Introduction
social insurance, pension and withholding across several states, shifts and a mixed on-roll and off-roll workforce, without a five-day monthly scramble.
Where multi-site payroll actually breaks
Payroll for a single-location company is largely arithmetic. Payroll across four sites in three states is a data collection problem: different biometric devices, different shift patterns, different Professional Tax slabs, and a contractor workforce whose attendance nobody owns.
The computation is rarely the failure point. Getting complete, accurate attendance in on time is.
Attendance is the real dependency
If attendance arrives as CSV exports from four devices, someone has to normalise them, chase the site that has not sent theirs, and handle the regularisation requests that arrive by phone. That is three days gone before any calculation begins.
Capturing attendance directly — from the biometric device, a geofenced mobile punch, or a WhatsApp check-in — removes the collection step entirely. Regularisation becomes a request in the system with an approver and an audit trail rather than a message someone has to remember.
Statutory variation across states
Professional Tax slabs differ by state. ESI applicability depends on wage thresholds that change. PF has its own set of rules for international workers and for employees above the wage ceiling. None of this is difficult, but all of it is easy to get slightly wrong at scale.
The practical requirement is that the rules are maintained centrally and applied automatically, and that when a rate changes you are told rather than expected to notice.
The off-roll workforce
Contractor and gig workers are usually handled outside the payroll system entirely, which creates two problems: no visibility into principal-employer compliance exposure, and invoice verification that nobody can perform because attendance was never captured.
Bringing off-roll attendance into the same system does not make them employees. It makes their invoices checkable.
What to aim for
Payroll should be a review exercise. The variance report should be ready before you ask for it, and it should name the employees who caused the variance rather than just showing a number. If your team is still assembling data on the twenty-eighth, the system is doing too little.
In summary
None of this requires a transformation programme. It requires deciding where each piece of data is created, making sure it is only created once, and letting everything downstream read from that record instead of keeping its own copy.
Aditya Kulkarni, Head of Product at easyto.work.
