CAL_PROTOCOL_01 // PROTOCOL_SPEC
Intent Reliability Weighting Model
Not all intent is equal. We weight signals by source reliability and buying-stage alignment so activation targets decision-grade accounts—not noisy researchers.
The Intent Noise Problem
Most intent providers over-report “in-market” behavior. Without reliability weighting, your programmatic spend chases broad curiosity instead of active evaluation.
CRITICAL FAILURE MODES
- ✕ High-intent audiences filled with low-value researchers
- ✕ Retargeting pools inflate without conversion lift
- ✕ Sales ignores “intent” because it rarely matches real deals
Reliability Scoring Logic
SIGNAL INPUTS
Source Type (publisher, network, first-party)
Topic Granularity (category vs solution)
Historical Conversion Correlation
Spam / Bot Risk
Cross-Source Confirmation
→
WEIGHTED OUTPUT
Decision-Grade Intent Score
Logic: [Source Reliability] × [Topic Depth] × [Validation] − [Noise Risk]
PLATFORM ON (6sense/Demandbase + DSP)
We map weighted intent into activation cohorts and enforce minimum reliability thresholds before an account enters high-cost sequences.
Rule: IntentScore ≥ 75 AND ConfirmedSources ≥ 2 → Activate Tier1_Proof
PLATFORM OFF (CRM/Sheets + DSP Partner)
We deliver a weighting rubric and apply it to your raw intent feeds, producing an activation-ready list with documented thresholds.
Rule: Publisher intent = 1.0, Network intent = 0.6, Unknown = 0.3
Artifacts You Receive
- ✓Reliability Weighting Rubric (source multipliers + definitions)
- ✓Validation Rules (cross-source confirmation logic)
- ✓Noise Suppression Rules (spam/bot topic filtering)
- ✓Activation Thresholds by Tier (what score unlocks what spend)
Implementation Steps
WK 1
Audit current intent feeds vs closed-won to quantify false positives by source.
WK 2
Define reliability multipliers and confirmation rules; publish minimum activation thresholds.
WK 3
Deploy weighted audiences, validate conversion deltas, and tune thresholds monthly.