Salesforce Multi-Org Architecture: Patterns, Governance & Data Sharing

A practical enterprise architecture guide to designing, integrating, governing, and sharing data across multiple Salesforce orgs

Running multiple Salesforce orgs is not simply a matter of creating additional environments. At enterprise scale, every additional org creates another boundary for data, identity, security, configuration, integrations, releases, reporting, and governance.

A multi-org Salesforce architecture can provide valuable business autonomy and regulatory isolation. It can also create fragmented customer data, duplicated capabilities, inconsistent processes, integration complexity, and governance challenges.

The architectural objective is therefore not to make every Salesforce org identical.

The objective is to create the right boundaries while maintaining the right enterprise-level capabilities across those boundaries.

This guide explains the major multi-org architecture patterns, how to define org boundaries, how to share data, how to integrate orgs, how to govern them, and how modern capabilities such as Data 360 and Data Cloud One change the architecture conversation.

Related decision guide: If you are still deciding whether your enterprise should use a single org or multiple orgs, start with the Single Org vs Multi-Org Architecture Decision Guide Click here.. This guide assumes that a multi-org strategy has already been selected or that the enterprise already operates multiple Salesforce orgs.

1. What Is Salesforce Multi-Org Architecture?

Salesforce multi-org architecture is an enterprise operating model in which an organization deliberately operates two or more Salesforce production orgs, with defined boundaries between them and explicit mechanisms for sharing identity, data, capabilities, and enterprise services.

Those orgs may represent:

  • Business units
  • Geographic regions
  • Brands
  • Product lines
  • Legal entities
  • Acquired companies
  • Regulated operations
  • Different operating models
  • Different customer populations
  • Separate functional domains

The important word is deliberately.

An enterprise with ten Salesforce orgs created independently over several years has a multi-org landscape.

That does not necessarily mean it has a multi-org architecture.

Architecture begins when the enterprise defines:

  • Why each org exists
  • What each org owns
  • What data belongs in each org
  • Which capabilities are shared
  • Which capabilities remain autonomous
  • How orgs communicate
  • How identities are managed
  • How enterprise data is governed
  • How changes are coordinated
  • How the overall landscape evolves

2. Why Multi-Org Architecture Is Difficult

A Salesforce org is more than a database.

It contains a combination of:

  • Data
  • Metadata
  • Business processes
  • Security
  • Users
  • Automation
  • Apex
  • Flows
  • Integrations
  • Reports and dashboards
  • Experience channels
  • Platform capabilities
  • AI capabilities

When you create multiple orgs, you create multiple instances of these architectural concerns.

For example:

                 ENTERPRISE
                     │
       ┌─────────────┼─────────────┐
       │             │             │
       ▼             ▼             ▼
    Org A          Org B          Org C
    North          Europe        APAC
       │             │             │
    Data A         Data B        Data C
    Process A      Process B     Process C
    Security A     Security B    Security C

The enterprise now has three autonomous Salesforce environments.

The architecture challenge becomes:

What should remain local, and what must behave like an enterprise capability?

That question should drive the design.

3. The Fundamental Multi-Org Principle

A useful architecture principle is:

Localize what requires autonomy. Centralize what requires consistency. Federate what requires collaboration.

This produces three architectural categories.

Local capabilities

Capabilities that should remain inside an org:

  • Business-unit-specific processes
  • Local workflows
  • Local automation
  • Local configuration
  • Local security policies
  • Local regulatory processes
  • Local user administration

Enterprise capabilities

Capabilities that should be governed consistently:

  • Identity
  • Customer identity
  • Enterprise data definitions
  • Security principles
  • API standards
  • Integration standards
  • Architecture standards
  • Data governance
  • Audit requirements

Federated capabilities

Capabilities that must operate across org boundaries:

  • Customer information
  • Orders
  • Cases
  • Product information
  • Account relationships
  • Marketing interactions
  • Events
  • Analytics
  • AI context

This distinction is one of the most important principles in a multi-org architecture.

4. Major Salesforce Multi-Org Architecture Patterns

There is no single multi-org pattern that fits every enterprise.

The most common patterns are below.

4.1 Regional Org Pattern

Each geographic region operates its own Salesforce org.

