一、认识统一客户画像
Unified Profiles • Individual DMO Types
Updating Profiles • Golden Records • Key Ring Analogy
学习目标
After completing this unit, you will be able to:
• Explain the benefits of unified profiles
• List the types of Individual data model objects
• Describe the difference between unified profiles
and golden records
Data 360 如何协调数据——统一画像概述
Data 360 HARMONIZES customer data across multiple systems
into unified profiles. This module explores the key data
unification concepts that help you make the most of Data 360:
• Unified profiles — linking data from disparate sources
• Data modeling — the Customer 360 Data Model
• Identity resolution — rulesets, match rules, and
reconciliation rules
• Data mapping requirements — what must be mapped for
identity resolution to work
Ready to get started?
数据与身份——概述与准备工作
First, let us watch an overview of data and identity.
[Image: Note]
Want to learn more about how to create your company’s
data strategy? Check out the Trailhead module:
[Customer Data Platform Strategy]
(https://trailhead.salesforce.com/en/content/learn/modules/customer-data-platform-strategy)
统一画像——链接多源数据创建单一客户视图
Unified profiles in Data 360 LINK data from MULTIPLE sources
into a single profile based on user-configured identity
resolution rules within a ruleset.
MEET RACHEL RODRIGUEZ, a customer and super fan of Northern
Trail Outfitters (NTO). NTO has data about Rachel in multiple
systems: Commerce, Marketing Engagement, Service, and more.
[Image: Rachel and the information we know about her from various sources.]
Each system has DIFFERENT information about her — different
email addresses, phone numbers, usernames. These unique pieces
of data are called CONTACT POINTS (phone, email, address).
For a marketer or service rep, connecting the dots to send
a campaign or find a single view of her support history
can be extremely tricky.
Data 360 统一画像的三层架构
That is where Data 360 data mapping and identity resolution
comes in. The process works in THREE layers:
1. MAP data to a standardized set of objects and fields
using the Customer 360 Data Model
2. CREATE an identity resolution ruleset with MATCH rules
(how to find relationships) and RECONCILIATION rules
(which data to prioritize)
3. Data 360 FINDS relationships between data based on
these rules. If the same data exists in multiple places,
profiles are linked together.
[Image: Unified individual ID for Rachel, with a single view of all her information.]
With identity resolution rules in place, NTO’s view of Rachel
includes a unified profile with data from ALL sources. You can
explore this in Profile Explorer. Rachel’s unified profile
UPDATES as new profiles are matched or existing ones change.
三种 Individual DMO 类型详解
DMOs are groupings of data in the Customer 360 Data Model
that describe an instance of a thing or action. Each DMO
has ATTRIBUTES — standardized pieces of data.
There are THREE types of individual DMOs. Let us look at
them through each phase of Rachel’s journey.
PHASE 1 — INGESTION: Rachel’s profile is ingested into
Data 360 and mapped to the INDIVIDUAL DMO.
INDIVIDUAL DMO:
• Contains source data ingested into Data 360
(e.g., Rachel’s Commerce profile is one instance)
• You know exactly which data stream the data came from
• Attributes: Individual Record ID, Data Source ID,
Data Source Object, First Name, Last Name, plus
other ingested and mapped values
• You have NO knowledge of the unified profile at this stage
身份解析运行后——Unified Link 与 Unified Individual
Identity resolution runs. Match rules LINK Rachel’s Commerce
profile to the unified profile. Reconciliation rules COMBINE
individual attributes with the unified attributes.
UNIFIED LINK INDIVIDUAL DMO:
• The JOINING POINT between source data and the unified
individual — you can traverse data in either direction
• Attributes: Individual Record ID, Data Source ID,
Data Source Object, Unified ID
UNIFIED INDIVIDUAL DMO:
• Contains RECONCILED data from all linked individuals
• A quick glance at sample values, NOT comprehensive
• NO data lineage — you cannot trace back to source
• Attributes: Unified ID (Individual ID), Reconciled
First Name, Reconciled Last Name, other reconciled values
The unified profile IS the combination of BOTH the unified
link individual AND unified individual. Together, you have
access to source data AND reconciled data.
