A customer may appear as a contact in Sales Cloud, a case requester in Service Cloud, a subscriber in Marketing Cloud, a website visitor, a buyer in commerce, and a record in an external system. Each record can be valid on its own. The challenge is understanding when they describe the same person or business.

Salesforce Data 360 identity resolution helps address that challenge by matching records into unified profiles and determining which source values to use when data conflicts. Done thoughtfully, this can improve reporting, audience segmentation, customer context, and the data available to AI. Done carelessly, it can merge distinct customers or leave important duplicates unresolved.

This article is part of our detailed guide to the Salesforce Data 360 product family. We’ll explain how Salesforce identity resolution works, what to consider when setting up rules, and how unified profiles can support other Data 360 use cases.

What Salesforce Data 360 identity resolution does

Identity resolution has two related jobs:

1. Matching: Determines which source records represent the same entity and should be grouped together.
2. Reconciliation: Determines which value to show for each field when the matched records disagree.

The result can be a Unified Individual or Unified Account, depending on the entity being resolved. These unified records give teams a consolidated view while retaining the relationship to the source records used to create them. Identity resolution does not, by itself, clean up or overwrite the original records in Sales Cloud or other connected systems.

Matching across devices, channels, or anonymous website activity depends on available identifiers and the way data is configured. An anonymous session does not become attributable to a known customer without a suitable signal to connect the records.

Salesforce matching rules: exact, fuzzy, and normalized criteria

A match rule is a set of criteria used to identify records that may belong to the same person or account. Rules can be designed for a particular data model object, such as Individual or Account, and can contain one or more criteria. Salesforce’s documentation describes matching and reconciliation as distinct steps in identity resolution rulesets.

Common matching approaches include:

• Exact matching: Compares values that are expected to be identical, such as a stable customer identifier. It can be reliable when the identifier is consistently populated and governed.
• Fuzzy matching: Looks for likely matches when values vary, such as misspellings in a name. It can help with inconsistent data, but broad matching can increase false positives.
• Normalized matching: Compares standardized values, helping account for differences in formatting. Standardizing addresses or phone numbers before matching can reduce avoidable variation.

A rule can use several criteria together (for example, name and email) so the match depends on the combined conditions rather than on a single loose resemblance. Multiple rules can provide alternative matching paths when the business has more than one trustworthy identifier. The right combination depends on the data and the consequences of an incorrect merge.

For consumer identity, a business might use email plus normalized phone as one path, with a more cautious fuzzy name-and-address path for records that lack those identifiers. For accounts, a stable business ID or verified domain may be more useful than a company name alone.

Reconciliation rules: deciding which value wins

After Data 360 groups records, reconciliation rules select the value to display for each field in the unified profile. Different fields can take values from different source records. For example, the profile could use a current email from one system and a preferred name from another.

Three common approaches are:

1. Last Updated: Selects the value from the most recently updated source record. This can suit fields where recency is a useful signal, provided source timestamps are dependable.
2. Most Frequent: Selects the value that appears most often across the matched records. Salesforce documents tie-breaking behavior for this option in its Most Frequent reconciliation guidance.
3. Source Priority: Selects the value from a source ranked as authoritative for that field. This is useful when, for instance, a CRM owns account status while a commerce system owns the latest order-related details.

These approaches are not interchangeable. Last Updated prioritizes recency, Most Frequent prioritizes agreement, and Source Priority prioritizes governance decisions about system ownership. Teams should choose field-level behavior with those distinctions in mind. For example, the marketing platform may be authoritative for subscription preferences, while the CRM is the source of truth for an account owner.

Rulesets: make identity resolution reusable

A ruleset brings match and reconciliation choices together for a defined identity-resolution use case. This makes the configuration easier to review and reuse, and allows different parts of the business to use different approaches when their data and risk profiles differ.

A strict B2B account ruleset might rely on a verified identifier or carefully validated combinations of domain, legal name, and address. A B2C consumer ruleset might allow more variation in names and addresses while relying on stronger confirmation from email or phone. Separate rulesets can help avoid applying a consumer-oriented matching approach to complex business accounts.

Need help reviewing how your rules should work? Schedule a Call to discuss your Data 360 identity-resolution use case.

Six practical safeguards for better resolution quality

1. Assess data quality first. Review missing identifiers, inconsistent formatting, shared email addresses, stale records, and duplicate patterns before tuning match criteria. Matching rules cannot compensate for every source-data problem.
2. Test on representative records. Include common cases and edge cases: changed emails, shared company phone numbers, duplicate contacts, subsidiaries, and records with incomplete addresses.
3. Watch for over-matching. An overly broad fuzzy rule can group distinct people or organizations. This can distort reporting, expose the wrong context to a user, or create an unsuitable basis for activation.
4. Watch for under-matching. Rules that are too restrictive can leave duplicate records separate, limiting the usefulness of the unified profile and downstream audience counts.
5. Govern rule changes. Document the purpose of each rule, its owner, expected effect, and approval path. Re-test changes before using them for business-critical segmentation or activation.
6. Measure results against reviewed data. Evaluate a sample of proposed matches against records reviewed by knowledgeable staff. Track false matches, missed matches, resolution coverage, and downstream effects, not just the number of profiles unified.

The appropriate balance depends on how your organization uses the result. A false match can be more consequential than an unresolved duplicate in some use cases, while a marketing audience may prioritize broad coverage with clear eligibility controls. Establish the risk tolerance with business owners before expanding a ruleset.

B2B identity resolution: accounts are more than company names

B2B data introduces relationships that a simple company-name match may miss. One organization can have a parent account, subsidiaries, multiple ship-to locations, several contacts, and relationships with distributors or dealers. At the same time, the same contact may interact with more than one account.

It helps to separate two questions:

• Are these records the same account or individual?
• How are these distinct accounts related in the business hierarchy?

Identity resolution can help group records that refer to the same entity; it does not automatically mean that related subsidiaries or ship-to locations should become one account. Preserve meaningful distinctions, and decide whether hierarchy, location, and dealer relationships need separate modeling. For distributors and dealers, teams may need to resolve both the business entities and the individual contacts associated with them.

How unified profiles support Data 360 use cases

Unified Individuals and Unified Accounts can provide a more consistent basis for downstream work:

• Segmentation: Build audiences using a consolidated view of customer attributes and interactions.
• Calculated insights: Aggregate activity or measures across data that would otherwise be split between separate records.
• Activation: Send appropriate segments to connected channels, subject to data governance and activation rules.
• Agentforce grounding: Give AI experiences better-organized customer context to draw on, rather than relying on fragmented records.

Unified profiles are one part of the wider Data 360 foundation, not a guarantee that every downstream result is correct. Data access, permissions, freshness, and business logic still matter. For more on Salesforce AI capabilities, visit our Salesforce Artificial Intelligence page. Salesforce also describes Data 360 as a source of unified context in its U.S. AIforce announcement.

Plan identity resolution around business outcomes

Salesforce identity resolution is a configuration discipline as much as a technical capability. Matching rules determine which records are grouped; reconciliation rules decide which field values represent those groups. Data quality, B2B modeling, testing, and ongoing governance influence whether the result is useful and trustworthy.

At CloudStreet, we design matching and reconciliation strategies, assess data quality, and validate resolution results against real customer data. For a detailed look at our delivery approach, see Salesforce Data 360 Implementation Services. We’re based in Houston, Texas, and work with customers locally and globally. If you’re evaluating Salesforce identity resolution, get a quote or contact our team to discuss your Salesforce environment and goals.