                    Enterprise
                        │
          ┌─────────────┼─────────────┐
          │             │             │
          ▼             ▼             ▼
       Americas        EMEA          APAC
          │             │             │
       Salesforce     Salesforce    Salesforce
          Org            Org          Org

Why enterprises use it

  • Data residency
  • Regional regulations
  • Local business processes
  • Regional autonomy
  • Local sales/service operations
  • Different legal entities

Architecture challenge

The enterprise still needs a global view of:

  • Customers
  • Revenue
  • Products
  • Orders
  • Service interactions
  • Account relationships

Therefore, regional separation should not automatically become enterprise data fragmentation.

4.2 Business-Unit Org Pattern

Each major business unit operates an independent Salesforce org.

Example:

                    Enterprise
                        │
       ┌────────────────┼────────────────┐
       │                │                │
       ▼                ▼                ▼
     Retail           Wealth          Insurance
       Org               Org              Org

This pattern is particularly common when business units have substantially different:

  • Products
  • Processes
  • Users
  • Security models
  • Regulatory requirements
  • Operating models

The architecture question becomes:

Which customer and enterprise data must cross the business-unit boundary?

4.3 Brand-Based Org Pattern

Each major brand operates its own Salesforce environment.

This can be appropriate where brands have:

  • Independent customer journeys
  • Different product models
  • Separate operating teams
  • Distinct security requirements
  • Independent commercial strategies

However, architects should be careful not to equate brand identity with data ownership.

Two brands may have different customer experiences while still needing an enterprise-level customer identity.

4.4 Acquisition-Based Org Pattern

An acquired company continues operating its existing Salesforce org.

Parent Company
      │
      ├── Existing Org
      │
      └── Acquired Company
              │
              └── Existing Salesforce Org

This is often the fastest post-acquisition strategy.

It avoids immediate migration risk while allowing the business to stabilize.

The long-term architecture may then move toward:

  • Consolidation
  • Federated multi-org
  • Hybrid architecture

The key is to avoid allowing a temporary acquisition architecture to become permanent architecture by accident.

4.5 Functional Org Pattern

Different Salesforce orgs are used for different functional areas.

For example:

Sales Org
Service Org
Marketing Org
Commerce Org
Industry Org

This can provide strong functional autonomy but often creates substantial integration and customer-identity complexity.

Functional separation should therefore be evaluated carefully.

A business process that crosses multiple functions may become an integration problem rather than a Salesforce configuration problem.

4.6 Hybrid Multi-Org Pattern

A hybrid model combines different boundaries.

For example:

                         Enterprise
                             │
              ┌──────────────┼──────────────┐
              │              │              │
           Americas         EMEA           APAC
              │              │              │
           Sales +        Sales +        Sales +
           Service        Service        Service
              │
              ▼
        Global Data 360

A hybrid architecture may provide:

  • Regional autonomy
  • Business-unit separation
  • Shared enterprise data
  • Centralized identity
  • Federated integrations
  • Enterprise analytics

For large enterprises, hybrid architectures are often more realistic than trying to force every business unit into one organizational pattern.

5. Define Org Boundaries Before Designing Integrations

One of the most common architectural mistakes is designing integrations before defining ownership.

The sequence should be:

Business Boundary
       ↓
Capability Ownership
       ↓
Data Ownership
       ↓
Security Boundary
       ↓
Integration Boundary
       ↓
Technology Implementation

Not:

"We have three orgs."
       ↓
"Let's connect them."

Integration should be a consequence of architecture, not a substitute for it.

6. The Org Boundary Model

For every Salesforce org, document five things.

1. Business ownership

Who owns the business capability?

2. Data ownership

Which system is authoritative for each major data domain?

3. Process ownership

Which org owns the business process?

4. Security ownership

Who determines access and policy?

5. Lifecycle ownership

Who controls releases, upgrades, and architectural changes?

A useful architecture artifact is an Org Boundary Matrix.

DimensionOrg AOrg BOrg C
Business ownerRegion ARegion BRegion C
Customer ownershipLocalLocalLocal
Global customer identityEnterpriseEnterpriseEnterprise
Sales processLocalLocalLocal
Product masterEnterpriseEnterpriseEnterprise
PricingLocalLocalLocal
IdentityEnterpriseEnterpriseEnterprise
Integration platformEnterpriseEnterpriseEnterprise
Release governanceFederatedFederatedFederated

The exact values will differ by enterprise.

The important thing is that ownership is explicit.

7. Data Sharing Is the Core Multi-Org Challenge

The hardest question in a multi-org architecture is usually not:

“How do we connect the orgs?”

