Most Salesforce orgs don't fail dramatically — they decay slowly. A field nobody uses. An automation that fires for the wrong record type. A duplicate account nobody has time to merge. Individually, each issue is minor. Collectively, they erode trust in your data and slow your team down in ways that are hard to quantify but impossible to ignore.
Here are five patterns I consistently see in orgs that need a cleanup — and what to actually do about them.
1. Your team works around Salesforce instead of in it
When reps start keeping their own spreadsheets, updating Salesforce "just for the manager," or defaulting to WhatsApp for deal updates — that's a signal. It means the system isn't serving them. Either the fields are wrong, the views are cluttered, or the data entry burden is too high.
Interview your heaviest users. Ask them what they actually need to do their job, what they skip in SF, and what they wish existed. Redesign your page layouts and required fields around real workflows, not ideal ones.
2. You have more than 20% duplicate records
Duplicates happen. Someone creates a new lead that already exists as a contact. An import runs without a dedup check. A name is spelled two different ways. But when duplicates hit double digits as a percentage of your total records, reporting becomes unreliable and automation misfires.
"Our reports showed 40 open opportunities with the same account — turns out 30 were duplicates. We were measuring momentum that didn't exist."
Use Salesforce's built-in Duplicate Management rules, or a tool like RingLead or DemandTools for deeper cleanup. Then build matching rules that prevent future duplicates at the point of entry.
3. Automation runs but nobody knows why
If you can't answer "what happens when a lead is converted?" without spending 20 minutes in Setup, your automation layer has grown without governance. Flows, Process Builder (legacy), Workflow Rules, and Apex triggers can all be firing on the same record — sometimes conflicting.
- Audit all active automations in Setup → Flows and Setup → Workflow Rules
- Look for record-triggered flows with no description and no owner
- Test a complete lead-to-close path in a sandbox to see what actually fires
- Document every automation: what triggers it, what it does, who owns it
4. Your field count has grown out of control
Every request becomes a field. Fields accumulate. Page layouts get crowded. Reps start ignoring sections. Report builders can't find what they need. I've seen orgs with 200+ fields on the Account object, fewer than 30 of which are actually used.
Run a field usage report using Salesforce Field Trips (AppExchange, free) or query the FieldUsageMetrics object. Any field with under 5% population that isn't formula-driven is a candidate for archiving or removal.
5. Your reports don't match your spreadsheets
When finance and sales have different revenue numbers, and both are "pulling from Salesforce," something is wrong with your data model or your field definitions. This is usually a combination of missing required fields, ambiguous picklist values, and date/currency field inconsistencies.
Pick your one most important report — pipeline, revenue, or activity — and trace every field it depends on back to its source. Fix the source first, then rebuild the report on clean data.
The bottom line
A Salesforce cleanup isn't a one-time event — it's a practice. The orgs that stay clean have governance: a naming convention, a change log, a person who owns the org architecture, and a process for adding new fields or flows.
If your org shows two or more of these signs, a structured audit is worth the investment. The cost of bad data compounds quietly until it's visible everywhere — in missed forecasts, broken automations, and teams that stopped trusting the system.