1. Decide what moves and what stays
A paper register holds two different things, and migration treats them differently:
- Identity and rules. Worker lists, sections, work codes, task norms, deduction slabs. This data moves into the system, checked line by line against the register.
- Historic muster marks. These stay in the books. The new system records from its go-live day onward. Retroactive data entry burns weeks of staff time and changes nothing going forward.
Making this split early keeps the migration short. Most estates that stall here are trying to re-enter history nobody will ever audit.
2. Move the identity data
Load workers with their permanent estate IDs and sections, then load the rules the field needs: tare, rainfall slabs, fine leaf settings, and task norms. Verify against the muster book section by section. Two checks catch most problems:
- Same name, different ID. Duplicate worker IDs across sections. Fix before enrollment, not after payroll.
- Rules that live in the clerk's head. Deductions applied differently by different sheds. Write down the actual slabs and make them explicit in the system.
3. Enroll and train
Enrollment and training follow the same practice as any rollout, but migration adds one rule: train supervisors on the manual fallback before the parallel run. When they know the app has a documented path for camera failures and unrecognized faces, resistance drops fast.
4. The parallel run
Run both systems for about a week. The comparison routine is simple:
- Mark muster on paper as usual.
- Record the same muster in the app.
- At end of day, the office compares worker counts and exceptions.
- Note the differences and their cause: enrollment quality, device placement, or an honest manual entry.
Differences should shrink every day. When a section's two records agree for a couple of days, it is ready to switch.
5. Switch, then close the books
Switch sections off paper one at a time. The register goes into the almirah the day the office reviews the app's record for that section without a difference. The first pay cycle on the new system is the real test, so keep support close until wages have run once from finalized app records.
After that, the paper register becomes what it always should have been: a record of last resort, not the system.