Determining Which Signals Deserve Confidence
Rules for weighting source reliability, recency, frequency, relevance, and corroborating evidence.
The signal-confidence model gives marketing, sales, revenue operations, and customer teams one shared method for evaluating evidence quality before an account signal creates action.
The Fairway Signal Confidence Index™
Assess the trustworthiness and interpretive strength of each signal before it changes account treatment.
Source Reliability
The authority, consistency, and known limitations of the source.
Identity Confidence
The degree to which the person, account, and role are known.
Recency
How recently the signal occurred relative to its useful window.
Freq. & Progression
Whether the signal is isolated, escalating, or moving to deeper behavior.
Commercial Relevance
How closely the evidence connects to the account’s fit and stage.
Corroboration
The degree to which independent sources support the same conclusion.
Why a Signal-Confidence Model Matters
Without explicit confidence rules, teams may overreact to anonymous or stale activity, underweight direct buyer evidence, and treat repeated events from the same source as independent corroboration.
Common Symptoms Solved
- × Every signal source contributes the same amount to priority
- × Anonymous activity is treated like known role engagement
- × Old behavior remains active indefinitely
- × Repeated low-value actions inflate confidence
- × Third-party topic surges route directly to sales
- × Teams cannot explain why an alert was generated
Strategic Decisions Enabled
- ✓ Which sources are reliable enough for automated action
- ✓ How known and anonymous activity should differ
- ✓ How quickly signals should decay
- ✓ When frequency strengthens confidence vs. creating duplication
- ✓ Which evidence combinations create corroboration
- ✓ Which confidence levels permit direct engagement
What Is Included
Confidence Dimensions
Major factors used to evaluate the quality and meaning of a signal.
Source Reliability
Ranking first-party, third-party, public, seller, and customer data.
Identity Confidence
Distinguishing known people, resolved accounts, and anonymous visitors.
Recency and Decay
Time windows determining how long signals remain commercially meaningful.
Corroboration Logic
Rules for combining independent evidence into stronger interpretation.
Governance
Ownership, testing, override, change, and performance-review model.
View the Sample Signal-Confidence Model
Explore the representative pages that evaluate AeroGrid’s signal sources, identity, recency, progression, relevance, corroboration, confidence, and permitted actions.
- Page 01 Signal-Confidence Summary
- Page 02 Confidence Arch. & Weights
- Page 03 Source Reliability Hierarchy
- Page 04 Identity Confidence
- Page 05 Recency and Decay
- Page 06 Frequency & Progression
- Page 07 Commercial Relevance
- Page 08 Corroboration Model
- Page 09 Calculation Examples
- Page 10 Caps & Overrides
- Page 11 Confidence-to-Action Matrix
- Page 12 Account Confidence Profile
- Page 13 Data Requirements
- Page 14 Confidence Review Workflow
- Page 15 Model Measurement
- Page 16 Governance
AeroGrid Signal-Confidence Model
A governed method for evaluating source reliability, identity, recency, frequency, relevance, and corroboration.
Signal-Confidence Summary & Architecture
The Model in One View
- Source Reliability Max 25
- Identity Confidence Max 15
- Recency Max 15
- Frequency & Progression Max 15
- Commercial Relevance Max 20
- Corroboration Max 10
Action: Direct qual/progression.
Action: Coordinate activation.
Action: Research/Nurture.
Action: Observe/Suppress.
- • Confidence is not Priority.
- • Direct Evidence Outweighs Inference.
- • Independent Evidence Matters.
- • Recent Signals Matter More.
- • Relevance Determines Meaning.
- • Confidence Must Be Explainable.
Source Reliability & Identity
Who Created the Signal and From Where
| Source Type | Reliability | Example | Default Use |
|---|---|---|---|
| Buyer-Confirmed | Very High | Discovery confirmation, Technical review | Direct qual/progression |
| Customer/Contractual | Very High | Contract date, Verified value | Lifecycle triggers |
| Known 1st-Party Activity | High | Known pricing visit, Webinar attendance | Person-level intent classification |
| Official Disclosure | High | Leadership announcement, Expansion plan | Context and corroboration |
| Trusted Ext. Research | Moderate | Industry database, Technology detection | Research signal or support |
| Anonymous 1st-Party | Low to Mod | Resolved account visit (unknown person) | Account monitoring |
| 3rd-Party Intent | Low to Mod | Topic surge, External research activity | Research signal or support |
Known Person, Confirmed Account, Confirmed Buying Role.
Anonymous Person, Resolved Account.
Known Person, Outdated Employment. (Invalid until corrected).
