Data quality: validation and duplicates

Clean data starts at entry. Validation rules block records that break business rules, picklists and required fields limit what can be typed, and duplicate and matching rules catch the same person or company being entered twice.

Validation rules

• How they work: A formula that returns true when the data is wrong. When it does, the save fails and the user sees your error message, at the top of the page or under a field.
• Where they apply: Every save: the user interface, imports, the API, and automation. They also block bad data loads, so plan for them during migrations.
• Common patterns: Require a field only in some cases (ISBLANK with a condition), keep a date in range, or check a format with REGEX.
• Bypass for integrations: A common pattern is a custom permission or field checked in the formula, so a data migration user can skip a rule.

Other entry controls

• Universally required fields: Marked required on the field itself, so every save must include a value, including through the API.
• Required on the layout: Only required on that page layout in the user interface. Imports and the API can still leave it blank.
• Picklists and restricted picklists: Stop free-text variations such as "USA", "U.S.", and "United States".
• Lookup filters: Limit which records can be chosen in a lookup field, for example only active accounts.

Duplicate management

• Matching rules: Define what counts as a duplicate, using exact or fuzzy matching on fields such as name, email, and phone. Standard matching rules exist for accounts, contacts, and leads.
• Duplicate rules: Decide what happens when a match is found: allow the save with an alert, or block it, separately for creates and edits.
• Reporting on duplicates: Duplicate rules can log matches as duplicate record sets, so you can report on them and merge them.
• Merging: Users with permission can merge duplicate accounts, contacts, leads, and cases, choosing which values to keep.