It is:

“What data should cross the org boundary, why, and who owns it?”

Without clear ownership, multi-org data sharing quickly becomes a synchronization problem.

8. Establish a System of Record

For every important data domain, define an authoritative source.

Example:

Data DomainSystem of Record
Customer IdentityEnterprise MDM / Data 360
AccountRegional Salesforce / MDM
ContactRegional Salesforce / MDM
ProductERP / Product Master
PricingERP / Pricing Platform
OrderCommerce / ERP
CaseSalesforce Service Org
EmployeeHR System
Marketing ConsentEnterprise Consent Platform

The exact architecture depends on the organization.

But there should be one authoritative owner for every critical data domain.

9. Avoid the “Everything Synchronizes With Everything” Pattern

This is one of the most dangerous multi-org architectures.

       Org A
      ↙  ↓  ↘
    Org B ↔ Org C
      ↘  ↓  ↙
       Org D

Every org becomes both producer and consumer.

This creates:

  • Circular dependencies
  • Duplicate records
  • Conflicting updates
  • Difficult error handling
  • Unclear ownership
  • Increasing integration costs

A better pattern is:

                Enterprise Data Layer
                       │
          ┌────────────┼────────────┐
          ▼            ▼            ▼
        Org A         Org B        Org C

The enterprise data layer may be:

  • MDM
  • Data 360
  • Integration platform
  • Enterprise data platform
  • Combination of these

The appropriate choice depends on the data domain and use case.

10. Data Sharing Patterns

There are several legitimate ways to share data across Salesforce orgs.

Pattern A — API-Based Data Sharing

One org exposes APIs consumed by another.

Org A
 │
 ▼
API Layer
 │
 ▼
Org B

Useful for:

  • Real-time lookups
  • Transactional interactions
  • Controlled data access

Advantages:

  • Explicit ownership
  • Real-time access
  • Good governance

Challenges:

  • Runtime dependency
  • Latency
  • Availability coupling
  • API limits

Pattern B — Event-Based Data Sharing

An org publishes events when data changes.

Org A
  │
  ▼
Event Platform
  │
  ├──> Org B
  ├──> Org C
  └──> Data Platform

This is often preferable when multiple consumers need the same business event.

Salesforce capabilities such as Platform Events and Change Data Capture can participate in event-driven architectures.

For enterprise integration, however, architects should distinguish between Salesforce event capabilities and the broader enterprise event backbone.

Pattern C — Integration Platform

An enterprise integration layer mediates communication.

             Integration Platform
              /       |       \
             /        |        \
          Org A      Org B     Org C

This provides a useful abstraction when there are many systems and orgs.

The integration layer can provide:

  • Transformation
  • Routing
  • Orchestration
  • Error handling
  • Monitoring
  • Security
  • API management

For enterprises already operating an API-led integration platform, this can become the principal integration boundary between Salesforce orgs.

Pattern D — Data Platform / Data 360

Instead of synchronizing every customer record directly between Salesforce orgs, the enterprise can establish a governed data layer.

Modern Salesforce architecture increasingly uses Data 360 for this purpose.

Salesforce’s current architecture guidance describes Data Cloud One as a multi-org pattern in which one Salesforce org acts as the Data 360 home org, while other Salesforce orgs connect as companion orgs. Shared data spaces and metadata allow companion orgs to access governed unified data.

This is fundamentally different from building a large collection of point-to-point synchronizations.

11. Data 360 and Multi-Org Architecture

Data 360 has materially changed the multi-org architecture discussion.

Historically, enterprises often approached multi-org data through:

Org A ↔ Integration ↔ Org B
Org A ↔ Integration ↔ Org C
Org B ↔ Integration ↔ Org C

With Data Cloud One, a different pattern becomes possible:

                    Data 360
                   Home Org
                       │
             ┌─────────┼─────────┐
             │         │         │
             ▼         ▼         ▼
        Companion   Companion   Companion
           Org A       Org B       Org C

Salesforce describes Data Cloud One as allowing multiple Salesforce orgs to use a single central Data 360 instance, with companion orgs accessing unified data and Data 360 capabilities.

This can reduce duplication and help establish an enterprise-level customer view.

But it does not mean every multi-org enterprise should centralize everything.