统一画像——链接多源数据创建单一客户视图
Unified profiles are MUTABLE — they change to improve
accuracy. A unified profile updates when:
• Source data changes
• New data sources are processed
• You modify identity resolution rules
EXAMPLE: Rachel has a Sales profile where the agent
misspelled her name as “Rochelle.”
Rachel’s Sales Profile:
• First Name: Rochelle (misspelled!)
• Last Name: Rodriguez
• Email: rrodriguez@example.com
ACTIVE MATCH RULE: Fuzzy First Name + Exact Last Name
+ Exact Email Address
“Rochelle” does NOT qualify as a fuzzy match with “Rachel”
→ Rachel ends up with TWO separate unified profiles.
Rachel gets the name fixed. On the NEXT identity resolution
run, her Sales profile LINKS to her unified profile. Now
she has ONE unified profile with ALL her data.
统一画像 vs 黄金记录——Data 360 与 MDM 的关键区别
Master Data Management (MDM) aims for a consolidated “best”
record called the GOLDEN RECORD. How does Data 360’s unified
profile compare?
UNIFIED PROFILE (Data 360):
• Links records into a single view WITHOUT overwriting sources
• Does NOT attempt to select a single “best” value
• ALL source data is retained and independent
• Data lineage is INTACT — you can trace back to origin
• Faster to implement and scale
• The unified ID CAN change over time
BOTH:
• Consolidate multiple records into a single record
• Assign unique identifiers (unified individual ID)
GOLDEN RECORD (Traditional MDM):
• Attempts to select the “best” single value — often no
“best” exists, so the record becomes oversimplified
• Stakeholders must AGREE on selection criteria — slow
• Source data is OVERWRITTEN — data lineage is LOST
• The unique identifier is PERMANENT
统一画像——链接多源数据创建单一客户视图
Think of a unified profile as a KEY RING.
[Image: Diagram of a key ring with house, SUV, and truck keys.]
A key ring LINKS your keys together (house key, car key).
It does NOT turn all keys into the same key or choose a
“best” key. Instead, it ORGANIZES them into a single object
you can grab easily. Keys can be moved between key rings.
[Image: Diagram of a unified profile as a key ring.]
Similarly, a unified profile LINKS your IDs from across
Salesforce. Each linked ID stays UNIQUE and you gain
access to data across all of Salesforce.
Marketing Engagement tracks subscriber details by contact
point, but its data model does not support more than one
email or phone per contact. If a customer interacts with
multiple contact points, you get MULTIPLE subscriber IDs.
Data 360 links all of them to ONE unified profile, so you
can access the contextual data for each subscriber ID.
二、创建统一画像
Implementation Steps • From Raw Data to Unified Profile
Discuss Your Data — Whiteboard Questions
学习目标
After completing this unit, you will be able to:
• Describe how to create unified profiles step by step
• Describe how to analyze your data and get it ready
for unified profiles
创建统一画像的四步旅程
It is helpful to understand these steps and concepts BEFORE
you begin data modeling and mapping. Here is the journey
from raw data to a unified profile.
STEP 1 — INGEST raw data from data sources:
• Data is added from bundles, data extensions, Amazon S3,
and other systems AS IS
• After raw data is added as a data stream, it needs
to be mapped to the data model
STEP 2 — MAP AND MODEL data:
• The Customer 360 Data Model standardizes data from
multiple sources into a readable, mappable format
• Data streams must be mapped to objects like Party
Identification and Individual for identity resolution
rulesets to work
STEP 3 — CREATE identity resolution rulesets:
• Match rules tell identity resolution what types of
data to MATCH across data streams
• Reconciliation rules tell it how to UNIFY profiles
STEP 4 — CREATE AND USE unified profiles:
• When identity resolution runs, it creates unified
profiles usable for segmentation and activations
• Add activation filters to filter audience members
based on their unified attributes
讨论你的数据——白板会议关键问题
[Image: A team gathered around a table and whiteboard to discuss data mapping.]
