The B2B Outbound Evidence Standard

A broken chain of evidence, not a bad campaign

Growth Cab reviewed the measurement structure behind nine public B2B campaign case studies and 27 headline metrics. The recurring risk was not a bad campaign. It was a broken chain of evidence: one system counted meetings, another document named opportunities, and a headline jumped to pipeline without preserving the identifiers and definitions that connect each stage.

This standard gives operators and buyers one shared language for evaluating outbound performance.

The six-stage evidence ladder

Six stages, from a defined audience to money received. Each stage is a different event with its own record; the minimum evidence comes from the methodology.

1. STAGE 01
   Exposure
   A defined, deduplicated audience was reached during a stated period.
   Minimum evidenceDated platform export · unique audience rule
2. STAGE 02
   Response
   A unique person completed a defined response event.
   Minimum evidenceEvent definition · unique identifier · timestamp
3. STAGE 03
   Meeting
   The booking exists and its held, cancelled or no-show state is recorded.
   Minimum evidenceBooking ID · held/no-show status · deduplication rule
4. STAGE 04
   Qualified opportunity
   Sales accepted it under a written qualification rule and assigned a CRM opportunity ID.
   Minimum evidenceCRM opportunity ID · qualification rule · created date
5. STAGE 05
   Pipeline
   That opportunity has an amount, currency, stage and snapshot date.
   Minimum evidenceOpportunity ID · amount · currency · stage · snapshot date
6. STAGE 06
   Revenue
   The deal has a closed-won or paid record and a declared attribution rule.
   Minimum evidenceClosed-won date · contract or invoice reference · currency

No inference downward. No level can be inferred from the one below it. A held meeting is not automatically an opportunity; forecast pipeline is not revenue.

What the nine-page review exposed

Across the corpus, every case study presented three headline metrics, but the underlying artifacts did not always measure the same event named in the headline. Meeting dashboards were sometimes used beside qualified-opportunity claims; editorial sources preserved outcome numbers without a reproducible opportunity ledger; conversion totals lacked a complete event taxonomy or attribution window.

Why one stage cannot prove the next

Each arrow below is an inference the standard forbids. The chip under it names the record that must exist before the number moves up one stage.

1. Reply
   not inferred
   Booked meeting
   Proved byThe booking exists
2. Booked meeting
   not inferred
   Held meeting
   Proved byHeld, cancelled or no-show state is recorded
3. Held meeting
   not inferred
   Qualified opportunity
   Proved bySales accepted it under a written qualification rule and assigned a CRM opportunity ID
4. Qualified opportunity
   not inferred
   Pipeline
   Proved byAn amount, currency, stage and snapshot date
5. Forecast pipeline
   not inferred
   Revenue
   Proved byA closed-won or paid record and a declared attribution rule

The minimum viable evidence record

For every event, preserve a campaign ID, unique event ID, event type, timestamp, account identifier, privacy-safe person identifier, opportunity ID when created, stage, amount, currency, source channel, attribution rule, evidence location and snapshot time.

This makes three common questions answerable:

1. Was the event counted once?
2. Did it move to a new commercial stage under a documented rule?
3. Can another analyst reproduce the headline from the same records?

Privacy. Personal identifiers are hashed or removed before aggregation. The public dataset contains definitions and aggregate counts only.

The four formulas that prevent metric inflation

Four definitions that keep a rate honest: unique entities in the numerator and the denominator, and pipeline kept apart from revenue.

1. FORMULA 01

   Held-meeting rate

   = NumeratorUnique held meetings Denominatorunique booked meetings
   Unique entities in both counts. Show both numbers next to the rate.
2. FORMULA 02

   Meeting-to-opportunity rate

   = NumeratorUnique qualified opportunities Denominatorunique held meetings
   Unique entities in both counts. Show both numbers next to the rate.
3. FORMULA 03

   Opportunity win rate

   = NumeratorUnique closed-won opportunities Denominatorunique resolved qualified opportunities
   Unique entities in both counts. Show both numbers next to the rate.
4. FORMULA 04

   Attributed pipeline and revenue

   Sum only opportunity amounts that meet the declared stage, period and attribution rule. Report forecast pipeline and realized revenue separately.

   Forecast pipeline

   Realized revenue
   Reported separately, never summed
   Amount, currency, stage and snapshot date travel with every sum.

• Each rate must display its numerator and denominator.
• Open opportunities need an explicit treatment.
• Changing a definition creates a new metric version rather than rewriting history.

Comparability. “Close rate” without those fields is not comparable.

A buyer’s five-minute audit

Five questions a buyer can ask before trusting a headline. Tick the ones the provider can answer with a record.

