Skip to main content

Audience Insights & Segment Intelligence

The Audience Insights module in DriveMetaData provides deep diagnostic intelligence for saved customer segments. Rather than treating an audience as an opaque list of IDs, Insights reveals who your users are, how they behave, where they drop off, and how viable the cohort is for third-party ad activation.

1. Module Overview & Business Outcomes​

Audience Insights helps growth marketers, analysts, and campaign leads make data-backed decisions prior to committing advertising budget:

  • Pre-Flight Campaign QA: Confirm whether your audience has sufficient reachability and purchase intent before syncing to ad platforms.
  • Discovering High-Affinity Sub-Cohorts: Uncover unexpected demographic clusters (e.g., discovering 80% of cart abandoners originate from two specific cities).
  • Detecting Audience Fatigue: Identify segments experiencing high ad exposure or recency decay.
  • AI-Guided Decision Making: Receive prescriptive "Next Best Actions" generated directly from cohort characteristics.

2. Prerequisites, Entitlements & RBAC​

RequirementDescription
Audience StateThe audience must be Saved or Estimated with a valid ID .
Viewer RoleInspect existing insight cards, size trends, and diagnostic tabs.
Editor / Admin RoleTrigger insight recalculations, accept AI recommended actions, and execute activation handoffs.
Data FreshnessBehavioral and commerce insights require that raw telemetry events have synced within the last 24 hours.

3. Product Workspace Anatomy​

The Insights screen is organized into high-impact diagnostic panels:

audience insights

4. Canonical Step-by-Step Walkthrough​

Step 1: Navigate to Insights​

  • From the Audience Builder, click the More actions (⋮) menu in the header and choose View insights.
  • Alternatively, navigate directly via the route: /insights/[audienceId].

Step 2: Review the Operating Context Strip​

At the top of the page, review four core indicators:

  • Confidence: High, Medium, or Low statistical confidence in the computed cohort metrics.
  • Definition Version: Confirms which rule version (v1, v2) was used for evaluation.
  • Execution Duration: Real-time analytical query time against ClickHouse (typically 150ms–450ms).
  • Reachability Rate: Percentage of profiles possessing verified identifiers eligible for third-party sync.

Step 3: Evaluate AI Summary Cards & Next Best Action​

  1. The AI Audience Summary highlights significant positive drivers and operational risks.
  2. Review the highlighted Next Best Action card (e.g., "Trigger WhatsApp Cart Recovery Campaign").
  3. Click the recommended action buttons to route directly to campaign creation or destination syncing.

Step 4: Analyze Historical Size Trend​

  1. Inspect the Audience Size Trend chart displaying the last 6 execution runs.
  2. Confirm whether audience membership is growing, stable, or decaying over time.
  3. Review the Current vs. Previous comparison metrics (+6.3%) to detect sudden anomalies caused by seasonal shifts or tracking drop-outs.

Step 5: Diagnose Drop-Off via the Filter Impact Funnel​

  1. Scroll to the Filter Impact panel.
  2. Trace how each condition card in your audience definition reduced the baseline customer population.
  3. Review the Biggest Drop-off card to pinpoint which specific exclusion or inclusion rule removed the largest percentage of users.

Step 6: Explore the 7 Diagnostic Breakdown Tabs​

TabKey VisualizationsWhat to Look For
DemographicsTop Countries, Top Cities, Device Platform Donut, Publisher mixGeographic clusters and mobile vs. desktop skew.
BehaviorPurchase Intent (Purchasers %, Avg. Order Value, Event Depth), Recency Risk, Activation GuardrailsConversion readiness and churn risk before campaign launch.
Traffic & AcquisitionInbound Source Mix, Publisher Concentration, Top Inbound CampaignsDominant acquisition channels feeding this audience.
EngagementEvent Frequency, Session Depth, Activity HeatmapUser activity volume across weekdays and hours.
CommerceAverage Order Value, Repeat Purchase Ratio, Product Category AffinityMerchandising opportunities and basket size trends.
TechnographicsOperating System (Android vs. iOS), Browser, App VersionTechnical compatibility for deep-linking and push notifications.
Custom EventsCustom Telemetry Trigger Frequency, Feature AdoptionTracking-specific milestones and workflow completion.

5. Operational Guardrails & Data Governance​

  • Cohort Sample Ceiling: If an audience contains fewer than 50 qualifying users, the insight engine flags the analysis as Low Confidence to prevent marketing teams from generalizing on statistically insignificant sample sizes.
  • Reachability Threshold: If the Reachability Rate drops below 40%, an Activation Guardrail badge (warning) is triggered, advising operators that paid ad network match rates will be severely degraded.
  • Stale Insight Invalidation: If the audience rules are modified in the Audience Builder, the backend definition version increments, immediately flagging current insight metrics as Outdated until recomputed.

6. Troubleshooting & Error Playbook​

SymptomProbable Root CauseActionable Resolution
Error: "Audience definition not found"Invalid audience ID in the route URL or audience was deleted in Postgres.Return to Audience Management and select a valid saved audience.
Trend chart displays "No trend data returned"Audience was created recently and has fewer than two historical execution runs.Trend data will automatically populate after subsequent evaluation or scheduled sync runs.
Reachability Rate shows 0%Customer profiles in this segment lack mapped identity columns (Customer.email, Customer.mobile).Check data source connections and identity resolution mappings in your CDP configuration.
Insight query fails with 504 Gateway TimeoutComplex nested filter across billions of event rows without lookback bounds.Edit the audience definition in Audience Builder and add a time window constraint (e.g., in last 30 days).

7. Downstream Handoff & Next Steps​