Measures worth tracking
Pick a small set and define each one in the agreement so that neither side argues about the arithmetic later.
| Measure | What it shows | How to count |
|---|---|---|
| Shortlist acceptance | Whether CVs match the brief | Candidates you approve for interview ÷ CVs sent |
| Trade test pass rate | Screening quality before the test | Passed ÷ tested |
| Selection to arrival | Process speed | Days from selection to landing, per worker |
| Drop-outs after selection | Candidate commitment | Withdrawals ÷ selected |
| Leavers in first 90 days | Fit and honesty of the offer | Early exits ÷ arrivals |
| Fee complaints | Fair recruitment | Confirmed complaints per batch |
Avoid measures that reward the wrong thing
Counting CVs encourages bulk forwarding. Rewarding speed alone can push an agency to skip reference checks. A pure cost measure invites hidden charges to candidates. Balance each speed or cost measure with a quality one so the agency cannot improve one by damaging another.
Be careful with targets the agency does not control. If your visa approvals are slow, selection-to-arrival time will look poor through no fault of the recruiter. Split the timeline into the steps each side owns.
Collect the data simply
A shared tracker with one row per candidate is enough: dates for CV, interview, test, selection, medical, visa, emigration clearance, travel and arrival, plus the outcome after 90 days. Ask the agency to update its columns and your site team to update theirs.
Ask arriving workers privately, through a short form in their own language, whether they paid anyone for the job. That one question gives you a fair-recruitment indicator that no agency report can.
Review and act
Hold a short review after each batch and a fuller one each quarter. Look for patterns rather than single failures: one trade with a low pass rate, one source region with early leavers, one step that always slips. Agree one or two changes and check them at the next review.
- Share the numbers before the meeting.
- Separate employer delays from agency delays.
- Record the agreed actions with owners and dates.
A worked example
Imagine a batch of thirty warehouse workers. The agency sent ninety CVs, you approved forty-five for interview, thirty-six passed the practical test, thirty were selected, two withdrew and twenty-eight arrived. Within ninety days, one left for family reasons and two left because night shifts had not been explained.
The numbers point to one clear fix: the job preview. Screening quality was reasonable and drop-outs were low, but the two early leavers who did not expect night work show that the briefing needs to be firmer. That is a conversation to have before the next batch, not a reason to change agency.
Linking KPIs to money
Some employers tie part of the fee to results, such as paying a balance after the probation period. Keep it simple and lawful, and never allow an incentive that could push costs onto the worker. Replacement guarantees often work better than penalties because they fix the problem rather than punish it.
Questions employers ask
How many KPIs are enough?
Four to six. More than that and nobody reads the report.
What is a good trade test pass rate?
It depends on the trade and how strict the test is. Track your own baseline and look for changes rather than chasing an external benchmark.
Should early leavers count against the agency?
Only where the cause was recruitment, such as a misleading offer or a skills mismatch. Exits caused by site conditions are the employer’s to fix.
Who should own the KPI report?
One person on the employer side, usually in HR or procurement, who collects data from both the agency and the site.
General information, reviewed 2026-11-21. Rules change and differ by country, so confirm the current position with the authority concerned or your legal adviser before acting.
Keep reading
Planning to hire from India?
Send the roles, numbers, destination and start date. We reply with a written scope, timeline and cost.
Send your requirement