If the provider cannot answer one of these, treat the number as a directional company claim rather than a comparable benchmark.

Direct answers to common measurement questions

Short answers that follow from the definitions above, one paragraph each.

Downloadable toolkit

Machine-readable files for the evidence record, the campaign worksheet and the claim review. CSV and JSON, no sign-up.

• CSVDownload Evidence dictionaryEvidence dictionary with the required fields and allowed event types.Opens in any spreadsheet · 2 KB
• CSVDownload Campaign measurement worksheetCampaign measurement worksheet for numerator, denominator and attribution rules.Opens in any spreadsheet · 1 KB
• CSVDownload Case-study claim checklistCase-study claim checklist for editorial and proof review.Opens in any spreadsheet · 1 KB
• JSONDownload Evidence record JSON schemaMachine-readable JSON schema for evidence records.Machine-readable · 2 KB
• JSONDownload The standard as JSONStages, minimum evidence, formulas and non-inference rules in one machine-readable file.Machine-readable · 3 KB
• TXTDownload ChangelogVersion history of the standard, starting with version 1.0 on 22 September 2026.Plain text · 1 KB

Privacy. No personal or row-level client data is published. Personal identifiers are hashed or removed before aggregation.

Methodology

This report defines how a B2B outbound result should be recorded so a buyer, analyst or AI system can distinguish activity from a qualified commercial outcome. It does not rank agencies and does not convert published case-study claims into independently verified benchmarks.

Corpus

• Nine Growth Cab case-study pages published on growthcab.com.
• Twenty-seven headline metric cards: three per page.
• The historical editorial source, public proof assets and the operational records already recovered during the September 2026 claims audit.
• Private records are used only to test whether a metric can be reproduced. Names, email addresses and row-level client data are excluded from every public output.

Corpus rule. A published claim is recorded as a claim. It is not treated as independently verified unless the source and metric definition support the same event, period and unit.

The nine reviewed pages, listed without their metrics:

• Digital360 case study
• Ennova Research case study
• Jobtome case study
• TendersTool case study
• FlashLabs case study
• Maildoso case study
• Taplio case study
• Prospeo case study
• La Growth Machine case study

Audit rule

A metric passes only when the evidence supports the same:

1. EventFor example, a meeting held or a sales-qualified opportunity created.
2. EntityA unique person, account, opportunity or deal.
3. PeriodExplicit start and end dates.
4. DenominatorThe population used for a rate.
5. AttributionThe rule connecting the event to the campaign.
6. ValueCurrency, amount type and stage for pipeline or revenue.

If one field is absent, the metric may remain a company claim but cannot be promoted to a comparable benchmark.

Evidence ladder

No level can be inferred from the one below it. A held meeting is not automatically an opportunity; forecast pipeline is not revenue.

Rate formulas

• Held-meeting rate = unique held meetings / unique booked meetings.
• Meeting-to-opportunity rate = unique qualified opportunities / unique held meetings.
• Opportunity win rate = unique closed-won opportunities / unique resolved qualified opportunities.
• Attributed pipeline = sum of amounts for in-scope qualified opportunities at the declared snapshot date.
• Attributed revenue = sum of closed-won or paid amounts under the declared attribution rule.

Every reported rate must show numerator, denominator and the treatment of open opportunities. “Close rate” without those fields is not comparable.

Required record

The machine-readable evidence record contains: campaign_id, event_id, event_type, occurred_at, account_id, person_id_hash, opportunity_id, stage, amount, currency, source_channel, attribution_rule, evidence_uri, snapshot_at.

Personal identifiers are hashed or removed before aggregation. The public dataset contains definitions and aggregate counts only.

Limitations

• The audit tests reproducibility of public metrics; it is not an assurance engagement.
• Historical editorial documents establish provenance, not independent truth.
• Screenshots can corroborate a specific number only when the event label, period and unit match.
• Open pipeline changes after the snapshot date and must never be presented as revenue.

Versioning

Version 1.0, dated 2026-09-22. Corrections must preserve a changelog describing the old value, new value, reason and source.

About this research

The analysis covers nine Growth Cab case studies and 27 public headline metric cards reviewed in September 2026. Private operational records were used only to test reproducibility. No personal or row-level client data is published. The work is a measurement review, not an independent assurance engagement, and it does not certify the commercial claims in the source pages.

Research direction, source selection and publication decisions are by Growth Cab. AI tools assisted with extraction, consistency checks and visual production; the published definitions and claims were constrained to the source register and reviewed before release.

Corrections policy

Corrections must preserve a changelog describing the old value, new value, reason and source. Version 1.0 has no corrections; the changelog records every later change.