Skip to content

PranavNagrecha/AwesomeSalesforceSkills / skills/admin/data-cloud-segmentation

data-cloud-segmentation

Verified

Create, filter, refresh, and activate audience segments in Salesforce Data Cloud: segment types (Standard, Real-Time, Waterfall, Dynamic, Data Kit), segment refresh schedules, activation field mapping, audience publishing. NOT for the SQL metrics a segment filters on - use admin/data-cloud-calculated-insights. NOT for ingestion, DLO to DMO mapping and identity resolution - use data/data-cloud-data-streams.

Data Cloud Segmentation

This skill activates when a practitioner needs to create, configure, publish, or troubleshoot audience segments in Salesforce Data Cloud. It covers the full segment lifecycle: building filter logic against Unified Profiles, choosing the right segment type and refresh schedule, mapping attributes to activation targets, and diagnosing why contacts are missing from downstream systems.


Before Starting

Gather this context before working on anything in this domain:

  • Confirm Data Cloud is provisioned and at least one data stream is ingested and identity-resolved. Segments without a populated Unified Individual DMO return empty populations.
  • Know the org's current Rapid Publish segment count. The hard org-wide limit is 20 Rapid Publish segments. Exceeding it silently prevents creation of new Rapid Publish segments without surfacing an obvious error.
  • Identify whether the target activation system (Salesforce CRM, Marketing Cloud Engagement, cloud storage) is already configured as an activation target in Data Cloud setup. Activation cannot proceed without a named target.
  • Clarify the segment's population size expectation. Segments with populations over 10 million Unified Profiles cannot use related attributes in activations.

Core Concepts

Segment Types

Data Cloud supports five segment types, each with different use cases and constraints:

TypeDescriptionKey Constraint
StandardFilter-based segment refreshed on a scheduleDefault; 12–24 hour refresh cycle
Real-TimeEvaluated near-continuously for use in real-time personalizationRequires real-time data stream; higher processing cost
WaterfallMutually exclusive priority-ordered bucketsContacts can appear in only one bucket
DynamicSegment membership driven by changes to a related data objectUseful for event-triggered audiences
Data KitSegments packaged and distributed via Data KitsCross-org distribution pattern

The org-wide hard limit is 9,950 total segments across all types.

Segment Refresh Schedules

Segment refresh and activation delivery are independently configured. Refreshing a segment more frequently does not accelerate when activation data lands in the destination system.

Three refresh options exist:

  • Standard refresh — runs every 12–24 hours; evaluates the full data history for the segment's date range.
  • Rapid Publish — runs every 1–4 hours but only considers the last 7 days of data. Hard org limit: 20 Rapid Publish segments total. Segments requiring history older than 7 days will silently under-count members if Rapid Publish is selected.
  • Incremental refresh — evaluates only records that changed since the last run; reduces compute but requires a change-detection-compatible data model.

Activation and Field Mapping

An activation connects a segment to a named activation target and specifies which attributes to publish alongside the segment membership record. Limits:

  • Maximum 20 related attributes per activation (attributes pulled from objects related to the Unified Individual, not direct profile attributes).
  • Maximum 100 activations with related attributes per org.
  • Segments with 10M+ population cannot include any related attributes; only core identity fields are supported.

Activation has its own refresh schedule that is separate from the segment refresh. A segment that refreshes every 2 hours on Rapid Publish may still deliver to a Marketing Cloud activation only once per day if the activation is configured for daily refresh.

Null Email Addresses and Contact Completeness

Data Cloud activates segment members regardless of whether the contact has an email address unless the segment filter explicitly excludes null email. This is a platform default that frequently causes unexpected deliverability failures in downstream systems. Always add an explicit filter: Email IS NOT NULL (or the equivalent on the activation target's required identity field) when the downstream system requires a valid contact identifier.


Common Patterns

Pattern: Standard Segment with CRM Activation

When to use: Sending a daily audience of qualified leads or contacts from Data Cloud back to a Salesforce CRM campaign or list for Sales Cloud follow-up.

