How to Clean Up a Messy Salesforce Org in One Afternoon

October 29, 2026

Most Salesforce orgs under 50 users are not badly built. They are ungoverned — built layer by layer by whoever needed something that week.

Direct answer: To clean up a messy Salesforce org, run the duplicate report first, merge accounts, then audit field usage with the Field Trip app, archive unused fields, and rebuild reports from scratch. Four hours gets you a system people will actually use.

Where do you start?

Duplicates first. Nothing else matters while a rep cannot tell which of three “Acme Inc” records has the real history.

Go to Setup → Data → Duplicate Management and run the standard Account duplicate report. Salesforce’s fuzzy matching catches name variants, so “Acme Inc” and “Acme Incorporated” score as likely duplicates.

Merge carefully. The merge tool keeps one record and moves child records over, but it does not merge all field values — you choose field by field. For accounts with many contacts and opportunities, review each merge rather than using bulk merge. Bulk merge has orphaned opportunity history for teams before.

Expect to resolve 5-15% of accounts in a list that has never been cleaned.

How do you find out which fields are dead?

Install Field Trip from the AppExchange. It is free, and it reports field fill rates and page layout usage.

Sort by fill rate. Anything under 10% fill rate is a candidate for deletion. In a typical messy org, that is 30-60% of custom fields.

Before deleting, check three things: is the field referenced by a report, a workflow, or an integration? Salesforce blocks deletion if it is, which is helpful. And export the field’s data first — deletion is permanent after 15 days in the recycle bin.

Deleting dead fields is the single highest-value hour you can spend in a Salesforce cleanup. Every dead field is a decision a rep has to make while entering data.

What should you do about required fields?

Two competing pressures. Too many required fields, and reps enter garbage to get past the screen. Too few, and you have no data.

The test: would a rep reasonably know this value at the moment they create the record? Opportunity Amount — yes. Lead Source — usually yes. “Expected Decision Criteria” — no, that is discovered later.

A workable rule set for Opportunity: Name, Stage, Amount, Close Date, Account. Everything else optional but prompted. Use page layout sections to group the optional fields rather than marking them required.

How do you rebuild reporting?

Old reports are usually worse than no reports, because people trust them.

  1. List every report in use — ask each rep which three they open weekly.
  2. Delete or archive everything else. Salesforce has no easy archive, so move them into a folder called _Archive and remove sharing.
  3. Rebuild the surviving reports as Lightning reports if they are classic. The filter logic differs and the old ones often quietly break.
  4. Add a single dashboard with four tiles: pipeline by stage, won this quarter, activity per rep, deals with no activity in 14 days.

That dashboard covers what a manager actually checks daily. More tiles do not help.

What is the cost of getting this wrong?

Salesforce pricing starts around $25 per user per month for Starter Suite and climbs past $100 for Sales Cloud Enterprise — plus add-ons for things like duplicate management on some tiers. A ten-person team is spending $6,000-15,000 a year before anyone writes a report.

A messy org with that price tag is a very expensive spreadsheet. The cleanup is not optional maintenance; it is the difference between the tool being used and being resented.

What should you not fix in one afternoon?

Data model changes. If your Account-to-Opportunity structure is wrong, that is a project, not an afternoon. Same for automation — a stack of 40 legacy workflow rules needs a migration plan to Flow, not a quick edit pass.

Also skip permission set redesign on day one. Touching sharing settings can expose data across the org, and that is a one-way mistake nobody forgives.

How does this compare to switching tools?

Usually switching is the wrong answer for a data problem. A team that cannot keep Salesforce clean will make Zoho CRM or Insightly messy within six months, just more cheaply.

Migration makes sense when the problem is structural — you sell three product lines and need three pipelines, or you need project handoff that Salesforce does not do without add-ons.

Quick Summary

  • Run the duplicate report first; merge accounts before anything else.
  • Install Field Trip and delete fields under 10% fill rate — often 30-60% of custom fields.
  • Only require fields a rep would know at creation time.
  • Rebuild reports instead of trusting inherited ones; four dashboard tiles is enough.
  • Do not touch the data model or sharing settings in a one-day cleanup.