Case Study: Ghost Trainees and Unverified Placements in a Skill Development NGO

Year after year, the utilisation reports improved: 600 trained and 350 placed in Year 1, 900 and 580 in Year 2, 1,200 and 840 in Year 3. A 70% placement rate is exceptional for skill development in India, where public programmes commonly report far lower rates.

The CSR head asked to meet 20 placed candidates and their employers. The NGO produced 8 candidates at its own training centre; the other 12 were "at work and unavailable". The 8 struggled to describe their jobs, and two named the same employer but described different work locations.

Trainee verification

Deduplication. The database held 1,200 entries with Aadhaar numbers. 82 entries repeated an Aadhaar number under a different name or in a different batch, leaving 1,118 unique entries.

Attendance registers. Northrop examined 8 of 24 batch registers (400 names). For 140 names, every day was marked present in identical handwriting, consistent with bulk completion rather than daily marking. 85 names were marked present on dates when the centre was closed, as shown by the NGO's own published holiday notices and the building's security log.

Telephone verification. 200 numbers were randomly selected from the 1,118 unique entries and called directly, not through the NGO.

The placement report listed 45 employers. Northrop contacted the 20 to whom the NGO attributed the most placements (610 of the 840), visiting premises or calling HR directly.

• 7 employers confirmed hiring through the programme: 120 placements in total.

• 5 employers said the NGO had approached them, but they had hired no one.

• 4 employers had no record of any contact with the NGO.

• 4 employers could not be found at the listed address.

So 120 of the 610 placements claimed at these employers were confirmed (20%). Applying the same rate to the 230 placements at the other 25 employers suggests roughly 45 more. The best estimate of genuine placements is therefore about 165, not 840. Against about 380 genuine trainees, that is a placement rate of 32% to 43%: respectable, but nowhere near the 70% reported to the funder's board.

Employers confirmed monthly salaries of Rs 8,000 to Rs 10,000 for the 120 confirmed placements. The NGO had reported Rs 12,000 to Rs 15,000.

Of 50 offer letters examined, 15 were verified with the employer, 20 could not be verified by the named company's HR team, and 15 carried no employer identification and were signed by the NGO's own placement officer.

The Rs 1.4 crore was reported as: trainers Rs 50 lakh, materials and equipment Rs 35 lakh, trainee stipends Rs 25 lakh, placement support Rs 15 lakh, administration Rs 15 lakh. Using the NGO's bank statements:

• Stipends, Rs 25 lakh. Rs 12 lakh went to accounts whose holder names matched trainees. Rs 8 lakh was recorded as cash paid to trainees without bank accounts, backed by self-prepared receipts, of 30 such recipients called, 12 confirmed receipt, 8 denied it and 10 were unreachable. Rs 5 lakh went to accounts whose holder names matched no one on the trainee database.
• Placement support, Rs 15 lakh. Paid entirely in cash against self-prepared vouchers. No employer visit reports, interview schedules or follow-up records were produced.

The larger finding is not financial. The funder had reported 840 placements to its own board on the strength of the NGO's report. The evidence supported about 165.

The funder ended the arrangement and redesigned its skill development programme: biometric attendance instead of manual registers, placements confirmed directly with employers by the funder's own HR team, and independent quarterly verification before each tranche is released.

An outcome number you report to your board is only as good as the verification behind it. If your placement figures come solely from the NGO, you are not reporting placements. You are reporting claims.