Signal-to-Noise Filter
Cut through intent data noise to find who's actually buying
Intent data is expensive and noisy. This skill aggregates signals from multiple sources, applies weighted scoring, and delivers a daily action list of accounts showing real buying behavior — not just content consumption.
INGREDIENTS
PROMPT
Create a skill called "Signal-to-Noise Filter". I'll configure my intent signal sources and weights. Aggregate signals across all sources for my target accounts. Apply weighted scoring: [pricing page visit = 10, demo request = 10, G2 comparison = 8, case study download = 5, blog visit = 1, third-party intent = 3]. Generate a daily prioritized list of accounts above my threshold score. For each account, show: composite score, individual signals that fired, trend direction (heating/cooling), and a "why now" summary I can use in outreach. Suppress chronic false positives that surge but never engage.
How It Works
Instead of treating every "surging" account equally, this skill layers multiple
signal types — website visits, content downloads, review site activity, job
postings, and third-party intent — into a composite score. Only accounts crossing
your threshold make the daily list.
What You Get
- Multi-signal aggregation: first-party + third-party intent combined
- Weighted scoring: pricing page visits outweigh blog reads
- Daily prioritized action list with "why now" context per account
- Signal trend tracking: is this account heating up or cooling down?
- False positive filtering: suppress accounts that always "surge" but never buy
Setup Steps
- Define your signal sources (website analytics, content platform, review sites, etc.)
- Set signal weights (e.g., pricing page = 10, blog visit = 1, G2 comparison = 8)
- Set your action threshold (e.g., composite score > 25)
- Configure daily delivery (Slack, email, or Telegram)
Tips
- First-party signals from your own site and content are usually more predictive than third-party intent alone
- Pricing page + competitor comparison page in the same week = highest-intent combo
- Review weekly: if your list is too long, raise the threshold
- Track conversion rates from the list to validate scoring weights