The Data Governance Framework Your Salesforce Needs

Data governance sounds like a corporate buzzword. In practice, it means: "rules that prevent your Salesforce from turning into a dumpster fire."

Without governance, every day makes your org slightly worse. A rep saves a Contact without an email. An integration imports 500 Leads with "test" in the company name. Someone creates a new picklist value called "Othr" because "Other" was already taken. Multiply this by 50 users, 3 years, and 5 integrations, and you have the average Salesforce org: 40% of data is incomplete or duplicated, nobody trusts the reports, and the admin spends half their time fixing preventable problems.

Layer 1: Validation Rules (Prevent Bad Data at Entry)

Validation rules are the front door security for your data. They prevent records from being saved unless specific criteria are met.

The 5 validation rules every org needs:

1. Email format on Contact and Lead. Formula: NOT(REGEX(Email, '[a-zA-Z0-9._%+-]+@[a-zA-Z0-9.-]+\\.[a-zA-Z]{2,}')) — prevents "test@test" and blank-but-not-empty email fields.
2. Phone minimum length on Contact and Lead. Formula: LEN(Phone) > 0 && LEN(Phone) < 10 — prevents partial phone numbers like "555" that are useless for outreach.
3. Amount required on Opportunity when Stage is past Prospecting. Formula: NOT(ISPICKVAL(StageName, 'Prospecting')) && ISBLANK(Amount) — prevents pipeline reports from including deals with no dollar value.
4. Close Date required on Opportunity. Formula: ISBLANK(CloseDate) — prevents Opportunities with no Close Date from corrupting pipeline-by-quarter reports.
5. Closed Lost Reason required when Stage = Closed Lost. Formula: ISPICKVAL(StageName, 'Closed Lost') && ISBLANK(Closed_Lost_Reason__c) — captures why deals are lost.

These 5 rules prevent roughly 60% of the data quality issues we see in typical orgs. They take 30 minutes to configure.

Layer 2: Required Fields on Page Layouts

Validation rules enforce data at the point of save. Required fields on page layouts enforce data at the point of creation — they mark the field with a red asterisk and prevent the record from being saved if it's empty.

Best practice: Make fields required on the page layout for the initial creation, and use validation rules for conditional requirements (like requiring Amount only after a deal passes Prospecting stage).

Layer 3: Picklist Standardization

Picklist values drift over time. Someone adds "Othr" because "Other" is taken. Another person adds "Website" and "Web Site" as separate values for Lead Source. The result: your reports group by picklist values, and the same category appears 3 times with slightly different names.

The fix:

1. Export all picklist values for each critical field (Lead Source, Industry, Stage, Case Reason, Closed Lost Reason).
2. Identify duplicates, misspellings, and unnecessary variations.
3. Merge and standardize. Use Data Loader to update existing records to the standardized values.
4. Remove the deprecated values from the picklist (Setup → Object Manager → Field → Values → Deactivate).
5. Restrict picklist values — check "Restrict picklist to the values defined in the value set." This prevents users from creating new values via API.

Layer 4: Field Hygiene

Custom fields accumulate like clutter. Someone creates a field for a one-time project, the project ends, and the field stays. Three years later, you have 200 custom fields on the Account object and nobody knows what 150 of them do.

The quarterly field audit:

1. Run the Field Usage report — identify fields where more than 90% of records have the field blank. These are unused fields.
2. For each unused field, determine: is it referenced in any Flow, validation rule, report, or formula? If not, it's safe to delete.
3. Remove unused fields from page layouts first (non-destructive). After 30 days with no complaints, delete the field.

Layer 5: Ongoing Monitoring

Governance without monitoring is just good intentions. Set up a "Data Health" dashboard with 4 components:

1. Total records by object — baseline. Track monthly.
2. Records created this month — inflow rate. Sudden spikes mean a new integration or import happened.
3. Duplicate record set count — should be declining after you implement duplicate rules.
4. Records with missing critical fields — email blank, phone blank, Amount blank on Opportunities. Should be declining.

Check this every Monday. 15 minutes. If any metric moves in the wrong direction, investigate immediately before the problem compounds.

The Implementation Order

1. Week 1: Add the 5 validation rules. Announce to the team with a brief explanation.
2. Week 2: Standardize your 3 most critical picklists (Lead Source, Industry, Stage).
3. Week 3: Run the field audit. Remove unused fields from page layouts.
4. Week 4: Build the Data Health dashboard. Set a recurring calendar reminder to check it weekly.

After this month, your data quality will improve measurably. New bad data is blocked at entry. Existing bad data is being monitored. And you have a framework for ongoing governance that prevents the slow drift back toward chaos.