How it works:

  1. In Data Cloud, navigate to Segments and create a new Standard segment.
  2. Define filter criteria on the Unified Individual DMO (e.g., LifetimeValue > 500, LastPurchaseDate within 90 days, Email IS NOT NULL).
  3. Set refresh schedule to Standard (every 12 hours is sufficient for CRM use cases).
  4. Navigate to Activations, create a new activation pointing to the Salesforce CRM activation target.
  5. Map the segment to the target CRM object (Contact or Lead), map required fields (Email, FirstName, LastName), and optionally add up to 20 related attributes.
  6. Set the activation publish schedule (independent of segment refresh).
  7. Publish manually to validate the first run, then confirm records appear in the CRM target list.

Why not a simpler approach: Reports and list views in CRM cannot pull cross-cloud data or apply Data Cloud identity resolution. Only activation delivers a refreshed, deduplicated list.

Pattern: Rapid Publish Segment for Real-Time Personalization

When to use: Feeding a segment into a personalization engine (e.g., Marketing Cloud Personalization or a connected CDP activation) where audience membership needs to reflect behavior from the last few hours.

How it works:

  1. Verify the org has fewer than 20 existing Rapid Publish segments before creating a new one.
  2. Create a segment with filter criteria scoped to attributes that change within a 7-day window (e.g., CartAbandonedDate within 2 days, Email IS NOT NULL).
  3. Set refresh to Rapid Publish (1–4 hour cadence).
  4. Do not use Rapid Publish for segments that require historical lookback longer than 7 days — lifetime spend, multi-year engagement — because older data is excluded from the evaluation window.
  5. Create the activation with a matching high-frequency publish schedule.

Why not Standard: Standard refresh runs every 12–24 hours; for cart abandonment or browse behavior, that lag misses the engagement window.


Decision Guidance

SituationRecommended ApproachReason
Daily campaign audience for CRM outreachStandard segment + Standard activationNo need for sub-24h freshness; avoids Rapid Publish quota
Cart abandonment or time-sensitive behavior in last 48hRapid Publish segmentRefreshes every 1–4 hours; fits 7-day lookback window
Need 30-day or longer historical lookbackStandard segmentRapid Publish only looks back 7 days
Org already has 20 Rapid Publish segmentsStandard segment with shortest allowed scheduleCannot exceed 20 Rapid Publish cap
Segment population > 10 million profilesStandard segment, no related attributes on activationRelated attributes not supported above 10M
Mutually exclusive VIP / Standard / Lapsed tiersWaterfall segmentEnsures each profile lands in exactly one bucket
Contacts missing from activation with no errorAdd Email IS NOT NULL filter, check activation refresh scheduleNull emails pass through; activation has its own schedule

Recommended Workflow

Step-by-step instructions for an AI agent or practitioner working on this task:

  1. Confirm prerequisites — Verify Data Cloud is provisioned, at least one data stream is ingested and identity-resolved, and an activation target is configured. Check the org's current Rapid Publish segment count if planning a high-frequency segment.
  2. Define segment type and filter logic — Choose the segment type (Standard, Real-Time, Waterfall, Dynamic, Data Kit) based on the audience use case. Build filter criteria against the Unified Individual DMO. Add an explicit Email IS NOT NULL filter (or equivalent required identity field) unless the downstream system can handle null identifiers.
  3. Select refresh schedule — Choose Standard (12–24h) for most use cases. Use Rapid Publish only if the use case requires sub-4-hour freshness AND the data lookback fits within 7 days AND the org has fewer than 20 existing Rapid Publish segments. Use Incremental for large data volumes where change detection is feasible.
  4. Create and configure the activation — Navigate to Activations, select the segment and activation target, map required identity fields, and add up to 20 related attributes. Set the activation's own publish schedule independently of the segment refresh.
  5. Publish and validate — Trigger a manual publish on the first run. Confirm member count in the segment preview matches expectations. Confirm records appear in the activation target within the expected delivery window.
  6. Monitor for limit violations — After publishing, verify segment population count, check activation delivery logs, and confirm Rapid Publish quota is not exceeded at the org level.

