Offline attendance needs a precise definition. A worker’s phone may be offline while a shared kiosk is online. That is different from a site with no connectivity at all. Establish which devices need a connection before choosing a check-in method.
Map connectivity at the point of work
Check the actual gate, warehouse floor, farm building, or temporary worksite where people arrive. Identify who has a phone, whether location services are available, and whether a kiosk can maintain a connection. Test weak connectivity as well as a complete outage.
Radin supports mobile GPS, printed QR, rotating kiosk QR, and an offline signed credential shown on a worker’s phone and scanned by an online kiosk. The last method does not require the worker’s phone to connect, but the kiosk remains an online attendance point.
Preserve the evidence behind each event
The time shown on a device is not automatically trustworthy. A clock can be wrong, a request can arrive late, or the same action can be retried. Ask how the system identifies the worker, records the capture method, and distinguishes the original event from a correction.
Walk through a delayed arrival, a repeated scan, a forgotten checkout, and a shift that crosses midnight. Check which time zone and attendance rules apply. A record that looks plausible on a dashboard still needs enough context to explain the resulting hours.
Design the exception and recovery process
Agree what supervisors should do when a kiosk loses connectivity or a worker cannot present a credential. Document the fallback, the person responsible for reviewing it, and how an approved correction becomes part of the attendance history.
Use the pilot to follow each exception through to payroll preparation. Radin keeps attendance and operational approvals connected, but the right setup depends on your devices and worksite. Demonstrate the exact outage scenario you need; do not assume every offline method supports every outage.
Questions to take into a demo
- Which device must remain online for each check-in method?
- How are duplicate events and incorrect device times handled?
- Who reviews missing punches and manual corrections?
- Can the original event and approved correction both be explained?