions.
You assign per agent, from a start date, with a duration and the name of whoever made the change. When a campaign moves someone from evenings to nights, the assignment history tab records who moved them and when. Six weeks later, that tab is what settles the dispute.
Client campaigns are what pull your windows around. A client whose afternoon is your night wants the queue open in their afternoon, and every client record shows that client's local time beside yours. At least the rota argument starts from the same clock.
How attendance is taken
Two paths. A ZKTeco fingerprint terminal at the floor entrance, reached by its IP on your local network, or browser check-in that an admin switches on. Fingerprints stay on the terminal. PeopleMuster receives only who punched and when.
Browser check-in opens inside a window around each agent's assigned shift. By default the card appears an hour before the start and closes two hours after it, and check-out opens half an hour before the end. Your admin can move every one of those edges, which matters when the handover between two windows is tight.
Late minutes read against the assigned window, not a company start time. An agent on the evening window arriving at two in the afternoon is early. That same clock time on the morning window is five hours late, and the record says so without anyone doing the arithmetic.
Missed punches are routine on a busy floor: a queue at the terminal at handover, a headset already on before the scan. Each one becomes a manual punch request with an approval, so the correction sits on the record instead of in a team lead's chat thread.
Every agent has a device user on the terminal, mapped to a system user. Your mapping page counts how many are still unmapped, and a new cohort of twenty hires shows up there first, on their first morning.
The number contact centres argue about
Adherence. Whether the seat was filled when the window opened, and for how much of the window it stayed filled.
Attendance gives you the measured half. An on-time rate for the period, late minutes per agent, a late-arrivals-only filter, and the split of present, absent, partial and leave. Partial is the bucket your team leads read first: an agent who left the window two hours early lands there, not under absent.
Hours short against the period is the other side, reported per agent and as a working-hours histogram. A window that consistently runs twenty minutes short is obvious on the histogram and invisible in a monthly total.
Shrinkage gets forecast rather than discovered. Your leave view lists who is out today, what is pending and who is away in the next seven days. Your capacity grid lays agents against days with leave and holidays already removed, so the seats you count on are seats you actually have.
Attendance and hours both export as CSV over any date range for any group of people, which is where your monthly per-campaign report starts.
Which modules run here
Switched on
- Shifts
- Morning, evening and night as separate named windows, assigned per agent with history.
- Attendance
- On-time rate per window, late minutes, the partial bucket and the punch correction queue.
- Leave
- Quotas per type with approvals, plus today's on-leave list and the next seven days.
- Resource allocation
- Agents against days, banded under, partial, full and over, with leave already out.
- Clients
- One record per client account, with that client's local time beside yours.
- Projects
- Each campaign as a project with a billable flag, so hours land against the right account.
- Daily recap
- Hours per client account for the people who work across several, with approvals.
- Checklists
- A cohort's joining steps as one template, started against each new agent.
- Devices
- Headsets, handsets and thin clients by serial, condition and current holder.
- Notifications
- Recurring Slack reminders: the monthly attendance check, the weekly hours submission.
Left off
- Meetings
- A call queue produces no meeting for Fireflies to file.
- Project boards
- A Kanban board suits delivery work. A queue is staffed, not boarded.
Questions people ask
Can each agent be on a different shift window?
Yes. Shifts are assigned per person from a start date, with a duration and the assigner's name. Two agents on the same team can sit on different windows, and the assignment history keeps every move.
What happens when an agent misses the check-in window?
Their browser card routes them to a manual punch request, which HR reviews and approves. Your record then shows the correction and who approved it, rather than a silent gap.
What does it cost for a floor of two hundred agents?
$3 per person per month, with every module included and no tiers. A floor of two hundred pays the same per-person rate as a team of ten.