What is a CRM data model? Objects and relationships
At some point after a CRM goes live, someone pulls a pipeline report and the numbers don’t add up. Marketing and sales end up using different definitions for the same term. A new integration breaks three sales pipeline reports that nobody knew shared the same data source.

本条来自 HubSpot Marketing Blog(marketing),聚焦 零售渠道、品牌营销、财报增长、设计创意。 Free CRM Software With Something for Everyone
At some point after a CRM goes live, someone pulls a pipeline report and the numbers don’t add up
- What is a CRM data model? Objects and relationships
- HUBSPOT'S FREE CRM SOFTWARE
- What is a CRM data model?
- HubSpot's Free CRM Software
- Free CRM Software With Something for Everyone
At some point after a CRM goes live, someone pulls a pipeline report and the numbers don’t add up
Marketing and sales end up using different definitions for the same term
A new integration breaks three sales pipeline reports that nobody knew shared the same data source
- Home
- Marketing
- What is a CRM data model? Objects and relationships
What is a CRM data model? Objects and relationships

HUBSPOT'S FREE CRM SOFTWARE
Free CRM Software With Something for Everyone

At some point after a CRM goes live, someone pulls a pipeline report and the numbers don’t add up. Marketing and sales end up using different definitions for the same term. A new integration breaks three sales pipeline reports that nobody knew shared the same data source.

Pipeline reporting discrepancies, conflicting data definitions, and broken integrations all trace back to the same root cause: a CRM configured without an intentional data model. According to Validity’s State of CRM Data report, 37% of CRM users have directly lost revenue due to poor data quality, and only 9% trust their data enough for confident reporting, which means the design work this guide covers is far more common a gap than most teams expect.
Today, I’ll guide you through what a CRM data model is, why they matter, and how to set up one that provides cross-team clarity over friction.
- What is a CRM data model?
- Why a CRM Data Model Matters for Data and CRM Outcomes
- Components of a CRM Data Model
- CRM Data Model Examples Using HubSpot Objects
- How to Design a CRM Data Model Step by Step
- CRM Data Model Patterns for B2B, B2C, and B2B2C
- Canonical Data Model vs CRM Data Model
- Govern and Optimize Your CRM Data Model
- How to Visualize and Document Your CRM Data Model
- AI Agents and Analytics Depend on a Clean CRM Data Model
- Signs Your CRM Data Model Is Working
- Frequently Asked Questions About CRM Data Models
What is a CRM data model?
A CRM data model is the structural blueprint that defines how customer data is organized within a CRM system. It specifies which objects exist, which properties each object has, how objects relate to one another, and the rules governing data entry and pipeline movement.
A CRM database stores the records; the data model defines what those records look like and how they relate to one another. Think of it the way you would a database schema: the schema tells the system what tables (objects) exist, what columns (properties) they hold, and how tables relate.
A CRM data model defines how customer data is structured in a CRM. It includes objects, properties, relationships, pipelines, activities, and IDs. A CRM database stores the records defined by a CRM data model.

