Skip to main content

πŸ”— Cross-Platform Identity Resolution

UniPulse automatically links the same person's profiles across Facebook, Instagram, TikTok, and other platforms into a single unified AudienceNode. This means you get one profile, one engagement score, and one conversation history per real person β€” not per platform account.


Why It Matters​

Without identity resolution, a customer who comments on your Facebook post and DMs you on Instagram looks like two separate people. This leads to:

  • Fragmented data β€” Engagement scores that undercount actual activity
  • Duplicate outreach β€” Contacting the same person through multiple channels
  • Lost context β€” Agents missing conversation history from other platforms
  • Inflated audience counts β€” Your "10,000 contacts" might be 7,000 real people

Cross-platform identity resolution solves all of these by merging profiles into a single AudienceNode with unified socialHandles, aggregate metrics, and a complete interaction timeline.


How It Works​

The resolution engine uses multiple signals with confidence-weighted matching:

SignalConfidenceDescription
Matching emailVery HighSame email address across platform profiles
Linked social accountsVery HighPlatforms that report linked accounts via API
Same display nameMediumExact or fuzzy name matching across platforms
Phone numberHighMatching phone numbers (when available)
Behavioral patternsLow–MediumSimilar interaction patterns, timing, and content

The Merge Process​

  1. Signal collection β€” When a new interaction arrives, the system extracts identifying signals (name, email, social handle)
  2. Candidate matching β€” Existing audience nodes are searched for matching signals
  3. Confidence scoring β€” Each match signal is weighted; a combined confidence score determines if profiles should merge
  4. Profile merge β€” If confidence exceeds the threshold, profiles are merged into a single AudienceNode:
    • socialHandles JSON is updated with all known platform handles
    • totalComments, totalDMs, totalPurchases, totalSpent are recalculated across all linked profiles
    • engagementScore is recomputed from the full interaction history
    • firstSeenAt takes the earliest date across all profiles
    • lastInteractionAt takes the most recent
  5. Ongoing linking β€” As new data arrives, previously unmatched profiles may be linked retroactively

What Gets Unified​

When profiles are merged, the resulting AudienceNode aggregates:

FieldResolution Strategy
displayNameMost complete name available
avatarUrlMost recently updated avatar
email / phoneAll known values stored
socialHandlesCombined JSON map of all platform handles
engagementScoreRecalculated from all interactions across all platforms
totalComments / totalDMsSum across all linked profiles
totalPurchases / totalSpent / avgOrderValueAggregated purchase data
tagsUnion of all tags from merged profiles
metadataDeep-merged JSON
Behind the Scenes

The AudienceInteraction records are never duplicated β€” each interaction retains its original platform and postId fields. The merge only updates the parent AudienceNode to point all interactions to the unified profile. This means your interaction-level analytics remain platform-accurate while your profile-level data is person-accurate.


Viewing Linked Identities​

On any audience profile in the Audience Graph, you can see:

  • Linked platforms β€” Which social handles are connected to this profile
  • Per-platform breakdown β€” Interaction counts and engagement metrics per platform
  • Merge history β€” When and why profiles were linked
  • Manual override β€” Split incorrectly merged profiles or manually merge profiles that the system missed

Privacy & Data Handling​

  • Identity resolution only uses publicly available information and data shared through official platform APIs
  • No data is shared between workspaces β€” resolution is scoped to your workspace
  • All data is stored securely and encrypted, within your workspace's isolated data boundary
  • Customers can be removed from the audience graph upon request (right to deletion)
  • The system respects platform API terms of service for data usage

  • Audience Graph β€” The unified profiles that identity resolution creates
  • Segments & Scoring β€” Accurate scoring depends on complete, merged profiles
  • Unified Inbox β€” Conversations linked to resolved identities via audienceNodeId