Direct answer
To judge a sourcing tool, divide its fully loaded cost for a pilot period by the number of unique evidence-fit leads it produced: people a recruiter reviewed against the requirements, found to fit, and could not have found through the sources already in use. Fully loaded means licence or credits plus the recruiter time spent using the tool and reviewing its results.
Cost per hire still matters at budget level. The ANSI/SHRM standard defines it as total internal plus external recruiting costs divided by total hires for the same period [S1][S2]. But hires are too few and too slow to judge one tool in a 60- or 90-day pilot. Cost per unique evidence-fit lead is the earlier signal that tells you whether the tool adds anything at all.
Start from cost per hire, then go one level down
The ANSI/SHRM cost-per-hire standard (06001.2012) is the common reference for recruiting cost: cost per hire = (total internal recruiting costs + total external recruiting costs) / total number of hires, with every input taken from the same period [S1][S2]. Secondary descriptions list external costs such as agency fees, job boards, assessments and recruiting software, and internal costs such as recruiting team compensation and referral bonuses [S3]. Industry coverage at the time described variants of the standard that treat hiring-manager time differently, so state which version you use [S2].
A sourcing tool is one line inside "external costs". Cost per hire can show that recruiting got cheaper or more expensive overall, but it cannot isolate one tool. Several sources contribute to most hires, attribution is contested, and a pilot rarely runs long enough for its hires to close.
The metric ladder for a tool pilot
| Metric | Formula | What it can mislead you about |
|---|---|---|
| Cost per profile or record | Pilot cost / profiles returned | Rewards volume; says nothing about fit or overlap |
| Cost per reviewed lead | Pilot cost / leads a recruiter actually opened and assessed | Still counts poor fits |
| Cost per evidence-fit lead | Pilot cost / reviewed leads that met the requirement standard | Counts people you already had from other sources |
| Cost per unique evidence-fit lead | Pilot cost / evidence-fit leads not surfaced by your existing sources | The best pilot signal, but sensitive to how carefully duplicates are matched |
| Cost per hire influenced | Pilot cost / hires where the tool surfaced the hired person first | Too few events in a short pilot; attribution judgment calls |
The fourth row uses the same idea as Unique Contribution Rate: the share of a source’s evidence-fit leads that no other tested source surfaced. Cost per unique evidence-fit lead is that unique count with a cost attached.
Count the full cost, including time
- Licence or subscription fees for the pilot period, pro-rated if the term is longer.
- Credits or usage charges actually consumed, plus credits bought but unused if they expire.
- Setup and integration time, at a loaded hourly cost your finance team agrees with.
- Recruiter time spent searching in the tool and reviewing its results. This is usually the largest cost and the one most often left out.
- Compliance and review work, for example a privacy or security review the tool required.
Worked example with illustrative numbers
All figures below are invented to show the arithmetic. They are not benchmarks, typical prices or results from any real tool.
| Line | Illustrative value | Note |
|---|---|---|
| Pilot licence and credits | $3,000 | Hypothetical quote for 90 days |
| Recruiter time using Tool X | 20 hours × $50 loaded = $1,000 | Hypothetical loaded hourly cost |
| Fully loaded pilot cost | $4,000 | Sum of the two lines above |
| Profiles returned | 400 | |
| Leads reviewed | 120 | |
| Evidence-fit leads | 30 | Met the same written standard used for every source |
| Unique evidence-fit leads | 12 | Not found by the existing stack in the same requisitions |
| Unique contribution rate | 12 / 30 = 40% | |
| Cost per evidence-fit lead | $4,000 / 30 = $133.33 | |
| Cost per unique evidence-fit lead | $4,000 / 12 = $333.33 | The decision metric |
Now test sensitivity. If careful duplicate matching shows only 3 of the 30 were truly unique, cost per unique evidence-fit lead becomes $4,000 / 3 = $1,333.33, four times higher, and the decision may flip. This is why the duplicate check deserves more care than any other step in the pilot.
Compare the result to a threshold your team sets before the pilot, for example what you would otherwise spend to produce one additional qualified, unique lead through another channel. Write the threshold down in advance so the decision is not fitted to the result.
Design the pilot so the numbers mean something
- Pick 2 to 4 live or recently closed requisitions that represent the work the tool is meant for.
- Write the evidence-fit standard per requisition before searching, and apply it identically to every source.
- Search the existing stack and the new tool for the same requisitions in the same window.
- Match duplicates across sources by person, not by name string, and record unresolved matches as unresolved.
- Log recruiter time per source as you go, not from memory at the end.
- Report all ladder metrics, with the unique count as the headline and the sensitivity case next to it.
Sourcing tool pilot ROI worksheet
One row per source per requisition, with columns for every cost and ladder count. The CSV includes the illustrative Tool X rows from this guide, clearly labelled, so you can check the formulas before replacing them with your own numbers.
Download the CSV worksheet (plain text, opens in any spreadsheet tool; contains no candidate data).
| Column | What to record |
|---|---|
source | Tool or channel name |
requisition_ref | Your requisition reference |
pilot_days | Length of the pilot window |
licence_and_credit_cost | From your quote or invoice, pro-rated |
recruiter_hours | Logged during the pilot |
loaded_hourly_cost | Agreed with finance |
fully_loaded_cost | licence_and_credit_cost + recruiter_hours × loaded_hourly_cost + other costs |
profiles_returned | Raw results |
leads_reviewed | Opened and assessed |
evidence_fit_leads | Met the written standard |
unique_evidence_fit_leads | Not surfaced by other sources |
cost_per_unique_evidence_fit_lead | fully_loaded_cost / unique_evidence_fit_leads |
data_label | REAL or ILLUSTRATIVE |
Related guides
Source ledger
The cost-per-hire formula comes from secondary descriptions of the ANSI/SHRM standard, reached through search-result summaries; the standard’s own text was not reviewed. Everything else on this page is method and illustrative arithmetic. A person must re-check the sources before publication.
| ID | Source | Claim it supports | Accessed / method | Human re-check |
|---|---|---|---|---|
| S1 | Precision Machined Products Association: New cost per hire standard: 50 pages of American National Standard | ANSI/SHRM 06001.2012 cost per hire = (internal + external recruiting costs) / total hires | 2026-10-25 · WebSearch summary; page not opened | Secondary source; confirm against the ANSI/SHRM standard text |
| S2 | ERE: SHRM task force close to official standard for cost per hire | The standard effort described internal and comparable variants that treat hiring-manager time differently, and a recruiting cost ratio | 2026-10-25 · WebSearch summary; page not opened | Pre-publication coverage; confirm which variants appear in the final standard |
| S3 | Workable: Recruiting costs FAQ: budget and cost per hire | Typical examples of internal and external recruiting costs | 2026-10-25 · WebSearch summary; page not opened | Vendor explainer; use only for examples of cost categories |