Now that you understand the concept behind unified profiles,
what is next? To be successful, spend time ANALYZING the
data you want to use in Data 360.
GRAB YOUR TEAM, A WHITEBOARD, AND DISCUSS:
• Where is your data located? (spreadsheets, S3, CRM,
Marketing Engagement, etc.)
• Do you have an asset inventory for each data source?
• How do you identify individuals in each source?
(email, name, birthday, system ID?)
• Do you use contact keys, lead IDs, or subscriber keys
as unique system identifiers?
• What data is SHARED across systems?
• What does your customer journey look like?
• What data do you truly NEED for audience segmentation?
• How is the DATA QUALITY in each source? Misspellings?
Missing birthdays, phone numbers, or other fields?
DO NOT SKIP THIS PART! Understanding your data is KEY
to a successful Data 360 implementation.
三、映射必需对象
Customer 360 Data Model Components • Required Mappings
Individual • Contact Points • Party Identification
学习目标
After completing this unit, you will be able to:
• Recognize Customer 360 Data Model components
• Describe individual, contact point, and party objects
• Identify mapping requirements for identity
resolution rulesets
Customer 360 数据模型组件回顾
The Customer 360 Data Model is Data 360’s standard data
model that makes data INTEROPERABLE — usable wherever you
need it. Understanding these components makes data mapping
and identity resolution setup much easier.
[Image: subject area diagram]
SUBJECT AREA (A Business Goal):
Groups DMOs according to business goals — marketing, product
support, etc. Subject areas include Party (unique identifiers),
Engagement data, Sales Orders, or Product information.
DATA MODEL OBJECT (DMO) (Groups of Data):
An object created by ingested data streams and insights.
Can be STANDARD or CUSTOM. Stores data such as leads,
product info, customer info, and so on.
ATTRIBUTES (Data About Your Contacts):
Unique bits of information about a contact from different
sources. These link an individual’s data together to build
a unified profile. For marketers, attributes are GOLD —
use them to create fine-tuned segments (e.g., contacts
under 25 who prefer weight lifting to running).
数据映射的重要性——身份解析的前提条件
In order to create unified profiles, you MUST map your
data correctly. Data 360 is like AI — it requires QUALITY
data and some human intervention to be most effective.
The system can ONLY unify profiles if they are mapped
correctly to TWO things:
1. The INDIVIDUAL OBJECT (required)
2. At least ONE other element:
• A CONTACT POINT OBJECT (email, phone, address, app)
— OR —
• A PARTY IDENTIFIER OBJECT
These are the minimum mapping requirements for identity
resolution to work. Let us examine each one in detail.
Individual 对象——最重要的映射对象
The INDIVIDUAL object is the most important — it holds all
the personal information you know about your customer.
CRITICAL RULE: Every data stream with customer information
MUST have a field mapped to the Individual ID field from
the Individual object in order to use identity resolution.
REQUIRED MAPPINGS FOR INDIVIDUAL:
• Individual ID (primary key) — ssot__Id__c
• First Name — ssot__FirstName__c
• Last Name — ssot__LastName__c
MAP TO relationships:
• Individual.ID → ContactPointAddress.Party
• Individual.ID → ContactPointApp.Party
• Individual.ID → ContactPointEmail.Party
• Individual.ID → ContactPointPhone.Party
• Individual.ID → PartyIdentificationId.Party
You can also map ADDITIONAL fields from your customer data
to standard and custom attributes — birth date, preferences,
and any other data that helps build a complete profile.
The Individual ID is the PRIMARY KEY for the Individual
object and is REQUIRED for mapping and identity resolution.
