Straight answer: contract labour is where attendance records break, because enrolment takes longer than the worker stays. The result is a paper list maintained by the contractor — the party with the least incentive for it to be accurate — while the principal employer may still carry obligations for those workers.
This is not a hypothetical gap. On most industrial sites we visit, permanent staff are on a biometric system and contract labour is on a register in someone's pocket.
Why the enrolment economics fail
Consider what enrolling one worker traditionally costs: a walk to the office, a supervisor's time, three or four fingerprint captures, a record created in a system, sometimes a card printed.
Call it ten minutes of two people's time. For a permanent employee staying three years, trivially worth it. For a mason on a two-week job, it is never going to happen — so it doesn't, and the register absorbs them.
Multiply by a gang of thirty arriving on a Monday and the reason for the paper list becomes obvious. Nobody is being lazy; the process simply costs more than the tenure justifies.
What the gap actually costs you
Headcount you cannot verify. The contractor bills for thirty; you have no independent record of how many walked through the gate. This is the most common commercial leak on Indian sites and it is almost never measured.
Compliance exposure. Where obligations attach to the principal employer, "the contractor maintained that record" is a weak position if the contractor's record turns out to be inadequate.
No incident reconstruction. When something goes wrong, you cannot establish who was on site. Not for blame — for basic accountability and, where relevant, for insurance.
Safety blind spot. In an evacuation, a headcount you cannot verify is a real problem, not a paperwork one.
What has to change
The variable is enrolment cost, not record-keeping discipline. Any system requiring per-worker admin time will lose to contract churn indefinitely.
Enrolment has to be fast enough to do at the gate on arrival: a guided face scan from a phone, under a minute per person, no office visit, no card, no supervisor tied up for an hour. At that cost, enrolling a thirty-person gang becomes a normal part of induction rather than a project.
Two properties then follow that matter specifically here. The record is photo-backed, so contractor-billed headcount can be reconciled against who actually passed the gate. And contract-worker records can be held separately from permanent staff, with retention ending when the contract does — which is also the right posture under the DPDP Act's storage-limitation principle.
We build this, so weigh the framing accordingly. The underlying diagnosis — that enrolment cost, not technology, is what excludes contract labour — holds regardless of vendor.
The check worth running this month
Take last month's contractor invoice and the gate record for the same period.
If you cannot compare them, that is the finding. If you can and they differ, that is a more expensive one.