12. When Data 360 Should Be Centralized

Centralized Data 360 can be attractive when the enterprise wants:

  • A unified customer profile
  • Centralized data governance
  • Cross-org analytics
  • Shared AI context
  • Common identity resolution
  • Enterprise-wide segmentation
  • Shared activation
  • Reduced data silos

Salesforce’s current architecture guidance specifically positions Data Cloud One as a multi-org connectivity architecture for organizations with business-unit, regional, product-line, or acquisition-driven org separation.

13. When Multiple Data 360 Instances May Be Appropriate

Centralization is not always correct.

Multiple Data 360 instances may be necessary where there are strong:

  • Regulatory requirements
  • Data residency constraints
  • Legal-entity boundaries
  • Security isolation requirements
  • Organizational autonomy requirements

Salesforce explicitly notes that some organizations may require more than one Data 360 instance because of legal or business requirements.

Therefore:

Multi-org does not automatically mean one Data 360 instance.

The architecture must consider the legal and data-processing boundaries of the enterprise.

14. Data Residency Must Be Designed Explicitly

Data residency is particularly important in global multi-org environments.

Consider:

North America Org
        │
        │
        ▼
   North America
    Data 360

Europe Org
        │
        │
        ▼
     Europe
    Data 360

versus:

           Global Data 360
                 │
       ┌─────────┼─────────┐
       ▼         ▼         ▼
      NA        EU        APAC

The second model may be operationally attractive but can introduce data-processing considerations.

Salesforce’s current Data 360 architecture guidance notes that connected orgs may process data across territories and recommends explicitly reviewing data residency requirements and applicable legal obligations.

Architects should therefore involve:

  • Legal
  • Privacy
  • Security
  • Compliance
  • Data governance

before defining the final cross-region data model.

15. Identity Architecture Across Multiple Orgs

Users should not experience the enterprise as a collection of disconnected Salesforce environments.

A multi-org strategy therefore needs an enterprise identity architecture.

Typical model:

                 Enterprise IdP
                       │
          ┌────────────┼────────────┐
          ▼            ▼            ▼
        Org A         Org B        Org C

Central identity can provide:

  • Single sign-on
  • Central lifecycle management
  • MFA policies
  • Conditional access
  • Role/group integration
  • Consistent authentication

However, authentication centralization does not automatically mean authorization centralization.

Each Salesforce org may still require its own:

  • Permission model
  • Role hierarchy
  • Sharing model
  • Profiles/permission sets
  • Business-unit access policies

Therefore:

Centralize authentication where appropriate; federate authorization according to business ownership.

16. Security Architecture

Multi-org environments should define security at three levels.

Enterprise level

  • Identity
  • Security standards
  • Encryption
  • Compliance
  • Audit
  • Privacy

Org level

  • Salesforce sharing model
  • Profiles and permission sets
  • Object access
  • Field-level security
  • Record access

Integration level

  • API authentication
  • OAuth
  • Named Credentials
  • Certificates
  • Secrets
  • Network controls
  • Integration-user governance

A common mistake is assuming that multiple orgs automatically provide sufficient security isolation.

They provide boundaries.

They do not replace a security architecture.

17. Integration Architecture

A multi-org Salesforce environment should have a deliberate integration strategy.

A useful reference model is:

                  Enterprise Systems
                         │
                  Integration Layer
                 /       │        \
                /        │         \
             Org A      Org B      Org C

The integration layer can provide:

  • API management
  • Transformation
  • Routing
  • Orchestration
  • Event processing
  • Observability
  • Security
  • Retry handling
  • Dead-letter handling

The principle should be:

Don’t use Salesforce orgs as an accidental integration hub for the entire enterprise.

Each org should own its business capabilities, while the enterprise integration architecture manages cross-domain communication.

18. Synchronous vs Asynchronous Communication

Not every cross-org interaction should be synchronous.

Use synchronous communication when:

  • A user needs an immediate response
  • A transaction depends on current information
  • A validation must happen in real time

Use asynchronous communication when:

  • A downstream system can process later
  • Multiple consumers need the event
  • High scalability is required
  • Temporary downstream unavailability should not block the source transaction

This distinction is especially important when multiple Salesforce orgs are involved.

19. Change Data Capture in Multi-Org Architectures

Change Data Capture can be useful when downstream systems need to react to Salesforce record changes.

Salesforce describes CDC as providing a stream of change events for record insertions, updates, deletions, and undeletions, with a versioned event schema and temporary event retention.