Contact Point 对象——Email、Phone、Address、App
Contact points (email, phone, address, device, social) all
have associated objects usable for identity resolution.
Similar to Individual ID, a Contact Point ID serves as the
PRIMARY KEY and is REQUIRED for mapping and identity
resolution. Whatever field in your data stream uniquely
identifies the customer should map to this ID.
CONTACT POINT ADDRESS:
• Primary Key: Contact Point Address Id (ssot__Id__c)
• Fields: Address Line 1, City, Party, Postal Code,
State Province
• Maps To: ContactPointAddress.Party → Individual.ID
CONTACT POINT APP:
• Primary Key: Contact Point App ID (ssot__Id__c)
• Maps To: ContactPointApp.Party → Individual.ID
CONTACT POINT EMAIL:
• Primary Key: Contact Point Email ID (ssot__Id__c)
• Fields: Email Address, Party
• Maps To: ContactPointEmail.Party → Individual.ID
CONTACT POINT PHONE:
• Primary Key: Contact Point Phone ID (ssot__Id__c)
• Fields: Formatted E164 Phone Number, Party
• Maps To: ContactPointPhone.Party → Individual.ID
Party Identification 对象——客户提供的标识符
Party identifier matching allows you to use your own
CUSTOMER-SUPPLIED IDENTIFIERS. This is especially important
with Marketing Engagement data bundles.
REQUIRED FIELDS:
• Party Identification ID (primary key) — any unique ID
from your customer data (ssot__Id__c)
• Party — foreign key, same as used in Individual object
(ssot__PartyId__c)
• Party Identification Type (required for mapping,
optional for identity resolution) — describes the
identifier, e.g., “Social” — BE DESCRIPTIVE, it is used
in match rule setup (ssot__PartyIdentificationTypeId__c)
• Identification Number — the ID used for identity
resolution comparison (ssot__IdentificationNumber__c)
• Identification Name (required for mapping) — name
of the ID space, e.g., “Mobile ID” or “LinkedIn ID”
— also used in match rule setup (ssot__Name__c)
Map To: PartyIdentificationId.Party → Individual.ID
DRIVER’S LICENSE EXAMPLE:
• Party Identification ID: 100a
• Party: 10016-00001
• Type: Driver License
• Identification Number: D1469256
• Name: CA Driver ID
数据映射案例——NTO 忠诚度计划与 Party 关系
PARTY RELATIONSHIPS use Many-to-One cardinality — you can
have MULTIPLE party fields mapped to ONE Individual object.
For example, three party identification types: LinkedIn,
Contact ID, and Marketing Cloud Subscriber Key.
[Image: Exact party identification configuration.]
NTO DATA MAPPING EXAMPLE:
To help visualize, let us review NTO’s data model. The data
stream “NTO Loyalty Program” (1) is mapped to the
Individual object (2) and Contact Point Phone object (3).
NTO created this mapping to use phone numbers as a match rule.
[Image: Mapping of Data source to contact point phone and individual.]
THREE REQUIRED PRIMARY KEY MAPPINGS in this example:
• SubscriberKey (from the data source)
• Contact Point Phone ID
• Individual ID
The SubscriberKey from the data source is mapped to the
other primary keys AND to Party — because the Party field
helps provide the RELATIONSHIP between objects and primary
keys found in data sources.
下一步：身份解析规则集——三大映射要求回顾
Now that you understand the importance of data mapping
and the required objects, we are ready to explore how
MATCH RULES and RECONCILIATION RULES work inside identity
resolution rulesets.
RECAP — THE THREE MAPPING REQUIREMENTS:
1. Individual Object — maps ALL customer data streams
to the Individual ID (REQUIRED for every stream)
2. Contact Point Object — email, phone, address, or app
(at least ONE required)
3. Party Identification Object — customer-supplied
identifiers (alternative to contact points)
Get your mappings right, and identity resolution can work
its magic. In the next unit, we look at how to configure
the rules that actually link profiles together.
