Recruiting Funnel Analyzer

About the Recruiting Funnel Analyzer

A recruiting funnel shows how many applications reach each stage of hiring — screen, interview, offer, hire — and how many pass from each stage to the next. Choose the export of your applicant tracking system (one row per application, CSV or Excel), check the columns it found, and the tool works out the funnel and its pass-through rates, the days spent in each stage (median and 90th percentile), time to hire and time to fill, the offer acceptance rate, source of hire and 90-day retention by source, broken down by job, department, source and recruiter.

Two planning aids use the same file: a hiring planner works back from a number of hires to the applications, screens and interviews they need at your own pass-through rates, and the cost per hire divides the period’s external and internal recruiting costs by its hires. When the export holds self-identified sex or race/ethnicity, the tool also shows the selection rate of each group at each step and its impact ratio against the most selected group, with the four-fifths flag — for your own monitoring, not as a bias audit.

Nothing is uploaded: the file is read in your browser. The free preview shows the funnel and the time to hire of your own file in full, and the first rows of the other tables; a Premium pass downloads the full workbook, the PDF report and the funnel chart.

Common uses

• Finding the stage where most candidates drop out, or wait longest.
• Comparing sources: which ones bring applications, which bring hires, and which hires stay.
• Planning how many applications a hiring target needs, and how long it will take.
• Monthly or quarterly recruiting reports for managers, as a workbook or a PDF.
• Monitoring selection rates of self-identified groups at each step, to see where a closer look is needed.

What is measured, and how

• Reached a stage: the application got at least that far — its furthest stage, or a later stage with a date, or a hire. Pass-through = reached this stage ÷ reached the stage before.
• Days in each stage: from the date a stage was entered to the date the next one was, for applications with both dates; the median and the 90th percentile (nearest rank: the smallest value with at least 90% of the values at or below it).
• Time to hire: application date → offer accepted (or the hire stage’s date). Time to fill: the job’s opening date → offer accepted.
• Offer acceptance = offers accepted ÷ offers made. Source of hire: the hires by source.
• 90-day retention by source: of the hires who started at least 90 days before the report date, those still there 90 days after their start.
• Cost per hire = (external recruiting costs + internal recruiting costs) ÷ hires in the period.
• Planner: the hires you need divided back through each stage’s pass-through rate, rounded up at every stage; the weeks are the median stage times one after another.

Selection rates by group, and the four-fifths rule

The selection rate of a group at a step is the share of the people considered at that step (who reached its first stage) who were selected (reached its second). The impact ratio divides a group’s selection rate by the rate of the most selected group. Under the US Uniform Guidelines on Employee Selection Procedures, a selection rate under four-fifths (80%) of the highest group’s rate “will generally be regarded” as evidence of adverse impact, while smaller differences can still matter and differences based on small numbers may not be significant (29 CFR 1607.4(D)). So each group also gets Fisher’s exact test against the most selected group, and groups with fewer than 30 people (or fewer than 5 selected or not selected) are marked as small.

Sex, race/ethnicity and their intersections are calculated separately; answers such as “Prefer not to say” are counted as unknown and left out of the ratios; and categories under 2% of the data are marked (an independent auditor may leave them out) — as New York City’s rules on bias audits of automated employment decision tools lay out (6 RCNY §§ 5-300, 5-301). This is internal monitoring, not an independent bias audit: where a law requires an audit, an independent auditor does it.

Getting the export right

• One row per application. A candidate who applied to two jobs has two rows.
• Furthest stage reached, not the current status: a column that says only “Rejected” cannot tell how far the application got. Where the export has a date for each stage, choose those columns: the tool counts the furthest stage with a date, which is usually the truest record.
• Outcomes (hired, rejected, withdrawn, offer declined) are read from a status column; a hire also counts as reaching the last stage.
• Dates in ISO form (2026-03-04), day first (04/03/2026) or month first (03/04/2026) — choose the order in the settings — and Excel dates are read.
• The Applicant Tracking Board keeps a history of every move: its JSON export or its encrypted backup gives the stage dates without any column matching.

Limitations

• The figures are only as good as the export: stages skipped in the system, or dates entered late, change the stage times. The data checks list what was left out and why.
• Selection rates by group are for internal monitoring: they are not a bias audit, legal advice or proof that a process is fair or unfair. Small groups give unstable ratios.
• The tool reads one sheet of a workbook (the first) and up to 100,000 applications.
• It does not connect to an applicant tracking system: export the file from it first.

Privacy

On this device; its files download once

Your export is read in your browser and never uploaded. The column matches for each file layout, the date order and your organisation’s name are remembered in this browser; the applications are not. A backup’s passphrase is used only to open it and is not kept.