A common architecture is:

Salesforce Org
      │
      ▼
Change Event
      │
      ▼
Event / Integration Layer
      │
   ┌──┴────┐
   ▼       ▼
Org B    Data Platform

But CDC should not be treated as a universal replication solution.

Architects still need to define:

  • Which objects are replicated
  • Which fields matter
  • Event ownership
  • Ordering requirements
  • Replay strategy
  • Idempotency
  • Error handling
  • Data retention
  • Consumer responsibilities

20. Avoid Bidirectional Synchronization Unless Necessary

Bidirectional synchronization looks attractive:

Org A  ⇄  Org B

But it introduces difficult questions:

  • Who wins when both sides change the record?
  • How are conflicts resolved?
  • What happens when one side is unavailable?
  • How are deletes handled?
  • How do you prevent loops?
  • How do you guarantee idempotency?

Prefer:

Authoritative Source
        │
        ├──> Consumer A
        ├──> Consumer B
        └──> Consumer C

unless genuine bidirectional ownership is required.

21. Enterprise Customer Identity

One of the hardest problems in a multi-org Salesforce architecture is determining whether:

Org A Customer 123

is the same customer as:

Org B Customer 987

This is fundamentally an identity resolution problem.

Do not solve it merely by matching Salesforce Account IDs.

Salesforce IDs are org-specific.

The enterprise should consider an enterprise customer identifier such as:

Enterprise Customer ID
        │
   ┌────┼────┐
   ▼    ▼    ▼
Org A Org B Org C

This enables:

  • Cross-org customer matching
  • Enterprise reporting
  • Identity resolution
  • Duplicate management
  • Customer 360
  • Cross-business analytics

Data 360 can play an important role in this architecture through identity resolution and unified customer profiles.

22. Product and Reference Data

Customer data is not the only shared domain.

Multi-org enterprises frequently need common:

  • Product
  • Price
  • Currency
  • Country
  • Industry
  • Territory
  • Employee
  • Location
  • Legal entity
  • Account hierarchy

Reference data should generally be governed centrally even when transactional data remains local.

A useful distinction is:

Enterprise Master Data
          │
     ┌────┼────┐
     ▼    ▼    ▼
    Org A Org B Org C

Local Transactional Data
     │     │     │
     ▼     ▼     ▼
    Org A Org B Org C

23. Reporting Across Multiple Salesforce Orgs

A multi-org strategy creates an immediate reporting challenge.

Executives often ask:

“Show me total enterprise pipeline.”

But the pipeline may exist across:

  • Americas org
  • Europe org
  • APAC org
  • Acquired-company org

Aggregating reports directly between orgs can become fragile.

A better architecture may use:

Salesforce Orgs
      │
      ▼
Enterprise Data Platform
      │
      ▼
Analytics / BI

The architecture should define:

  • Enterprise metrics
  • Local metrics
  • Data latency
  • Metric definitions
  • Currency conversion
  • Hierarchy rules
  • Data quality

The biggest problem is often not technology.

It is inconsistent definitions.

24. Governance Model

Multi-org architecture requires governance at both enterprise and local levels.

A strong model is:

              Salesforce COE
                    │
        Enterprise Architecture
                    │
        ┌───────────┼───────────┐
        ▼           ▼           ▼
      Org A       Org B       Org C
      Team        Team        Team

The Center of Excellence should define enterprise standards without unnecessarily controlling every local decision.

25. What Should Be Centralized?

Typically centralize:

  • Architecture principles
  • Integration standards
  • API standards
  • Identity standards
  • Data governance
  • Security principles
  • Naming standards
  • Enterprise data definitions
  • Observability standards
  • DevSecOps standards
  • AI governance

26. What Should Remain Local?

Typically allow local autonomy for:

  • Business processes
  • Local workflows
  • User-specific configuration
  • Regional compliance processes
  • Local sales/service processes
  • Business-unit-specific reporting
  • Local operating procedures

The exact boundary depends on the enterprise.

The goal is not maximum centralization.

The goal is controlled autonomy.

27. Multi-Org DevOps

A multi-org architecture also requires a multi-org release strategy.

Without one, every org becomes its own development island.

A mature model includes:

                Enterprise DevOps
                       │
        ┌──────────────┼──────────────┐
        ▼              ▼              ▼
      Org A           Org B           Org C
        │              │              │
       CI/CD          CI/CD           CI/CD

Governance should cover:

  • Source control
  • Branching
  • Deployment
  • Testing
  • Dependency management
  • Release calendars
  • Emergency changes
  • Metadata standards
  • Environment strategy

Not every org needs identical release timing.