四、理解身份解析规则集
Rulesets • Match Rules (Exact, Fuzzy, Normalized)
Reconciliation Rules • Object and Field Level
学习目标
After completing this unit, you will be able to:
• Define rulesets and their purpose
• Create match rules using exact, fuzzy, and
normalized methods
• Use reconciliation rules to prioritize data
身份解析规则集——Match 与 Reconciliation 规则
RULESETS allow you to configure MATCH rules and
RECONCILIATION rules about a specific object, such as
Individual. The system follows these rules to link
together multiple sources of data into a unified profile.
[Image: Ruleset Properties]
Regardless of what objects you use, carefully REVIEW your
data requirements to make sure your source data COMPLIES
with the mapping requirements BEFORE ingestion.
It is far EASIER to fix a data stream BEFORE ingestion
than to update it AFTER ingestion.
Let us review your match rule options so you can make an
informed decision about what will work for your account.
匹配规则——精确、模糊与规范化三种方法
Match rules are CUSTOMIZABLE based on your business needs.
[Image: Match rule criteria — object, field, and match method.]
To create a match rule, you select THREE things:
1. OBJECT (1) — from Individual, Contact Points (email,
app, phone, address), Device, or Party Identification
2. FIELD (2) — attributes available based on the object
you selected
3. MATCH METHOD (3) — how to compare values:
• EXACT — matching on an exact match. No typos or
alternative formats allowed.
• FUZZY — matching on a similar match. Typos and
slightly different spelling are OK. Available ONLY
for First Name.
• NORMALIZED — matching on the same exact info,
regardless of formatting. Available for Email,
Phone, and Address.
You can create COMBINATIONS of match rules based on
standard and custom attributes. Give each rule a DESCRIPTIVE
name like “Fuzzy First Name and Custom Field 2.”
The MORE rules you configure, the MORE mapping
requirements you must follow.
协调规则——Last Updated、Most Frequent 与 Source Sequence
While match rules LINK data together, RECONCILIATION RULES
determine the logic for which data gets DISPLAYED in the
unified profile when the same value exists in multiple sources.
THREE RECONCILIATION RULE OPTIONS:
LAST UPDATED:
• The most recently updated value is selected
• Consider: what data gets updated most regularly?
Customer service data or marketing preference data?
MOST FREQUENT:
• The most frequently occurring value is selected
• If 4 out of 5 sources say “Rachel” and 1 says “Rochelle,”
“Rachel” wins
SOURCE SEQUENCE:
• You SORT your data sources in order of most to least
preferred for inclusion
• This is based on your CONFIDENCE in each data source
• Example: use Commerce data FIRST, S3 data LAST
Reconciliation rules can be set at BOTH the object level
and the individual FIELD level. What happens to data you
do NOT select? All unique values are STILL STORED for each
customer — you never lose important data.
下一步：NTO 实战案例——Rachel 的六条记录
In the next unit, we apply everything we have learned to
a REAL USE CASE. We will see how NTO configures identity
resolution rulesets for Rachel Rodriguez, walking through
her six disconnected records and the specific match rules
that bring them together into one unified profile.
RECAP — MATCH VS. RECONCILIATION:
• MATCH rules determine WHICH profiles get LINKED together
• RECONCILIATION rules determine WHICH VALUES get
DISPLAYED in the unified profile
Both are essential. Match rules build the connections;
reconciliation rules make the connections useful.
五、配置身份解析规则集
Create Rulesets • NTO Use Case (Rachel’s 6 Records)
Identity Resolution Home Page • Best Practices Wrap-Up
学习目标
After completing this unit, you will be able to:
• Create identity resolution rulesets step by step
• Follow best practices for identity resolution
configuration and monitoring
身份解析规则集——Match 与 Reconciliation 规则
With requirements in hand, you can begin data ingestion
and modeling. After your data has been ingested and mapped,
you are ready to create up to TWO identity resolution
rulesets from the Identity Resolutions tab.
