Public benchmark · updated in real time

The one benchmark in this category you can actually rerun.

Every competitor claims 99%. None publishes the methodology. We publish ours — the exact SQL, the exact sample size, the exact formulas — and split precision into three separately-defined numbers instead of collapsing them into one convenient headline.

Strict verified
 
found / total
Reframable
 
(found + protected) / total
Returned subset
 
found / (found + protected + catch_all)
Coverage
 
(found + protected + catch_all) / (total - no_mx)

 

found
1,240
protected
565
catch_all
590
not_found
715
no_mx
250
deferred
20
error
20

The three precision numbers, defined.

Same underlying data, three published metrics. Pick which one honestly matches your workflow.

Strict verified

found / total

Email confirmed via SMTP RCPT-TO handshake with the destination server. Most conservative. Matches the definition independent auditors use in category comparisons.

Reframable precision

(found + protected) / total

Includes M365/Proofpoint/Mimecast domains that accept our probe but don't confirm existence. Matches how every enterprise vendor implicitly reports.

Returned-subset precision

found / (found + protected + catch_all)

Precision on the subset we chose to return. This is the same shape most category headline claims are computed on — precision only against the emails a vendor chose to return, not against the full input list.

Methodology.

  • Sample: 3,400 real LinkedIn Sales Navigator leads uploaded through the production dashboard by a paying customer. No synthetic data, no vendor-friendly seed lists, no cherry-picking.
  • Verification path: SMTP RCPT-TO probe on the destination mail server across our distributed verification fleet. Jittered Gmail-throttle backoff. Reason codes returned honestly: found, protected, catch_all,not_found, no_mx, deferred, error.
  • Query: SELECT status, COUNT(*) FROM leads WHERE status NOT IN (NULL, 'pending') GROUP BY status
  • Refresh: live. Every uploaded lead updates these numbers within minutes.
  • Endpoint: api.coalstoke.com/v1/benchmark/summary — public, unauthenticated, hit it with curl.

Compared to the 99% club.

Every published category headline (typically 95–99%) is subset-precision on the emails a vendor chose to return. Independent whole-list audits by third-party researchers put effective precision far lower than the headline once the denominator includes the full input list — which is why we publish three separately-defined numbers instead of one.

VendorClaimIndependent auditCoverage
Category headline (typical)95–99%subset-only
Category same-methodology (typical)varies by list
CoalstokeSegmented (three numbers, no headline)Self, live, aboveAbove

Precise vendor-by-vendor numbers vary week to week and by input sample. Rather than freeze fabricated audit numbers into a comparison table, we publish our own three metrics live and invite direct head-to-head runs. Bring your own list and API keys — the numbers we return are reproducible.

Head-to-head — 15 randomly-sampled leads.

Same 15 leads, three engines. Pulled at random from the 3,400-lead D1 sample — no cherry-picking. Sample size is small so treat this as directional, not a leaderboard. The point is showing where each engine wins and where a combined waterfall outperforms any single provider.

EngineVerifiedEmail returnedCost / verified
Coalstoke SMTP only8 / 15 (53%)13 / 15 (87%)$0.004
Hunter API only8 / 15 (53%)10 / 15 (67%)~$0.05 rack
Icypeas API only (projected)~7 / 15~9 / 15~$0.02 rack
Coalstoke + Hunter BYOK fallback12 / 15 (80%)13 / 15 (87%)~$0.03
+ Icypeas as 3rd fallback (projected)~13 / 1514 / 15~$0.04

"Verified" = engine confirmed the email is deliverable via SMTP handshake or provider-side verification signal. "Email returned" includes probable-guess emails on catch-all and protected domains that no single-engine test could confirm. Icypeas row is projected from their published methodology pending API access.

Where we win, where we lose.

Same benchmark, sliced by mail infrastructure. If your list is 90% Google Workspace, Hunter alone gets you most of the way. If it's a mixed enterprise list with heavy catch-all / Proofpoint / niche self-hosted, the combined waterfall pulls ahead by 25+ points.

Mail infraBest engineWhy
Google WorkspaceCoalstoke SMTP + HunterSMTP RCPT works cleanly, Hunter has strong pattern data — nearly tied at 90%+.
Microsoft 365 (protected)Combined waterfallO365 accepts anything → we guess pattern, BYOK confirms via provider records.
Proofpoint / MimecastBYOK fallback (Hunter/Icypeas)SMTP probe blocked at edge — provider records are the only signal.
Catch-all domainsCombined waterfallSMTP says "accept" on everything — BYOK is the tiebreaker with real deliverability data.
Niche SME / self-hostedCoalstoke SMTP aloneSmall vendors aren't in most finders' scrape corpus — SMTP is the only shot.
Deep enterprise (Fortune 500)Combined + human reviewNo engine hits 90%+ — SMTP probes rate-limited, pattern data is stale. Set expectations at 60–70%.

Selling verified emails yourself?

Wholesale API pricing starts at $0.005 per verified at enterprise volume — undercuts every published API rack in the category while keeping your retail margin healthy. Attach behind your brand or take the full white-label stack at verify.yourbrand.com.

Starter

$0.015

< 10K / mo

Growth

$0.010

10K – 50K / mo

Scale

$0.007

50K – 250K / mo

Enterprise

$0.005

250K+ / mo

Full wholesale terms + intake form

Want to run this benchmark against your own list? Sign up for 50 free credits — drop a CSV, get the four numbers back for your own data. That's the entire product.

CoalstokeIndependent · Verification infrastructure
Made in Missouri
Essential cookies only · why