Attribution Integrity Audit

Id

RLR-33

Link

6

Link Name

Attribution + Decision Intelligence

Role

audit

Loop Step

Diagnose

Purpose

Determine whether the business can reliably trace an opportunity from its original source, through every stage of the Revenue Chain, to realized revenue and — where available — gross profit. The audit inspects real records and systems, measures how much of the chain is actually connectable, and returns an explicit attribution-confidence level. It does not fabricate benchmarks and does not treat low-confidence attribution as precise measurement.

Governing Principle

You cannot manage what you cannot trace. Link 6 exists to answer 'where did this revenue come from, and what happened to the opportunity at every stage?' A number the business cannot trace source-to-revenue is not a measured result — it is an estimate at best and a guess at worst. Attribution CONFIDENCE (how trustworthy the trace is) is never the same as attribution ACCURACY, and an Unknown source is never silently converted to zero or to a disqualified source.

Progression

• CAPTURE
• CONNECT
• MEASURE
• INTERPRET
• DECIDE
• ACT
• RE-MEASURE

Critical Distinctions

Original vs latest source

ORIGINAL SOURCE (the first touch that created the opportunity) is not LATEST SOURCE (the most recent touch before an action). Both are legitimate; the audit fails any system that silently overwrites original source with latest activity, because that erases the very trace Link 6 depends on.

Measured vs modeled

A MEASURED value is reconciled from real records (CRM, telephony, billing). A MODELED value is produced by the Revenue Leak Calculator or another planning model. The audit keeps them in separate columns and never presents a modeled estimate as a measured actual.

Confidence vs accuracy

Attribution CONFIDENCE describes how completely and consistently the chain connects; it is not proof the attribution is ACCURATE. High field completeness with broken cross-system IDs can still be low-confidence. The audit reports confidence, not certainty.

Unknown vs zero

An UNKNOWN source, revenue, or outcome is recorded as Unknown and routed to a measurement plan. Unknown is never treated as 0, as 'no revenue', or as a disqualified source.

Reporting vs decision

This is not reporting for its own sake. The audit's job is to establish whether the data is trustworthy enough to drive a cut/keep/fix/scale decision. Data that is too weak to decide on is an explicit output, not a hidden gap.

Handoff From Link5

Note

Link 6 consumes Link 5 (Booking + Handoff) output and the outputs of Links 1-4 that flow through it. It does NOT recreate Link 5 booking, confirmation, no-show, or handoff logic.

Consumes

original_source; qualification_classification; booked_status; booked_date; appointment_type; assigned_owner; show_no_show_status; reschedule_status; appointment_outcome; next_stage_outcome; retained_won_status_where_available; revenue_where_available

Traceability Chain

• Original Source
• Inquiry
• Contact
• Qualification
• Follow-Up
• Appointment
• Show
• Sale / Retained Client
• Revenue
• Gross Profit (where available)

Systems Inspected

• forms
• inbound calls
• missed calls
• SMS
• website chat
• paid media
• organic traffic
• referrals
• LSAs / marketplaces where relevant
• social channels
• offline leads
• CRM
• calendars
• sales/intake systems
• revenue/billing data where available

Data To Collect

Scope

A representative sample of recent opportunities that reached at least the inquiry stage, drawn across sources and outcomes (won, lost, no-show, unknown) so the audit sees the whole chain — not only closed-won records. Join each record across every system that holds part of its trace.

Integrity Checks

Metrics

Note

Compute only where data permits, and report the record coverage beside every rate so thin data is never presented as a confident result.

Attribution Confidence Levels

Benchmark Policy

Approved Benchmarks Available

false

Rule

Do not fabricate benchmark thresholds for attribution completeness, unknown-source rate, or confidence. Report the business's own measured coverage. Where an approved ShiFt benchmark exists, label it explicitly and never present it as this business's audited result.

Measured Vs Modeled

Rule

Every value the audit produces is tagged MEASURED (reconciled from real records) or MODELED (from the Revenue Leak Calculator or another planning model). Modeled estimates never appear in a column labeled as actual results, and the two are never summed together.

Missing Data Rules

Count unknowns explicitly and report them beside every rate. A record with no resolvable source is an Unknown routed to a measurement plan — never converted to zero revenue or to a disqualified source. A won record with missing revenue is a data gap, not proof of a low-value source.

Compliance Guardrail

The audit inspects attribution and revenue data under the business's approved data-handling, privacy, retention, and consent policy. It does not instruct the operator to store more personal or financial data than necessary, does not bypass provider restrictions or opt-outs, and treats suppressed/consent-limited records as legitimate constraints, not coverage failures. Jurisdiction- and industry-specific requirements are applied by the operator.

Feeds Into

Note

The audit exposes clean, measurable inputs to the rest of the Link 6 system without duplicating them.

Closed Loop

Note

Attribution findings are the entry point of the closed loop: once the chain is trustworthy enough to interpret, decisions (RLR-37) feed source allocation and process repair back to Link 1 (Demand + Source) and to whichever link is leaking, so demand spend follows evidence instead of guesswork. Link 6 does not terminate at a dashboard.

What To Measure

Whether opportunities can be traced original-source → inquiry → contact → qualification → follow-up → appointment → show → sale → revenue (and gross profit where available), measured as source completeness, source-to-appointment and source-to-sale traceability, revenue-to-source traceability, unknown-source and duplicate rates, field completeness, and an overall attribution-confidence level — reported with per-record coverage.

Data Required

A representative, cross-outcome sample of recent opportunities with the per-record fields above, joined across CRM, telephony/call-tracking, forms, scheduler, and revenue/billing systems.

How To Interpret

Read every rate beside its coverage and its confidence level. High field completeness with broken cross-system IDs is still low confidence. A low source-to-revenue rate driven by unreconciled offline sales is a data leak, not a bad source. Never let an Unknown source become a zero, and never treat a low-confidence trace as a measured result.

Action That Follows

Repair the highest-impact attribution gaps by implementing the Revenue Attribution Field Map (RLR-34), normalizing definitions against the Revenue Chain KPI Dictionary (RLR-35), and standing up the Revenue Chain Dashboard (RLR-36). Prioritize with the Revenue Leak Priority Matrix (RLR-06); gaps with INSUFFICIENT_DATA get a measurement plan before any recovered-revenue claim.

Owner

A named revenue-operations or data owner runs the audit; each identified attribution gap gets a named owner at repair time via the Repair Plan (RLR-39) and the Weekly Operating Review (RLR-38).

How To Implement

Pull the cross-outcome sample, join it across every system that holds part of the trace, populate the field table, run each integrity check, compute the metrics with coverage, assign an overall attribution-confidence level, and record findings on the Revenue Leak Repair Plan (RLR-39).

How To Test

Trace a handful of known won deals end-to-end by hand — from the billing record back to the original source — and confirm the systems reproduce the same trace. Where the manual trace and the systems disagree, the confidence level is overstated and must be lowered.

When To Remeasure

Re-run the audit 14 and 30 days after attribution repairs (RLR-34/RLR-09) are installed, using the same sample definition so the before/after comparison is valid.

How To Know It Improved

The leak is repaired only when re-measured source completeness, source-to-sale and revenue traceability, unknown-source rate, and attribution-confidence level improve on real records — not because a dashboard or field map was created.