QUICK STEPS:
1. Click New
2. Select the Individual entity
3. Add an optional ruleset ID of up to four text
characters (once set, it CANNOT be changed)
4. Click Next
5. Name your ruleset and add an optional description
to help identify its properties
6. Click Save
[Image: Note]
Once published for the FIRST time, your account creates
unified profiles within 24 hours based on the rules
assigned. After initial creation, any changes made to
your rules are processed on a DAILY basis.
NTO 实战案例——Rachel Rodriguez 的六条记录连接分析
Rachel has 6 TOTAL RECORDS across NTO’s systems: Agentforce
Service, Marketing Engagement, Commerce, and Loyalty
Management. Let us review the connections to determine
what match rules can unify her profile.
[Image: Rachel and her information from various sources.]
MATCHING CONNECTIONS ACROSS RECORDS:
• Records (1) & (4) — fuzzy first name “Racheal” vs.
“Rachel” + exact address (Joachimsthaler Str. 1-4, Berlin)
• Records (2) & (6) — fuzzy first name “Rachele” vs.
“Racheal” + exact party identifier (Twitter: NTOfanRachel)
• Records (3) & (6) — exact name + exact email address
• Records (4) & (5) — exact name + normalized phone number
• Records (5) & (6) — fuzzy name “Rachel” vs. “Racheal”
+ work email (rachel@mystyle.com)
匹配规则——精确、模糊与规范化三种方法
[Image: Unified individual ID for Rachel, with all her information.]
Here are the MATCH RULES NTO used to create Rachel’s
unified profile. Each rule uses a COMBINATION of criteria
connected by AND, with rules connected by OR:
• Fuzzy first name AND exact normalized address
— OR —
• Fuzzy first name AND exact party identifier
— OR —
• Exact first name AND exact normalized email
— OR —
• Exact first name AND exact normalized phone
The OR logic means ANY rule that matches will link the
profiles. Each rule combines two criteria with AND
to increase match confidence. The fuzzy matching on first
name is critical — without it, records with misspelled
names like “Rochelle” or “Racheal” would never connect.
This flexible combination of rules gives NTO the coverage
needed to link records across all their systems while
maintaining match accuracy.
Identity Resolution 监控主页——查看处理历史与优化
After configuration, you can MONITOR and EDIT your rulesets
from the Identity Resolutions tab.
[Image: Identity Resolution page with properties, details, and processing history.]
From the individual ruleset page, you can view:
• Ruleset PROPERTIES
• Ruleset DETAILS
• PROCESSING HISTORY (1) — including last processing date,
individual sources, matching rates, and consolidation rates
KEY INSIGHTS FOR OPTIMIZATION:
• To INCREASE the consolidation rate → add MORE match rules
• To REDUCE the consolidation rate → remove SOME match rules
• Match rules CAN be edited after creation — monitor your
results regularly and iterate
Use the TWO available rulesets for A/B TESTING — compare
different match and reconciliation strategies side by side.
五大最佳实践总结
As you begin using Data 360 and identity resolution tools,
follow these BEST PRACTICES:
• Use RULESETS to COMPARE AND TEST — once your first
ruleset is configured, create a second to conduct A/B
tests on match and reconciliation rules
• Consider the ACCURACY AND CLEANLINESS of your data
— can you clean any data BEFORE importing it into
Data 360? Clean data makes better matches.
• Determine matching rules and requirements BEFORE
you begin mapping your data — the rules drive the
mapping requirements, not the other way around
• Unified profiles are only as trustworthy as the
SOURCE SYSTEM DATA — identify which data source has
the most up-to-date information and use that to guide
your match and reconciliation rules
• REVIEW your unified profiles REGULARLY from Profile
Explorer — see if tweaks need to be made to your ruleset
With the right data mapping and identity resolution
configuration, Data 360 is set up to help you get the
MOST out of your data.