Your Recruiter Outreach Has a Denominator Problem

The dashboard that rewards the wrong behavior

Imagine two sourcing campaigns. Campaign A sends 100 emails, records 70 opens, eight replies and two qualified conversations. Campaign B sends 60, records only 35 opens, ten replies and six qualified conversations. The 'winning' campaign changes depending on whether you optimize opens, replies or genuine next steps. Open tracking can also be distorted by privacy protections and automated image fetching. It is not a dependable proxy for candidate intent.

Four denominators, four different questions

Record attempted, accepted for delivery, delivered/valid when verifiable, and unique people reached. A message can bounce; a follow-up can create multiple sends to one person; a reply can arrive after the reporting period. LinkedIn defines its InMail response rate with a specific time window and product scope, so it should not be compared casually with email reply rates. Define the unit, channel and observation window before benchmarking anything.

Separate interest from politeness

A reply of 'not now', a request to unsubscribe and a genuine conversation are all replies but imply different outcomes. Build mutually exclusive labels: interested and role-relevant, interested but not suitable, timing mismatch, declined, referral, opt-out, ambiguous. Do not use sentiment analysis to override a clear candidate statement. Measure positive qualified conversations per unique person reached, alongside complaints and opt-outs as guardrails.

The contact-data failure that looks like bad copy

If messages never arrive, rewriting the subject line is wasted effort. Check verified addresses, bounce classifications, domain authentication, sender reputation and throttling before A/B testing language. Greenhouse's deliverability documentation describes email verification and SPF/DKIM considerations. Distinguish a technical delivery problem from a value-proposition problem. Otherwise a recruiter may send more volume and worsen the very reputation issue causing low responses.

Run a small fair experiment instead of a spray-and-pray blast

Within one job family and seniority band, randomize eligible prospects to two honest messages. Hold sender, channel, role, follow-up cadence and observation window constant. Make one meaningful change, such as opening with a concrete role challenge instead of a generic compliment. Do not personalize with inferred private traits. Pre-register the primary metric: qualified conversations within 14 days per unique contact. Track bounce and opt-out rates separately.

The surprising advantage of a deliberate stop rule

After a fixed, respectful number of contacts, stop unless the candidate initiates engagement or a separate lawful basis supports future contact. More reminders can inflate total reply counts while damaging trust. A simple stop rule reduces duplicate messaging across recruiters and makes the denominator cleaner. Keep a shared contact history so two agency teams do not unknowingly approach the same person on consecutive days.

Worked example: the false 20% reply win

Synthetic campaign: 120 attempts, 15 bounces, 105 unique delivered contacts, 21 replies and seven qualified conversations. Reply rate per attempted message is 17.5%; per delivered contact it is 20.0%; qualified conversation rate is 6.7% per delivered contact. All three numbers can be accurate. None should be labeled 'conversion' without naming the denominator and outcome. Compare only cohorts with similar role attractiveness and time windows.

An operational playbook for Monday morning

Audit the last 50 distinct people approached for a single role. Check duplicate sends, delivery status, reply classifications, opt-outs and whether the opportunity matched the prospect's known skills. Recode five ambiguous replies with a second reviewer. Then fix the largest leakage point: address quality, proposition clarity or scheduling handoff. Preserve opt-out preferences and applicable privacy requirements. The metric is a diagnostic, never an individual recruiter's performance verdict.

Run the operational check

This calculator uses only synthetic or aggregate values and runs entirely in your browser. It cannot score an individual candidate.

Implementation checklist and guardrails

Assign a named owner, agree the data definitions, document exclusions and check a small representative sample before changing the workflow. Review outcomes by job family and process stage. Protect candidate privacy, provide accessible alternatives and do not automate adverse decisions from these calculators. Recheck the process after two cycles and retain the results that contradict the initial hypothesis.

Research basis and further reading

Twenty topic-specific and adjacent methodological references. Some describe regulations or assessment standards, others are vendor studies or operational definitions. They do not all validate the proposed framework; source scope and jurisdiction must be checked before applying it.

1. Greenhouse — Sourcing measurement
2. Greenhouse — Sourcing email deliverability
3. Greenhouse — Sourcing email templates
4. LinkedIn — InMail response rate methodology
5. Ashby — Source attribution definitions
6. Ashby — Candidates versus applications
7. Ashby — Recruiting operations benchmarks
8. Ashby — Job dashboard definitions
9. Ashby — Source taxonomy
10. Ashby — EMEA recruiter productivity
11. Greenhouse — 2024 state of job hunting
12. Greenhouse — 2025 AI in hiring report
13. CIPD — Resourcing and talent planning
14. CIPD — Workforce planning
15. CIPD — Recruitment overview
16. ICO — Recruitment and selection data
17. EDPB — Data protection principles
18. W3C — WCAG 2.2
19. W3C — Preliminary accessibility checks
20. CIPD — Workforce planning topics

The decision worth changing this week

Pick one measurable, reversible change, state what success and harm would look like, and publish the result with the denominator. Recruiting quality improves when evidence survives the handoff from recruiter to hiring manager.