HubSpot's Free CRM Software
Free CRM Software & Tools for Your Whole Team
- Sales
- Marketing
- Operations
- Customer Service
Why a CRM Data Model Matters for Data and CRM Outcomes
Accurate reporting, reliable handoffs, and clean pipelines all depend on a well-structured data model. According to research, 76% of organizations say less than half of their CRM data is accurate and complete, 37% have directly lost revenue from poor data quality, and only 9% of businesses trust their data enough for confident reporting.
I’ve seen this pattern repeatedly: a company imports a spreadsheet into its CRM, and six months later, nobody trusts the pipeline numbers. The root cause is almost always a data model that was never properly designed.
A well-structured model gives every team a shared vocabulary for customer data, so a deal handoff includes all the context the next team needs. For what to do when trust has already broken down, see HubSpot’s guide to fixing untrustworthy reporting data.
Components of a CRM Data Model
Six elements make up a CRM data model:
- Objects: the main categories of data, including contacts, companies, deals, tickets, and custom objects.
- Properties (fields): the individual data points stored on each object, such as email, lead source, or deal amount.
- Associations: the relationships between objects. Contacts can be associated with companies. Deals are associated with contacts and companies. Tickets can be linked to contacts and companies.
- Pipelines and stages: the defined progression of deals or tickets through a workflow. Pipelines track the stages of deals or tickets.
- Activities::calls, emails, meetings, and notes that capture interactions across CRM records.
- Unique IDs: system-generated identifiers that support deduplication, synchronization, and record matching across tools.
Governance rules (who can create fields, which naming conventions apply, and when to retire properties) form the structural layer that keeps objects, properties, associations, and pipelines working reliably over time.
CRM Data Model Examples Using HubSpot Objects
HubSpot’s Smart CRM is built around four standard objects: Contacts, Companies, Deals, and Tickets. Enterprise edition adds custom objects for business-specific entities that standard objects don’t cover (such as a Subscription, Location, or Project). Custom objects represent business entities that standard CRM objects don’t cover.
Individual people: leads, customers, partners
First name, email, lifecycle stage, lead source
Organizations that contacts belong to
Company name, industry, annual revenue
Revenue opportunities in a pipeline
Deal name, amount, close date, deal stage
Customer support cases
Subject, status, priority, ticket owner
Business-specific entities (Enterprise)
Subscription, Location, Project (examples)
HubSpot’s CRM customization tools let teams configure object names, required fields, and association labels directly in the UI, without requiring developer resources.
How Relationships and Associations Work
In HubSpot, associations determine which records connect to which.
- A contact can be associated with one or more companies.
- A deal links to both the contacts and company involved.
- Tickets associate with the contact who raised the issue.
Associations support labels, so you can distinguish a Decision Maker from a Technical Evaluator on the same deal, which is especially useful in B2B models where multiple stakeholders influence a buying decision.
How to Design a CRM Data Model Step by Step
HubSpot’s data model builder gives you a visual canvas to see, configure, and document your CRM structure in one place.
Step 1: Open the builder and read the canvas.
In your HubSpot account, go to Data Management > Data Model, then click Edit data model. Before changing anything, click each object card and watch which connections highlight. This view shows you what your model actually looks like, which is often different from what teams assume.
Step 2: Activate the objects you need.
Contacts, Companies, Deals, and Tickets are on by default. Optional objects (Appointments, Courses, Listings, Services) can be turned on inside the builder.
- Click Activate [object] on the canvas, then click Confirm.
- Once active, the object is available for associations, workflows, and segments.
- Administrators should deactivate any optional objects without a documented business use case. Unused objects clutter the canvas and confuse new users.
Best for: Teams performing a CRM migration. Activate only what you have real use cases for before moving data over. Adding objects later is simple; cleaning up records that landed in the wrong place isn’t.
Step 3: Add properties to each object.
Click any object on the canvas to expand its properties. Click + Create property, fill out the definition in the right panel, and save.
- Teams should use dropdown fields instead of free-text fields wherever possible to improve data quality. Every open-text field is a future data quality problem.
- Proceed to mark critical fields as required at specific pipeline stages, not account-wide.
- Teams should maintain a data dictionary that documents every property, its definition, owner, and acceptable values.
Pro tip: Breeze Assistant can create properties from plain-language prompts. Type something like “Create the contact property Secondary email” and it builds the field for you. Faster than the manual panel for large property sets.
Step 4: Configure associations.
Click the Associations tab in the left sidebar to manage how objects connect.
- Click + next to any object to add a new association, then select it from the dropdown.
- To add association labels (for example, Decision Maker versus Technical Evaluator on the same deal), click an existing association line, then click + Create association label.
- To limit how many records can be associated, click 1-to-many next to an association and set your limit.
Solid association design is what keeps customer data integration clean. If your integrations don’t respect your association structure, records end up orphaned.
Step 5: Add custom objects for business-specific entities.
When a core part of your business doesn’t fit a standard object, create a custom one. In the left sidebar, click + Create a custom object, fill out the right panel, and click Create.
Custom objects are available on Enterprise plans.
Common examples: Subscriptions, Locations, Projects, Contracts.
Pro tip: Pressure-test the use case with at least two teams before creating a custom object. Restructuring once records accumulate is expensive.
Step 6: Export, document, and validate your data model.
Before launch, run test records through every pipeline stage and confirm associations work the way you expect.
- Export a canvas snapshot: click the three-dot menu next to Data Model, then Export as an image — this becomes your ERD.
- Check model health: click View details on any object card. The Usage tab shows property fill rates, empty properties, and potential duplicates.
- Finalize your documentation: ERD, data dictionary, and a change log that records every modification with a date and reason.
Want to see how HubSpot’s data model fits your use case? Get a demo from the HubSpot team.
HubSpot's Free CRM Software
- Sales
- Marketing
- Operations
- Customer Service
CRM Data Model Patterns for B2B, B2C, and B2B2C
B2B CRM Data Models
B2B CRM models center on companies, contacts, and opportunities. Because a typical B2B buying group includes 6 to 10 decision-makers (per Gartner), B2B models need association labels and multi-contact deal tracking, not just a single contact per record. Customer lifecycle management across multiple stakeholders is central to B2B model design.
B2C CRM Data Models
B2C CRM models center on individual contacts and lifecycle data. Companies are often irrelevant. The key relationships are between a contact and their transaction history, subscription status, or lifecycle stage. B2C models prioritize high-volume performance and fast segmentation across millions of contacts.
B2B2C CRM Data Models
B2B2C CRM models often require intermediary entities such as partners or locations. A company that sells through channel partners needs to track the end customer, the partner organization, and the agreement between them. These intermediary entities typically become custom objects with explicit associations on both sides, plus their own consent and SLA properties.
Reliable customer data integration is especially important in B2B2C, where data flows across more systems and business relationships.
Canonical Data Model vs CRM Data Model
本条目归入「Brand Marketing」垂直,涉及真实话题:零售渠道、品牌营销、财报增长、设计创意。
· 市场:关注 零售渠道、品牌营销 对相关品类与竞争格局的潜在影响。
· 消费者:受众行为与偏好变化值得追踪。
· 品牌:本动向对品牌资产建设的启示。
· 渠道:内容分发与触点组合(社媒 / 电商 / 线下)的协同值得复盘。
· 核心话题:零售渠道、品牌营销、财报增长、设计创意。
· 可思考:如何把「零售渠道」的洞察,转化为可衡量的内容与增长动作?
面试中可引用「What is a CRM data model? Objects and relationships」:围绕 零售渠道、品牌营销、财报增长、设计创意,说明你对行业动向的判断与可落地动作。
本条目相关英文术语可在「商务英语」模块按话题检索,用于外企面试表达训练。
At some point after a CRM goes live, someone pulls a pipeline report and the numbers don’t add up. Marketing and sales end up using different definitions for the same term. A new i…
Pipeline reporting discrepancies, conflicting data definitions, and broken integrations all trace back to the same root cause: a CRM configured without an intentional data model. A…
First name, email, lifecycle stage, lead source…