Skip to content
LeadExample LeadExample B2B Conversion Benchmarks · Since 2019
Methodology · Vol. 04, Issue 02 Published January 2025 · Refreshed weekly

On the Record

How we measure what we publish.

A working disclosure of how LeadExample collects, anonymizes, qualifies, and verifies every B2B lead-conversion number that enters the public library — written for the marketing-ops reader who wants to see the gates a record clears before it shows up in a benchmark.

Author of record Lena Marquez, Co-founder
Verified by Dun & Bradstreet · RBAR customer panel
Cadence 312 quarterly updates · 0 skipped quarters

Verified inputs, on the record

  • 12,400+ Real B2B campaigns benchmarked since 2019
  • 1,820 Active institutional customers (D&B-verified, Jan 2025)
  • 312 Quarterly benchmark updates · 0 skipped quarters
  • 1.8M Anonymized qualified-lead records, refreshed weekly

Independently verified by Dun & Bradstreet institutional counts and the Recursive Benchmark Audit Round (RBAR) customer panel.

Chapter 01 · Definitions

Definitions: what we count as a converted B2B lead.

The single most important section on this page is also the shortest. If a number in our library does not satisfy the three conditions below, it does not enter the dataset and it is not published. Every benchmark you read — whether from a public report, a vendor case study, or a competitor deck — should be asked the same three questions before you treat it as comparable.

The unit of measure: LCR
Lead Conversion Rate (LCR) is the percentage of raw form-fill submissions that reach a defined qualified-pipeline stage within 90 days of capture. We score it on a single denominator across the library so every vertical reads against the same yardstick. LCR is the only lead-conversion metric we publish under the LeadExample brand.
What counts as "qualified"
A lead is qualified when it has cleared all three of: (a) a verified work email on a non-disposable domain, (b) a BANT or equivalent qualification score of 60+ from the customer's own scoring model, and (c) at least one human action — meeting booked, RFP opened, demo attended — within 90 days. MQL-only definitions are explicitly excluded; they inflate the rate by 2–4× and break cross-vertical comparability.
Cohorts and channel exclusions
We report LCR broken down by four cohort axes: industry (47 verticals), ICP tier (SMB / Mid / Ent), deal-size band, and acquisition channel (organic, paid search, paid social, outbound, partner, event). Affiliate traffic, incentivized clicks, and any campaign with a sample size under 400 form-fills per quarter are excluded from published medians.
Why 47 verticals instead of one number
A single B2B-wide LCR is statistically meaningless: a fintech demo request and a manufacturing RFQ behave nothing alike. We publish 47 vertical-specific micro-benchmarks because the lift you can ship against a reference depends on whether your reference matches your motion. The top-decile rate for verticalized SaaS demo funnels is roughly 2.6× the all-B2B median — that gap is the actionable number.

Chapter 02 · The pipeline

From raw submission to published benchmark: six gates a record must clear.

Walk a single anonymized campaign record through the six gates. Each gate is owned by a named role inside the LeadExample team, each produces an artifact that is filed alongside the record, and each can reject a submission from the published library. The whole pipeline takes a median of 11 days from raw form-fill to public weekly refresh.

  1. 01

    Capture

    The raw form-fill is ingested via the customer's webhook (Marketo, HubSpot, Salesforce, Pardot, or a direct REST push). At this stage we hold only the email, timestamp, source, and form schema — never the message body or any free-text fields. Capture is automated and runs every 15 minutes.

    Owner: Data Ingestion team

  2. 02

    Anonymize

    PII is hashed with a per-customer rotating salt, the customer's brand string is replaced with a stable opaque tenant ID, and all free-text fields are stripped. The anonymized record carries no email, no company name, no IP-derived geolocation above city, and no UTM parameter that could re-identify a campaign under < $50k/mo spend.

    Owner: Privacy Engineering

  3. 03

    Qualify

    The record is matched against the customer's own scoring model — never LeadExample's. We ingest the score and the qualifying-event timestamp directly from the customer's CRM; if the score is below 60 or the event timestamp is missing, the record exits the pipeline as unqualified and contributes to the denominators (raw form-fills) but never to the numerator (qualified leads).

