skills/architect-audience-sync/SKILL.md
Cross-platform audience synchronization design across advertising platforms including Facebook, Google, TikTok, LinkedIn, and programmatic networks.
npx skillsauth add delta-and-beta/braze-agency architect-audience-syncInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
★ Insight ─────────────────────────────────────
Nick's two-layer content hierarchy means this generated SKILL.md acts as the synthesis layer — it references atomic topic files (the references/ subdirectory) rather than embedding all their detail. The skill file itself should be lean and directive, leaving raw platform docs in the topic files for on-demand loading.
─────────────────────────────────────────────────
Here is the generated skill body:
This skill guides the design and implementation of Braze Audience Sync integrations — the Canvas-native mechanism for pushing first-party user segments from Braze directly into advertising platform audiences. Use this skill when architecting how behavioral CRM data flows from Braze into paid media channels: Facebook Custom Audiences, Google Customer Match, TikTok Audiences, LinkedIn Matched Audiences, Snapchat Customer Lists, Pinterest Audiences, Criteo segments, and The Trade Desk's programmatic network.
The architectural lens is data flow design: how user identity maps across platforms, how Canvas orchestration controls sync timing and membership, how to handle consent requirements and data residency constraints, and how to avoid common integration pitfalls that cause audience drift or compliance exposure.
Apply this skill when:
Audience Sync is a Canvas component, not a standalone broadcast tool. It executes within user journey logic — sync decisions are made per-user as they traverse a Canvas, which means:
The Canvas Audience Sync step fires after a user enters that step in the flow. This has latency implications: audience population on the ad platform side can take minutes to hours depending on the platform's ingestion pipeline.
Each platform uses a different primary identity for matching. Design data collection and normalization to support the matching identifiers required by target platforms:
| Platform | Primary Match Keys | |---|---| | Facebook | Email, phone, mobile device ID (IDFA/GAID), Facebook UID | | Google | Email, phone, mobile device ID | | TikTok | Email, phone, mobile device ID | | LinkedIn | Email, LinkedIn profile ID | | Snapchat | Email, phone, mobile device ID | | Pinterest | Email, mobile device ID | | Criteo | Email, mobile device ID | | The Trade Desk | Email, phone, UID2 |
Design principle: Collect and store hashed email and phone at the profile level in Braze. These are the highest-coverage match keys across all platforms. Hash normalization (lowercase → SHA-256) happens inside Braze before transmission — store the plaintext, let Braze handle hashing at sync time.
All Audience Sync destinations use OAuth-based authorization. The integration pattern is:
Architectural implication: A single broken OAuth credential blocks all Canvas steps using that account. Build monitoring and alerts around token health. Reconnection requires admin action in the dashboard — it cannot be automated from the Canvas side.
Facebook Audience Sync supports Custom Audiences via the Marketing API. Key design decisions:
Google Audience Sync maps to Google Customer Match. Architectural notes:
TikTok Audience Sync uses the TikTok Audiences API. Key considerations:
LinkedIn Audience Sync maps to LinkedIn Matched Audiences. Architectural notes:
The Trade Desk Audience Sync feeds programmatic DSP targeting. Architectural notes:
These platforms follow the same OAuth + customer list pattern. Key notes:
Design consent controls as upstream Canvas gates, not post-sync filters:
Removal architecture: Audience Sync supports both add and remove operations. Build explicit removal logic into Canvas for: opt-outs, post-conversion suppression, and audience refresh cycles.
Audience population follows a two-phase latency model:
Design campaign activation timing to account for this pipeline. Do not schedule ad campaigns to launch immediately after Canvas audience population — build in a buffer.
Model audience membership as a stateful lifecycle, not a one-time export:
Platform API failures silently drop sync operations by default. Build observability into the architecture:
This skill draws from the following reference topics:
audience_sync — Canvas Audience Sync step mechanicscanvas_audience_sync_overview — Feature overview and getting startedaudience_sync_getting_started — Initial configuration and account connectionfacebook_audience_sync — Facebook Custom Audiences integrationgoogle_audience_sync — Google Customer Match, EU consent requirementstiktok_audience_sync — TikTok Audiences APIlinkedin_audience_sync — LinkedIn Matched Audiencessnapchat_audience_sync — Snapchat customer listspinterest_audience_sync — Pinterest Audiencescriteo_audience_sync — Criteo segment synctrade_desk_audience_sync — The Trade Desk programmatic integrationaudience_sync_faq — Timing, troubleshooting, and platform-specific edge cases★ Insight ─────────────────────────────────────
Platform identity heterogeneity is the central design constraint in audience sync architecture — each destination uses a different primary key (email, phone, device ID, UID2). The skill surfaces this as a data collection imperative: store plaintext email/phone at the Braze profile level and let Braze handle hashing, rather than pre-hashing at the source which would break cross-platform compatibility. Canvas as control plane is the other key pattern: sync decisions live inside journey logic, not in batch exports, which means timing, consent gating, and membership removal all flow through the same Canvas orchestration layer.
─────────────────────────────────────────────────
development
Defines cross-cutting API patterns for authentication, provisioning, preference management, and content delivery.
development
Covers API basics, authentication, rate limits, error codes, endpoint overview, data retention policies, and Postman collection usage.
development
Integration architecture for AI model providers including OpenAI, Google Gemini, and Anthropic within Braze messaging workflows.
development
Designs workspace structures, manages platform settings, rate limits, and API configurations.