Recency, Frequency & Progression
How Time and Repetition Affect Confidence
- • Direct Request: Full 0-7 days | Reduced 8-30 days
- • Pricing/Tech Eval: Full 0-14 days | Reduced 15-45 days
- • 3rd-Party Surge: Full 0-14 days | Expire unless corroborated
- • Leadership Change: Full 0-90 days | Reduced 91-180 days
Rule: Base decay on event date, not ingestion date.
- • Single Low-Value: Minimal confidence (Awareness)
- • Repeated Identical: Limited confidence (No progression)
- • Topic Progression: Mod/Strong (Moving to evaluation)
- • Multi-Role Progression: Very Strong (Cross-functional eval)
Rule: Progression and breadth are more informative than repetition.
Commercial Relevance & Corroboration
Why it Matters & Independent Support
Fit doesn't change whether signal occurred, but if it is useful.
| Level | Example | Effect |
|---|---|---|
| None | One anonymous topic surge. | No increase. |
| Repeated Same-Source | Several anonymous visits from resolved account. | Limited increase. |
| Cross-Source | Official announcement + known facilities engagement. | Meaningful increase. |
| Cross-Category | New facilities leader + pricing activity. | Strong increase. |
| Multi-Role Support | Facilities, finance, and IT engage evaluation content. | Very strong increase. |
Calculation Examples & Caps
Applying the Weighting Rules
Evidence: Confirmed identity/role. Three related 1st-party interactions. Progression from problem to approach. No current external trigger.
Action: Account-owner review & role-specific engagement.
Evidence: Relevant topic increase. Resolved to account. No known person. No corroborating event.
Action: Research and monitor.
- • Unknown account identity → Maximum Low Confidence
- • Resolved account but unknown person/role → Maximum Moderate Confidence
- • 3rd-Party intent without corroboration → Maximum Moderate Confidence
Action Matrix & Account Profile
What Each Confidence Level Permits
| Confidence | Allowed Actions | Restricted Actions |
|---|---|---|
| Low | Retain for context, Aggregate to trends, Low-cost nurture. | No direct sales alert, No auto-opp creation. |
| Moderate | Create research task, Trigger-specific nurture, Monitor. | No automatic exec outreach, No assumption of eval. |
| High | Notify acct owner, Coord marketing/sales, Validate readiness. | Do not assume purchase authority. |
| Confirmed | Direct qualification, Specialist engagement, Action planning. | Progression still requires buyer evidence. |
Signals: New VP Facilities (High), Modernization Initiative (High), Facilities/Finance Content Progression (High), Architecture Request (Confirmed).
Account Confidence: Confirmed. Independent org, ops, behavioral, and tech evidence support the same conclusion.
Data Requirements & Review Workflow
Operationalizing Confidence
- •
signal_id(Unique identifier) - •
signal_type(Controlled picklist) - •
source_reliability(Verified...Unknown) - •
identity_confidence(Known Person/Role...Unknown) - •
commercial_relevance(High...None) - •
confidence_level(Confirmed, High, Mod, Low)
- Ingest the signal
- Validate the source
- Resolve identity
- Evaluate relevance
- Apply recency & progression
- Find corroborating evidence
- Assign confidence & route action
Measurement & Governance
Maintaining Trust in the Model
- • False-positive rate
- • False-negative review
- • Precision by source and role
- • State-to-outcome correlation
- • Cadence: Weekly high-confidence review, quarterly performance analysis.
- • Triggers: High false-positive rate, weak correlation to opps.
- • Rule: Model is shared infrastructure. Keep it transparent, testable, and versioned.
Note: AeroGrid Systems and all associated confidence weights, thresholds, source rankings, and treatments are fictional. This sample demonstrates the structure of a Fairway engagement.
How Teams Use the Signal-Confidence Model
The model gives marketing, sales, revenue operations, customer success, and leadership one shared standard for determining which signals can support commercial action.
Product Marketing
Define signal relevance, buying-stage meaning, evidence standards, and confidence rules.
Marketing Operations
Evaluate sources, apply decay, deduplicate events, and manage confidence-based audiences.
Sales Development
Prioritize research and outreach according to signal quality rather than activity volume.
Sales
Understand what happened, how reliable it is, why it matters, and what still needs validation.
Revenue Operations
Implement confidence calculations, caps, routing, overrides, history, and reporting.
Leadership
Evaluate whether signal investment is improving account timing and opportunity quality.
Signal-Confidence Model Review Checklist
Is the Evidence Strong Enough to Support the Action?
Explore Related Artifacts
Intent Definition Framework
Defines the fit, interest, intent, and readiness states that the confidence model evaluates.
Trigger Prioritization Model
Combines signal confidence with account fit, value, and urgency to determine commercial priority.
Trigger Event Framework
Return to the complete signal, activation, routing, and optimization system.