But enterprise dependencies must be visible.

28. Shared Services Architecture

A mature multi-org environment should identify capabilities that can operate as shared enterprise services.

Examples:

  • Identity
  • Customer identity
  • Product master
  • Integration
  • API management
  • Data 360
  • Enterprise analytics
  • Notification services
  • Document services
  • Audit
  • AI governance

This creates an architecture where Salesforce orgs are business capability platforms, not isolated technology islands.

29. Common Anti-Patterns

Anti-Pattern 1: Org Proliferation

Creating another org because:

“This business unit is different.”

Difference alone is not sufficient justification.

Anti-Pattern 2: Point-to-Point Integration Explosion

A ↔ B
A ↔ C
A ↔ D
B ↔ C
B ↔ D
C ↔ D

This becomes exponentially harder to govern.

Anti-Pattern 3: Duplicate Master Data

Every org maintains its own version of:

  • Customer
  • Product
  • Employee
  • Territory

without enterprise governance.

Anti-Pattern 4: Global Customer ID Based on Salesforce IDs

Salesforce IDs are local to an org.

They should not become the enterprise identity model.

Anti-Pattern 5: Centralization at All Costs

A multi-org architecture should not become a disguised single-org architecture.

Some boundaries exist for legitimate reasons.

Anti-Pattern 6: Local Autonomy Without Governance

The opposite extreme is equally dangerous.

Every org independently chooses:

  • APIs
  • Data models
  • Security patterns
  • Integration technologies
  • Naming conventions

The result is an architecture that becomes impossible to operate consistently.

Anti-Pattern 7: Treating Data 360 as a Universal Data Warehouse

Data 360 can provide powerful unified customer data and activation capabilities, but it does not eliminate the need for enterprise data architecture.

The enterprise still needs to define:

  • Systems of record
  • Data ownership
  • Governance
  • Retention
  • Residency
  • Data quality
  • Integration boundaries

30. Reference Multi-Org Architecture

A mature enterprise architecture might look like this:

                         ENTERPRISE
                              │
        ┌─────────────────────┼─────────────────────┐
        │                     │                     │
        ▼                     ▼                     ▼
   Enterprise ID        Enterprise Data       Integration
      Provider              Platform             Platform
        │                     │                     │
        └─────────────────────┼─────────────────────┘
                              │
                         Data 360
                       Home / Hub
                              │
             ┌────────────────┼────────────────┐
             │                │                │
             ▼                ▼                ▼
          Org A             Org B             Org C
        Americas            EMEA               APAC
             │                │                │
             └────────────────┼────────────────┘
                              │
                    Enterprise Analytics
                              │
                              ▼
                       BI / AI / Agents

This is not a prescription.

It is a reference architecture showing how local execution can coexist with enterprise-level data, identity, integration, analytics, and AI capabilities.


31. Multi-Org Architecture Decision Matrix

Architecture DimensionCentralizedFederatedHybrid
Business processLow autonomyHigh autonomyBalanced
Data governanceStrong central controlDistributedFederated
Regional autonomyLowHighHigh
Integration complexityLowerHigherMedium
Data sharingEasierMore difficultManaged
Regulatory isolationLimitedStrongStrong
Enterprise reportingEasierMore complexModerate
Release managementCentralizedIndependentCoordinated
Customer 360EasierDifficultStrong with data layer
Business-unit autonomyLowerHighHigh
Governance complexityLowerHighHigh
M&A flexibilityLowerHighHigh

The table is intentionally directional.

The correct architecture depends on the enterprise’s business, regulatory, data, and operating-model requirements.

32. A Practical Multi-Org Architecture Framework

Before approving a multi-org architecture, answer these questions.

Business

  1. Why does each org exist?
  2. Who owns it?
  3. What business capability does it represent?
  4. What autonomy is required?

Data

  1. Who owns customer data?
  2. Who owns product data?
  3. What is the enterprise customer identifier?
  4. Which data must cross org boundaries?
  5. What data must never cross boundaries?

Integration

  1. Which interactions are synchronous?
  2. Which should be event-driven?
  3. Where should transformation occur?
  4. Where should orchestration occur?

Identity

  1. Is identity centrally managed?
  2. How are users provisioned?
  3. How is cross-org authorization handled?

Security

  1. What regulatory boundaries exist?
  2. What data residency constraints exist?
  3. What information must remain isolated?

