CDP Audience Taxonomy

200+ Segment Framework

ConversionOS defines a structured audience taxonomy that translates ML scores into actionable CDP segments. This framework has been deployed with 200+ active segments across ActionIQ, Segment, and Treasure Data.

Taxonomy Structure

Audiences are organized in a hierarchical taxonomy:

Level 1: Intent Category (Acquisition / Retention / Winback)
  Level 2: Score Tier (High / Medium / Low)
    Level 3: Behavioral Modifier (Channel preference, lifecycle stage, value tier)

Core Segment Framework

Acquisition Segments

Segment ID	Name	Definition	Activation
ACQ_HIGH_INTENT	High Intent Prospects	Propensity ≥ 0.7 + product page visits in 7d	Google Ads, Meta (value: $240)
ACQ_MED_NURTURE	Medium Nurture	Propensity 0.4-0.7 + email engaged	Email drip, retargeting (value: $50)
ACQ_LOW_AWARE	Low Awareness	Propensity 0.15-0.4, limited engagement	Broad awareness (value: $10)
ACQ_EXCLUDE	Suppression	Propensity < 0.15 or existing customer	Suppress from all paid (value: $1)

Retention Segments

Segment ID	Name	Definition	Activation
RET_LOYAL_HIGH	High-Value Loyal	Churn risk < 0.2 + LTV top quartile	Loyalty rewards, upsell
RET_AT_RISK_HIGH	High-Value At Risk	Churn risk > 0.6 + LTV top quartile	Proactive retention, personal outreach
RET_AT_RISK_MED	Medium At Risk	Churn risk > 0.6 + LTV mid-range	Automated retention journey
RET_DECLINING	Declining Engagement	Engagement recency ratio < 0.3	Re-engagement campaign

Composite Segments (examples from 200+)

Segment	Combination Logic	Size (typical)
PREMIUM_ACQ_DIGITAL	High propensity + digital channel preference + high credit tier	2-5% of prospects
WINBACK_RECENT_HIGH	Churned < 90 days + was high-LTV + winback propensity > 0.5	5-10% of churned
UPSELL_ENGAGED_MID	Active 30d + mid-tier product + engagement score > 0.7	10-15% of customers

CDP Configuration

ActionIQ Setup

{
  "audience_sync": {
    "source": "bigquery.conversionos.scored_prospects",
    "schedule": "daily_6am_est",
    "match_key": "hashed_email",
    "destinations": [
      {
        "platform": "google_ads",
        "list_type": "customer_match",
        "segment_mapping": {
          "ACQ_HIGH_INTENT": "conv_value_240",
          "ACQ_MED_NURTURE": "conv_value_50",
          "ACQ_LOW_AWARE": "conv_value_10"
        }
      },
      {
        "platform": "sfmc",
        "sync_type": "data_extension",
        "trigger_journeys": true
      }
    ]
  }
}

Segment Hygiene

• Mutual exclusivity: A user should appear in at most one segment per category
• Refresh cadence: Daily for acquisition, real-time for retention triggers
• Size monitoring: Alert if any segment grows/shrinks >20% day-over-day
• Overlap analysis: Monthly audit of cross-segment overlap rates