Pipeline Conversion Rate

Two kinds of drop: candidates leave or you screen them out

Not every lost candidate is the same. Separate stage losses into:

• Candidate-driven loss: people who don't complete the screen, don't book, don't attend or decline the offer. These are tracked as candidate drop-off rate and usually point to speed, friction or a weak offer.
• Employer-driven loss: people screened out for not meeting requirements. A low pass rate here may be fine, or a sign that the ad attracts the wrong people or the criteria are stricter than the job needs.

In the example above, the 55% completion rate on the online screen and the 74% attendance rate are candidate-driven. The 38% interview-to-offer rate is employer-driven. They call for completely different fixes.

Finding the leak

Lay the stages out in order, as in the recruitment funnel, and look for the stage with the biggest gap between where you are and where you could reasonably be. In high-volume hiring, the usual suspects are:

• Application completion: long forms on mobile lose people before you ever see them.
• Screen completion: an assessment sent two days after applying converts far worse than one offered immediately.
• Booking: if candidates must wait for a recruiter to call, many never book. Self-scheduling right after the screen typically raises this.
• Attendance: confirmations and reminders by text the day before and an hour before help most.
• Offer acceptance: pay below local market or slow offers.

Fix the biggest candidate-driven leak first. It is usually cheaper than buying more applicants.

Worked example: what a fix is worth

Using the example funnel, suppose the team cuts the online screen from 25 minutes to 10 and sends it by text immediately, and completion rises from 55% to 70%. With every later stage unchanged, starters go from 25 to about 32 from the same 400 applicants, a 27% increase with no extra advertising. Repeating the math for each stage shows where an improvement is worth the most, which is how to prioritize. Funnel math like this is the same calculation behind yield ratio planning, applied to diagnosis rather than forecasting.

Measuring it properly

• Use cohorts. Follow a group of candidates who entered a stage in the same period, and wait until most have finished it before calculating. Mixing candidates still in progress understates conversion.
• Segment by role, location, source and device. Mobile and desktop applicants often convert very differently.
• Track time in stage alongside conversion. Long waits are often the cause of low conversion.
• Check fairness. Compare pass rates for each stage across demographic groups where you collect that data. A stage with a much lower pass rate for one group needs review for adverse impact under the four-fifths rule and Title VII.

Common mistakes

• Only measuring applicant-to-hire, which hides where the loss happens.
• Treating a low pass rate at a screening stage as a problem to fix by loosening criteria, when the real issue is the ad attracting the wrong applicants.
• Changing several stages at once, so nobody knows which change worked.