Governance

  1. What does the Salesforce COE control?
  2. What can individual org teams decide?
  3. How are architecture exceptions managed?

DevOps

  1. How are releases coordinated?
  2. How are shared dependencies tested?
  3. How are enterprise changes communicated?

Data 360

  1. Is a centralized Data 360 appropriate?
  2. Which org should be the home org?
  3. Which orgs should be companions?
  4. Are multiple Data 360 instances required?
  5. What data spaces should be shared?

Salesforce’s current guidance emphasizes selecting the Data 360 home org and companion orgs early, especially because changing a home-org configuration later can be a significant undertaking.

33. The Architecture Principle I Recommend

For enterprise architects, the following principle is useful:

Don’t optimize for the fewest Salesforce orgs. Optimize for the right business boundaries with the fewest unnecessary dependencies.

A three-org architecture with clear ownership and a strong enterprise data layer can be healthier than a single org overloaded with incompatible business processes.

Conversely, ten orgs with uncontrolled replication and point-to-point integrations can become an architectural liability.

The number of orgs is therefore not the architecture.

The boundaries, ownership model, integration model, data model, and governance model are the architecture.

34. How Data 360 Changes the Future Multi-Org Model

The traditional multi-org conversation was largely:

“How do we synchronize data between Salesforce orgs?”

The modern conversation increasingly becomes:

“How do we provide governed access to a unified enterprise data foundation while allowing each Salesforce org to remain operationally autonomous?”

That is a fundamentally different architectural model.

Data Cloud One supports this direction by allowing multiple Salesforce orgs to connect to a central Data 360 home org as companion orgs, with shared data spaces and metadata.

This is particularly relevant as enterprises expand their use of:

  • Data 360
  • Analytics
  • AI
  • Agentforce
  • Customer 360
  • Cross-org personalization
  • Enterprise decisioning

The architecture should therefore be designed with future data and AI requirements in mind, even if today’s requirement is only CRM.

35. Final Architecture Checklist

Before implementing or expanding a Salesforce multi-org landscape, verify that you can answer all of these questions:

  • Every org has a clearly documented business purpose.
  • Every org has an accountable business owner.
  • Data ownership is explicitly defined.
  • Enterprise customer identity is defined.
  • Master data ownership is documented.
  • Cross-org data-sharing requirements are documented.
  • Integration patterns are standardized.
  • Point-to-point integrations are minimized.
  • Synchronous and asynchronous interactions are distinguished.
  • Identity architecture is centrally governed.
  • Security boundaries are documented.
  • Data residency requirements are understood.
  • Enterprise reporting architecture is defined.
  • Salesforce COE responsibilities are documented.
  • Local-vs-global decision rights are clear.
  • Multi-org DevOps processes are defined.
  • Shared services are identified.
  • Data 360 strategy is documented.
  • Home and companion org decisions are deliberate where Data Cloud One is used.
  • AI and Agentforce implications have been considered.
  • The architecture has a defined evolution path.

Conclusion

A Salesforce multi-org strategy should never be reduced to:

“We have multiple Salesforce orgs.”

The real architecture is the system of boundaries and relationships surrounding those orgs.

A well-designed multi-org environment allows individual business units, regions, brands, or acquired companies to maintain the autonomy they genuinely need while providing the enterprise capabilities they genuinely share.

The strongest architectures establish:

Clear org boundaries.

Explicit data ownership.

Enterprise identity.

Governed integration.

Controlled data sharing.

Federated governance.

Shared enterprise services.

And increasingly:

A unified data and AI foundation.

The goal is not to eliminate every boundary.

The goal is to ensure that every boundary exists for a reason—and that the enterprise remains connected where connection creates value.

Related Salesforce Architecture Guides

Start here:

[Single Org vs Multi-Org Architecture Decision Guide]

Then explore:

  • Salesforce Multi-Org Architecture: Patterns, Governance & Data Sharing
  • Salesforce Org Consolidation & Migration Architecture
  • Salesforce Org Separation & Carve-Out Architecture
  • Salesforce Data Architecture Decision Guide
  • Salesforce Integration Architecture Decision Guide
  • Salesforce Architecture for Agentforce

Author

Mahesh Raja Vandyala

Enterprise Architect | Salesforce & Integration Architecture

Writing about enterprise Salesforce architecture, integration, data, AI, architecture decision frameworks, and scalable digital platforms.