Review Checklist

Run through these before marking work in this area complete:

  • Segment filter includes explicit null exclusion on required identity field (Email IS NOT NULL or equivalent)
  • Segment type is appropriate for the use case (Standard for daily; Rapid Publish only for <7-day lookback and org count <20)
  • Activation refresh schedule is explicitly configured and matches the delivery SLA (not assumed to inherit segment schedule)
  • Related attribute count per activation does not exceed 20
  • Segment population is below 10M if related attributes are mapped in the activation
  • Org-wide Rapid Publish count verified before selecting Rapid Publish refresh
  • Org-wide segment count checked against the 9,950 limit for large deployments

Salesforce-Specific Gotchas

Non-obvious platform behaviors that cause real production problems:

  1. Rapid Publish silently under-counts when data is older than 7 days — Rapid Publish evaluates only the last 7 days of data regardless of how the segment filter is written. A filter for LastPurchaseDate within 30 days will appear to work but will only evaluate events from the past 7 days, silently missing contacts whose last purchase was 8–30 days ago.
  2. Segment refresh schedule and activation publish schedule are independent — Increasing segment refresh frequency does not change how often the activation delivers data to the downstream system. Both must be configured separately. A Rapid Publish segment paired with a daily activation schedule delivers no faster than once per day.
  3. Null email addresses are activated by default — Data Cloud includes contacts with null email in segment membership. Unless a filter explicitly excludes nulls, the activation will attempt to send contacts with no email to the downstream system, causing silent failures or malformed records in Marketing Cloud or CRM.
  4. 9,950 segment limit is org-wide and hard — There is no UI warning as the org approaches the limit. At exactly 9,950 segments, creation fails with a generic error. Large orgs with automated segment generation must track count proactively.
  5. Related attribute cap blocks activation setup for large segments — The 20 related attribute limit per activation and 100 total activations-with-related-attributes limit per org are enforced at save time. For segments over 10M profiles, the activation form will not allow related attribute selection at all.

Output Artifacts

ArtifactDescription
Published segmentA named Data Cloud segment with filter criteria, population count, and refresh schedule
Activation configurationMapping of segment to activation target with field mappings and publish schedule
Segment filter specificationDocumented filter logic including null exclusions and date range constraints
Refresh strategy decision recordRationale for Standard vs. Rapid Publish vs. Incremental choice

Related Skills

  • admin/data-cloud-identity-resolution — identity resolution must be completed before segment population reflects unified profiles; run this skill first
  • admin/data-cloud-provisioning — org must be provisioned with Data Cloud before segments can be created
  • admin/data-cloud-calculated-insights — calculated insights can be used as filter attributes in segments; use when segment criteria require aggregated metrics
  • admin/data-cloud-activation-development — for custom activation targets beyond Salesforce CRM, Marketing Cloud, and native connectors
deprecation-and-migrationaddyosmani/agent-skillsManages deprecation and migration. Use when removing old systems, APIs, or features. Use when migrating users from one implementation to another. Use when migrating a database schema in production, such as renaming or dropping a column without downtime (expand/contract). Use when deciding whether to maintain or sunset existing code.Stars: 119kUpdated Claude CodeCodexVerifiedInstall
performance-optimizationaddyosmani/agent-skillsOptimizes application performance across frontend, backend, queries, and databases. Use when performance requirements exist, when you suspect performance regressions, when Core Web Vitals or load times need improvement, when N+1 query patterns need fixing, or when profiling reveals bottlenecks.Stars: 119kUpdated Claude CodeCodexVerifiedInstall
Agent Cost Reportthedotmack/claude-memBelievable agent cost report for any period, default the last 7 full days PT, not counting today. Measured tokens from Claude Code transcripts priced at OpenRouter list prices (ESTIMATED), measured provider spend when a sanctioned source exists, note-taker cost separate, Timing-style HTML/PDF plus report.json, line-items.csv, evidence.json.Stars: 118kUpdated Claude CodeCodexVerifiedInstall
Agent Cost Reportthedotmack/claude-memBelievable agent cost report for any period, default the last 7 full days PT, not counting today. Measured tokens from Claude Code transcripts priced at OpenRouter list prices (ESTIMATED), measured provider spend when a sanctioned source exists, note-taker cost separate, Timing-style HTML/PDF plus report.json, line-items.csv, evidence.json.Stars: 118kUpdated Claude CodeCodexVerifiedInstall