    Owner: Customer Success Engineering

  4. 04

    Cohort

    The qualified record is bucketed on the four cohort axes — industry (47), ICP tier (3), deal-size band (5), and channel (6). A record can sit in up to 60 cohort cells; each cell holds its own minimum-sample rule. Cells with fewer than 400 qualified records in a rolling 90-day window are suppressed from the public report and replaced with a footnote.

    Owner: Research team

  5. 05

    Verify

    A 5% random sample of every customer's records is sent back through the Recursive Benchmark Audit Round (RBAR) — the originating customer's own ops team re-reconciles the lead against their CRM and certifies the LCR calculation. Customers flag discrepancies; flagged records are either corrected or excluded before publication. RBAR has been running continuously since Q3 2020.

    Owner: RBAR Panel · 19 participating customers in the audit pool

  6. 06

    Publish

    Records that survived gates 01–05 enter the LeadExample library on the next weekly refresh cycle (Tuesdays at 09:00 CT). Every published number is tied to a quarterly snapshot ID that links forward to the next refresh and backward to source records on request. The library has shipped every Tuesday for 312 consecutive quarters with no skipped cycles.

    Owner: Editorial · LeadExample, Inc.

The shape of the library

The shape of the library, as of January 2025.

A tonal pivot. The four numbers below are not marketing copy — they are the measured scale of the underlying dataset that every benchmark on this page is drawn from.

12,400 Real B2B campaigns benchmarked across SaaS, fintech, manufacturing, and professional services since 2019.
1.8M Anonymized qualified-lead records in the open dataset — the largest of its kind in B2B marketing.
47 Vertical-specific micro-benchmarks, each broken down by ICP tier, deal size, and acquisition channel.
312 Customer implementations verified through the Recursive Benchmark Audit Round (RBAR) panel.

Chapter 03 · Reader questions

Questions marketing-ops teams ask before adopting the data.

These four questions turn up in nearly every evaluation call we run with a VP of Marketing or Demand Gen lead. The answers below are quoted from the public RBAR audit pack — the same documentation we send under NDA during a procurement review.

How does RBAR verification actually work?

Every quarter, we select a 5% random sample from each customer's anonymized records and send the originating records back to the customer's own ops team — not to a third party. The customer re-reconciles the lead against their CRM, certifies the LCR arithmetic, and either confirms or flags the record. Flagged records are corrected or excluded before publication. The audit pack, including the sample-selection methodology and the discrepancy log from the most recent cycle, is available on request under NDA. RBAR has been continuous since Q3 2020 with no suspended cycles.

How does anonymization protect customer identity?

Every record carries no email, no company name, no free-text, no IP-derived geolocation above city, and no UTM parameter below a $50k/mo spend threshold. Each customer gets a stable opaque tenant ID and a per-customer rotating salt that is reissued quarterly. The hash function is one-way and salted, so a record cannot be traced back to a specific submission without access to both the salt and the customer's source — neither of which leaves LeadExample infrastructure. No raw record has ever been published, exported on request, or made available via API to non-customer accounts.

Why does weekly refresh matter?

B2B lead behavior shifts on the order of weeks, not quarters. A benchmark that was true in Q3 can drift by 4–8 percentage points by mid-Q4 once a major account changes its form length, switches ICP, or pauses paid channels. Weekly refresh gives the median a 90-day rolling window of data and lets the library reflect seasonal and campaign-level shifts inside a single quarter. Quarterly-only refreshes — the cadence most competitors ship — produce numbers that are 6–12 weeks stale by the time you read them.

How are D&B-institutional customer counts verified?

Once per quarter we submit a sealed list of active institutional customer DUNS numbers to Dun & Bradstreet for an independent count. D&B returns a reconciliation report flagging any DUNS that has been delisted, merged, or downgraded to a non-active status. The most recent reconciliation, dated January 2025, returned 1,820 active institutional customers. The D&B verification letter is included verbatim in the appendix of every quarterly benchmark report and is the single most cited credibility artifact on customer due-diligence calls.

Take the next step

See the library this methodology produces.

A weekly drop of the qualified-lead conversion rates, form-length splits, and channel-level performance figures behind the numbers on this page — delivered every Tuesday at 09:00 CT, 100% gated by your work email, no agency fluff.

We send one email per week. Unsubscribe at the bottom of any issue. No agency creative deck — only the underlying numbers and the templates behind them.