How to stop ghost workers on a construction site

Ghost workers — names that draw pay without a real person behind them — thrive on one thing: a payroll that can’t be checked against physical reality. If you suspect your site has them, you don’t need to accuse anyone or launch a dramatic investigation. You need a way to compare the names you’re paying against the bodies that actually showed up. Here’s how to do that, step by step, and how to close the gap so the ghosts can’t come back.

Step 1: Run a surprise headcount

Pick an ordinary working day and physically count the people on site, by role, without warning the supervisors in advance. Then compare that count to how many people are on the payroll for that day. A gap doesn’t prove fraud on its own — people are legitimately absent — but a large, persistent gap between “paid” and “present” is the first sign that your records and your reality have drifted apart.

Step 2: Reconcile the payroll against real identities

Go down the payroll and confirm that every name maps to a real, identifiable person — one national ID, one phone number, one human being. This is exactly how large payroll audits uncover ghosts: when Ghana’s government reviewed its payroll in 2025, it found more than 67,000 names that failed this test, many of them people who had left long ago but were never removed. The same pattern hides on construction sites: workers who finished months ago but are still “on the book,” and names with no verifiable identity behind them at all.

Step 3: Look for the classic patterns

Ghost workers tend to leave fingerprints. Watch for several workers sharing one phone number or one payment destination, names that are always present but never seen at the gate, attendance that’s suspiciously perfect (real people miss days), and crews whose headcount never matches the supervisor’s invoice. None of these is proof by itself, but together they tell you where to look closely.

Step 4: Separate the record from the people who benefit

Here’s the root cause behind most ghost workers: the person who records attendance is often the person who benefits from padding it. As long as a supervisor can write the register and also draw from it, no amount of auditing fixes the incentive. The durable solution isn’t to trust harder — it’s to take the pen out of the equation, so attendance is captured by a system, not authored by a person.

Step 5: Move to scan-based attendance

This is where a QR-card system like Twese closes the gap for good. Each worker gets a durable QR card and scans in and out at the gate; every scan is timestamped and stored off-site. The rule is simple and unbreakable: no scan, no record, no pay. A ghost can’t scan a card he doesn’t hold at a gate he never visits. A worker who left last month stops scanning last month and drops off the payroll automatically. The list on payday becomes who the system actually saw — not who someone remembered, or wanted, to be there.

Step 6: Make it self-cleaning

The quiet advantage of scan-based attendance is that it stays clean without effort. On paper, removing a departed worker is a task someone has to remember; forget it, and the ghost lingers. With scanning, absence is the default — you only appear on payday if you physically scanned in — so departed and invented workers fall off on their own. You’re not fighting the same battle every month; you’ve removed the ground the ghosts stood on.

The bottom line

You stop ghost workers by making attendance something that can be checked against physical reality, and then by making sure the record is captured by a system rather than written by hand. Do the headcount, reconcile the identities, and then take the pen away. Twese is built to do exactly that on Rwandan construction sites — turning “who worked this month?” from an opinion into a fact.

Where these numbers come from

Ghana’s 2025 payroll audit finding of 67,311 ghost names was reported by the Ghana Audit Service and announced in the government’s 2025 mid-year budget review. Figures are cited as reported and rounded for readability.