\nCompliance & Safety<\/strong><\/p>\nThis method is for legitimate recruiting outreach only. Always respect candidate privacy, opt-out requests, and local data laws. Heartbeat does not provide medical advice or legal counsel.<\/p>\n<\/blockquote>\n
<\/span>Framework: The Accuracy Stack: Identity \u2192 Channel Validity \u2192 Answerability<\/span><\/h2>\nTeams argue about \u201caccuracy\u201d because they mix three different checks. In recruiting ops, you need all three\u2014measured separately\u2014so you can fix the right failure mode.<\/p>\n
\n- Identity<\/strong>: the record belongs to the right person (correct person-to-credential-to-employer match).<\/li>\n
- Channel Validity<\/strong>: the phone number connects to a live line; the email address delivers.<\/li>\n
- Answerability<\/strong>: a human answers the connected call; a delivered email earns a reply.<\/li>\n<\/ul>\n
In this article, \u201caccuracy\u201d is an observed, per-attempt outcome by channel (per 100 dials and per 100 sent emails). Engagement is measured separately so you don\u2019t blame data for a messaging problem (or vice versa).<\/p>\n
The trade-off is\u2026 you can buy \u201cmore records\u201d or you can buy \u201cmore usable attempts.\u201d Recruiting throughput is driven by usable attempts.<\/p>\n
<\/span>Step-by-step method<\/span><\/h2>\n<\/span>Step 1: Use channel-specific definitions (don\u2019t use one blended \u201caccuracy\u201d number)<\/span><\/h3>\nUse these definitions in your scorecard so everyone measures the same thing.<\/p>\n
\n- contact data accuracy definition<\/strong>: The percent of outreach attempts where (1) identity is correct and (2) the chosen channel works as intended for that attempt. Always state the channel and denominator (per 100 dials or per 100 sent emails).<\/li>\n
- Identity accuracy (ops definition)<\/strong>: Per 100 connected calls, the share that reach the intended person (not a wrong person). A practical proxy is Wrong-person rate = wrong-person connects \/ connected calls<\/strong> (per 100 connected calls).<\/li>\n
- mobile accuracy definition<\/strong>: Per 100 dials to a \u201cmobile\u201d field, the share that connects to the intended person\u2019s mobile line (not disconnected, wrong person, or business main). This is channel validity; answerability is separate.<\/li>\n
- email accuracy definition<\/strong>: Per 100 emails sent to an address, the share that is delivered (not bounced). This is channel validity; replies are separate.<\/li>\n
- deliverability definition<\/strong>: Deliverability Rate = delivered emails \/ sent emails<\/strong> (per 100 sent emails).<\/li>\n<\/ul>\n
Define the related metrics so you don\u2019t \u201cfix\u201d the wrong layer:<\/p>\n
\n- Connect Rate = connected calls \/ total dials<\/strong> (per 100 dials). \u201cConnected\u201d means the call reached a live line (human, voicemail, IVR, or gatekeeper), not a failed attempt.<\/li>\n
- Answer Rate = human answers \/ connected calls<\/strong> (per 100 connected calls).<\/li>\n
- Bounce Rate = bounced emails \/ sent emails<\/strong> (per 100 sent emails).<\/li>\n
- Reply Rate = replies \/ delivered emails<\/strong> (per 100 delivered emails).<\/li>\n<\/ul>\n
<\/span>Step 2: Instrument \u201cper 100 attempts\u201d in your ATS\/CRM<\/span><\/h3>\nIf you don\u2019t log attempts, you\u2019ll end up debating anecdotes (\u201cthis data is bad\u201d) instead of fixing a measurable bottleneck.<\/p>\n
Measure this by\u2026<\/strong> logging every dial and email as an attempt, then calculating each metric per 100 attempts for a defined time window and segment (specialty, geography, source, campaign).<\/p>\nWorked example (fill in your numbers; do not guess):<\/strong><\/p>\n\n- Per 100 dials: __ connected calls; per 100 connected calls: __ human answers; __ wrong-person connects<\/li>\n
- Per 100 sent emails: __ delivered; __ bounces; per 100 delivered emails: __ replies<\/li>\n<\/ul>\n
<\/span>Step 3: Separate identity errors from channel errors<\/span><\/h3>\nWhen a recruiter says \u201cbad data,\u201d it usually means one of these:<\/p>\n
\n- Identity mismatch<\/strong>: wrong person, outdated employer, duplicate profiles merged incorrectly.<\/li>\n
- Phone channel failure<\/strong>: disconnected, wrong number, business main, call routing tree that never reaches the candidate.<\/li>\n
- Email channel failure<\/strong>: hard bounce, domain rejects, mailbox full, spam filtering.<\/li>\n
- Answerability failure<\/strong>: connects but no human answers; delivered but no reply.<\/li>\n<\/ul>\n
Identity problems require record-level remediation. Channel problems require verification\/refresh and suppression. Answerability problems require better timing, sequencing, and role-based messaging.<\/p>\n
<\/span>Step 4: Treat recency as a first-class field<\/span><\/h3>\nRecency<\/strong> is how recently the contact channel was observed as working. In recruiting, recency is what keeps \u201caccurate last quarter\u201d from becoming \u201cdead this week.\u201d Put a date on it.<\/p>\n\n- Store last_verified_phone_date<\/strong> and last_verified_email_date<\/strong> (or equivalent) per record.<\/li>\n
- Store verification_method<\/strong> (observed outreach outcome vs. validation tool).<\/li>\n
- Store source<\/strong> and source_date<\/strong> so you can compare decay by source.<\/li>\n<\/ul>\n
<\/span>Step 5: Decide what \u201cgood enough\u201d means for your workflow<\/span><\/h3>\n\u201cAccurate\u201d depends on what you\u2019re trying to do:<\/p>\n
\n- High-urgency outreach<\/strong>: prioritize phone channel validity and connect rate to compress time-to-first-conversation.<\/li>\n
- Pipeline building<\/strong>: prioritize email deliverability and reply rate to scale touches without burning call blocks.<\/li>\n
- Ops QA<\/strong>: prioritize identity accuracy and recency to prevent wasted recruiter hours and compliance risk.<\/li>\n<\/ul>\n
Heartbeat.ai is built for this reality: you\u2019re not buying a spreadsheet\u2014you\u2019re buying a workflow you can audit and improve, including ranked mobile numbers by answer probability<\/em> when you need to prioritize dials.<\/p>\n<\/span>Diagnostic Table:<\/span><\/h2>\nUse this to diagnose what \u201caccuracy\u201d problem you actually have. Copy\/paste into a QA sheet.<\/p>\n
![\"54195\"](\"http:\/\/hc.heartbeat.ai\/wp-content\/webp-express\/webp-images\/uploads\/2026\/02\/what-is-contact-data-accuracy-e4abf162.png.webp\")
<\/p>\n