Customer Data Model
Design CRM objects, required fields, and data quality rules.
Key takeaways
- The data model defines the unit of customer reality the company can see; a vague model lets segments, automation, and dashboards drift.
- Core objects include Account, Contact, Lead, Opportunity, Activity, Subscription/Order, and Consent, each with example fields.
- Make a field required only when it drives an operating decision, such as lead source when channel quality is measured.
- Data quality rules: one field means one thing, integrated fields need source-of-truth and sync direction, and duplicate merge rules must be written down.
- Document each field in a data dictionary with type, allowed values, source, owner, and where it is used.
The CRM data model defines what unit of customer reality the company can see. If the model is vague, segments, automation, dashboards, and accountability all drift.
Core Objects
| Object | Meaning | Example fields |
|---|---|---|
| Account | Company, household, store, or organization | Industry, size, tier, contract status |
| Contact | Person or user | Role, email, consent, preferred channel |
| Lead | Demand not yet validated as an opportunity | Source, interest, fit, recent behavior |
| Opportunity | Purchase or adoption opportunity | Amount, stage, expected close, decision maker |
| Activity | Interaction with customer | Meeting, email, call, ticket, campaign response |
| Subscription/Order | Contract or purchase record | Product, amount, term, renewal date |
| Consent | Privacy or marketing consent | Consent type, source, timestamp, withdrawal |
Required Field Rule
A field is required only when it drives an operating decision.
| Field candidate | When it should be required |
|---|---|
| Customer name | Almost always |
| Email/contact | Required when outreach is part of the process |
| Lead source | Required when channel quality will be measured |
| Industry/size | Required when ICP fit is used |
| Product interest | Required for routing or personalization |
| Owner | Required when SLA exists |
| Marketing consent | Required when campaigns are sent |
Data Quality Rules
- One field should mean one thing.
- Dropdowns should be granular enough for reporting, but not more.
- Free text is useful for discovery but weak for automation.
- Integrated fields need source-of-truth and sync direction.
- Duplicate merge rules must be written down.
Data Dictionary Template
| Item | Example |
|---|---|
| Field | lifecycleStage |
| Description | Customer lifecycle stage |
| Type | select |
| Allowed values | Lead, Qualified, Opportunity, Customer, Active, Risk, Dormant |
| Source | CRM input or automation rule |
| Owner | CRM Owner |
| Used by | Dashboard, automation, weekly pipeline review |
Common Problem
CRM data problems are discovered during interpretation, not during input. Every additional field adds quality cost, so start with the minimum data needed for decisions.